Re: [GTALUG] AI - Llama 2 an open source AI that can run on a Raspberry PI

[David Mason via talk]

The things that make it not OSI compliant are:
1) an application with more than 700 million monthly active users (i.e.
Google, Bing, Amazon, and Apple) require getting a licence
2) some moral restrictions (like criminal acts)

I can live with that...

../Dave

On Sat, 13 Jan 2024 at 05:39, Evan Leibovitch via talk 
wrote:

> On Wed, Jan 10, 2024 at 10:47 AM Colin McGregor via talk 
> wrote:
>
>> As noted last evening I am running the open source (but NOT GPL) Llama
>> 2 AI on a Raspberry PI 5.
>
>
> As discussed at the meeting Tuesday night, Llama 2 is *not* Open Source.
> It conforms to neither the OSI Open Source Definition nor the FSF Four
> Freedoms.
> This post from IBM explains how and why it's not:
> https://www.ibm.com/topics/llama-2#Is+Llama+2+open+source?
>
> While the license is broadly open, it's not open source or free software
> in much the same way as the Creative Commons "NC" license is not; it
> restricts use and requires certain entities to get a paid license.
>
> - Evan
>
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list
> https://gtalug.org/mailman/listinfo/talk
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] ot: sort of, is it really impossible to get real cable anymore?

[David Mason via talk]

It looks like this:

https://www.walmart.ca/en/ip/hdmi-to-rca-cable-hdmi-to-rca-converter-adapter-cable-1080p-hdmi-to-av-3rca-cvbs-composite-video-audio-supports-for-amazon-fire-stick/PRD700H6NJLEWBM
will convert a digital HDMI to an analogue signal that could be used for a
VCR or old TV.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Linux Unicorns

[David Mason via talk]

Hugh wrote:
There is room for innovation in infrastructure but it is hard to
become a unicorn there since the area is mostly occupied by giants:
- Amazon
- Microsoft
- Google
- Oracle...
Being cloud platform agnostic is smart.  How you do it isn't something
I know much about.

I wonder if Apple uses Linux for anything or if all their services run on
MacOS farms. I couldn't find any useful intel from a Google search. I
wonder if anyone here knows.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] brands matter; Lenovo's brands

[David Mason via talk]

As one of those academics (in Computer Science, no less), I use a Macbook
Pro that I connect to a Dell widescreen at work, and a Phillips 328E1 4K
widescreen at home, which I mostly using in Landscape mode. It's
blisteringly fast, works everywhere, runs forever (like 1/2 a day) on
battery, and has all the software available that I could want (almost
completely FOSS). It's not Linux, but it's Unix in a terminal window. The
majority of my colleagues run Macs, with Windows and Linux about tied for
the rest. My machine was a bit pricey, but many people have Macbook Airs,
which are also great at a great pricepoint.

My servers all run Linux, and I used a Linux laptop until (a) my Linux
wouldn't connect to a projector at a conference, and (b) Apple put Unix
under the GUI. So I've been a happy Mac/Linux user for 2 decades.

../Dave

On Sun, 17 Sept 2023 at 16:01, D. Hugh Redelmeier via talk 
wrote:

> | From: Peter King via talk 
>
> A wonderfully clear and evocative elegy!
> All rational.
>
> | To: D. Hugh Redelmeier via talk 
> | Cc: Peter King 
> | Date: Sat, 16 Sep 2023 13:11:24 -0400
> | Subject: Re: [GTALUG] brands matter; Lenovo's brands
> |
> | All points about brands/branding noted and appreciated.  I got the Lenovo
> | Legion T5 because I wanted a desktop, and it seemed to me that the niche
> for
> | high-quality and well-built desktops -- what used to be called
> "enterprise" or
> | "business" models -- had largely collapsed, being supplanted by either
> | high-end laptops that business users would tote around and if necessary
> plug
> | into a docking station, or by cheap consumer-grade desktops that were
> shoddily
> | built, under-powered, and meant to be thrown away in a few years (more
> | economical than investing in long-lasting hardware that would be
> outmoded too
> | quickly).  So what is someone who wants a good desktop unit to do?
>
> I don't mean to push Lenovo.  I pick them to discuss because I'm slightly
> more familiar with them (mostly on paper).
>
> As far as I can tell, there are still full-sized business towers.  Here's
> a page from Lenovo.
>
> <
> https://www.lenovo.com/ca/lenovopro/en/desktops/subseries-results/?visibleDatas=1033%3ATower%2CTower%25C2%25A0=bestSelling
> >
>
> They aren't particularly inexpensive.  Some might be missing important
> features.  I favour AMD processors but they are under-represented.  None
> is offered with Linux.
>
> Interestingly, they include your model (newer variants).
>
> Lenovo has a separate category "Workstation".  It includes desktop
> ThinkStations and notebook ThinkPads.
> <
> https://www.lenovo.com/ca/lenovopro/en/d/deals/workstations/?sortBy=priceUp=1035%3AThinkStation
> >
>
> Some of their servers (ThinkServer) look a lot like "desktops" (towers
> that go under desks):
> <
> https://www.lenovo.com/ca/lenovopro/en/d/deals/servers/?tabkey=Server%20%26%20Storage%20Deals
> >
> The one inexpensive model has a Celeron processor.
>
> ==
>
> Perhaps you are inferring too much about the quality of your Legion box
> from a few datapoints.
>
> ==
>
> My desktop is a decade old (HP Envy).  The only things I changed inside
> the box in that time (mostly in the first year):
>
> - replaced the GPU (to drive my high resolution displays)
>
> - added RAM
>
> - added an SSD
>
> For my next desktop, I expect to live with the iGPU.  That leaves disk
> bays the only issue with SFF or smaller boxes.
>
> Personally, I don't really fill up modern disks (they are big!).  I do
> want backups, but they need to be separate from the computer anyway.
>
> | A few years ago one of my desktop units failed.  I replaced it with a
> miniPC,
> | a minisform model I put more RAM and two 2TB SSDs into, and it runs just
> | fine.  Maybe that is the way to go.
>
> Ah, so you are quite familiar with that form factor.
>
> |   (I have a portable high-resolution LCD
> | screen now, and I think I'll eventually just carry around miniPCs rather
> than
> | laptops.)
>
> A laptop really is more portable than mini PC + display + keyboard +
> mouse.
>
> | But then again I also have a 14-year-old ThinkPad that still runs
> | like a dream once I put in an SSD; one of the last models with the
> "real" IBM
> | keyboard in it.
>
> 14 year old laptops have processors that use a lot of power when
> sleeping.
> Kind of annoying in a laptop.  They probably don't have USB 3, also
> annoying.  They don't have good displays.  They are surely heavy.  They
> won't have HDMI-out.  The DisplayPort probably cannot drive UltraHD.  By
> now, the battery is probably worn out and it isn't easy to get decent
> replacements.  The oldest processors generation that I'm happy to use in
> notebooks is "Haswell" (launched 10 years ago).
>
> We still use a ThinkPad T530 (Ivy Bridge), but always plugged in and
> without an external monitor.
>
> | Perhaps mistakenly, I thought that the combination of new hardware with
> the
> | rough requirements gamers have for their 

Re: [GTALUG] internet service speed test tip

[David Mason via talk]

I understand the vendors may optimize parameters, but I figured that
UofToronto wouldn't be affected by that, and the numbers were comparable...
but perhaps they tune at a different point in the pipeline.

../Dave

On Fri, 25 Aug 2023 at 17:14, o1bigtenor  wrote:

> On Fri, Aug 25, 2023 at 10:06 AM David Mason  wrote:
> >
> > Speedtest (the app) gives consistent results for a variety of test
> sites, including UofToronto, Bell Canada, Primus. All are significantly
> higher numbers than testmy.net (320Down, 450Up versus 190Down, 215Up on
> testmy.net Toronto site).
> >
> >
> AIUI 'speedtest' is most often optimized by the various vendors to
> give 'great looking numbers'.
> Testmy.net, imo anyway, seem to be more 'real world numbers.
> The graphs of the test I find very very revealing - - - - after a
> while one gets a really good idea of the
> connection possibilities.
>
> Thanks for sharing
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] internet service speed test tip

[David Mason via talk]

Speedtest (the app) gives consistent results for a variety of test sites,
including UofToronto, Bell Canada, Primus. All are significantly higher
numbers than testmy.net (320Down, 450Up versus 190Down, 215Up on testmy.net
Toronto site).

../Dave

On Fri, 25 Aug 2023 at 08:53, o1bigtenor via talk  wrote:

> On Fri, Aug 25, 2023 at 7:39 AM Alvin Starr via talk 
> wrote:
> >
> > On 2023-08-24 10:50, D. Hugh Redelmeier via talk wrote:
> > > Rogers and Bell have web sites for testing your internet speed.  The
> > > advantage of using one should be that it is a pretty pure test of the
> > > ISP's infrastructure.
> > >
> > I know that Bell has tuned their traffic management rules to prioritize
> > speed test sites.
> >
> > I would be very surprised if Rogers is not also doing the same thing.
> >
> > The Bell and Rogers hosted sites are the most likely to take advantage
> > of traffic management to enhance their stats.
> > So you may get results that are close to the maximum that the network
> > can support but nowhere near what the network is able to deliver under
> > its current load.
> >
> >
> > Network performance testing is best done with tools like iperf.
> >
> I found a website called testmy.net  that seems to do a very good job.
> Gives graphing of the results and shows the variability of connection
> quite well imo.
>
> HTH
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list
> https://gtalug.org/mailman/listinfo/talk
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] The good old days of oreilly.com

[David Mason via talk]

I had purchased several books over the years.

I can successfully login in to members.oreilly.com, but it shows 0
purchases.

...sucks...

On Sun, 30 Apr 2023 at 12:41, sciguy via talk  wrote:

> On 2023-04-30 12:10, James Knott via talk wrote:
> > On 2023-04-30 12:04, sciguy via talk wrote:
> >> I am having very little luck poking around. And to correct what I said
> >> earlier, searching specific titles leads me to a "403" error
> >> (Forbidden), not a 404 as I had said. I searched on Programming Perl,
> >> which I have the second and fourth editions, and got a "403" when I
> >> clicked on the title. The message below reads: "Your free O’Reilly
> >> trial has ended, making this content unavailable to view. For
> >> unlimited access to all O’Reilly has to offer, purchase a membership
> >> here." I also notice that on the titles I see, there are no "purchase"
> >> links, just "read" links. It sounds like for the $50 or so monthly
> >> fee, you can read anything you like. Still not sold on the idea, but
> >> clearly they have changed their marketing model.
> >
> > Search on Google for the book title.  Also, you can try
> > https://members.oreilly.com/ and log in with your ID and password.
>
> Bingo! The link to members.oreilly.com got me what I wanted! Thanks!
>
> Paul
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list
> https://gtalug.org/mailman/listinfo/talk
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] New WiFi router?

[David Mason via talk]

At the suggestion of someone here, I installed a UniFi router (where it
comes into the house and supports several wired Ethernet connections) and
remote Access Point at a central point in the house. Couldn't be happier!

../Dave

On Wed, 8 Mar 2023 at 07:20, Alex Kink via talk  wrote:

> I would suggest one of those ridiculous looking TP-Link units with several
> antennas. I've installed them here and there and have one of the "WiFi6"
> capable units at home and they perform very well for what they cost.
> Not suggesting any particular model, just whatever fits in the budget.
>
> > On Mar 8, 2023, at 01:09, William Park via talk  wrote:
> >
> > Hi all,
> >
> > I was upgraded (by my ISP) from Cable 75 (75M/10M) to Cable 100
> (100M/30M).  It turns out my router (Asus RT-N66U) maxes out at 85M/30M.
> Connecting directly to the modem, I get full 100M/30M.
> >
> > So, it's time for a new router.  It's been long time since I shopped for
> a router.  Which one do you recommend?
> > ---
> > Post to this mailing list talk@gtalug.org
> > Unsubscribe from this mailing list
> https://gtalug.org/mailman/listinfo/talk
>
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list
> https://gtalug.org/mailman/listinfo/talk
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Weird pivot from the Linux Foundation: Overture Maps Foundation

[David Mason via talk]

I think this is actually much more positive than you think:

https://wiki.osmfoundation.org/wiki/CDLA_permissive_compatibility

CDLA Permissive 2 seems very open, and OSM seems to agree.

../Dave
On Dec 15, 2022 at 2:59 PM -0500, Stewart Russell via talk , 
wrote:
> On Thu, Dec 15, 2022 at 12:28 PM Evan Leibovitch via talk  
> wrote:
> > > From the FAQ:
> > >
> > > > Data contributed to ODbL licensed datasets will be contributed under 
> > > > both the ODbL and CDLA permissive v2. Contributions to CDLA permissive 
> > > > v2 datasets will be contributed under the CDLA permissive v2.
> > >
> > > I don't know these licenses. Are they open enough such this project's 
> > > data can be used by OSM?
>
> ODbL is the licence used by OSM. Its background is more from European 
> database copyright than open source. It's a share-alike licence with required 
> attribution. CDLA I'd never heard of, but it seems like a "let's make the MIT 
> license, but for data". Many VC-funded startups balk at the share-alike and 
> attribution requirements of OSM, because they just want the free data and 
> earn money from contributors' work.
>
> This looks like Overture can consume OSM data, but OSM can't use Overture's 
> data easily. What a surprise.
>
>  Stewart
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] tpl Linux training?

[David Mason via talk]

Does this include old Macs, too? I have several that are sitting in a closet 
because they don’t run a currently supported MacOSX.

../Dave
On Sep 1, 2022, 6:34 PM -0400, Joseph Rocklin via talk , wrote:
> Hi,
>
> I'm running a charity called Rocklin Reboots. We're CRA recognized. The gist 
> of what we do is receive machines no longer needed by private citizens and 
> industry, use mitary data destruction for the hard drives, and put Linux on 
> the machine to give to those in need of functional computers for work or 
> study but can't afford new, off-the-shelf computers. We give tax receipts for 
> donations of computers and also from any financial contributions from 
> recipients receiving computers with Linux.
>
> The reason for my mentioning this here is that I was giving thought to 
> reaching out to FGT to see how we might integrate. I just can't see how yet. 
> Any ideas?
>
> Also, please note it is taking me a while to get my act together with a good 
> website and advertising. But so far you can look at 
> https://RocklinReboots.com.
>
> You can reach me at rock...@rocklinreboots.com if you have any computer 
> donations you'd like to make.
>
> Thank you for your time folks,
> Joseph Rocklin
> Founder and Administrator at Rocklin Reboots
>
>
> Sept 1, 2022, 6:08 p.m. by talk@gtalug.org:
> > Yes, Colin was who I recalled as well.
> >
> >
> >
> > On Thu, 1 Sep 2022, Stewart C. Russell via talk wrote:
> > > On 2022-09-01 11:16, Karen Lewellen via talk wrote:
> > > >
> > > > I have a memory of someone from Toronto freegeeks  > > > around?>
> > > > who used  to teach Linux classes, showcasing I believe an option outside
> > > > of windows.
> > >
> > > Yup, Free Geek Toronto (FGT) are still around
> > > (https://www.freegeektoronto.org/) but I don't think educational part of 
> > > what
> > > they do has been running for a couple of years. I believe that Colin 
> > > McGregor
> > > - who used to be very active in GTALUG - taught some classes at FGT.
> > >
> > > Stewart
> > >
> > > ---
> > > Post to this mailing list talk@gtalug.org
> > > Unsubscribe from this mailing list 
> > > https://gtalug.org/mailman/listinfo/talk
>
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] LLVM JIT Talks

[David Mason via talk]

I am building a Smalltalk interpreter/compiler (in a great new systems language 
called Zig), and I would like to generate executable code on the fly.

../Dave
On May 25, 2022, 11:57 PM -0400, Nicholas Krause via talk , 
wrote:
> Greetings
>
> I've misdeleted the other thread, so sorry for starting a new one. For David, 
> I would be happy to try if you let me know what the interest/use case is. On 
> a side note, for Alan I've
> asked him to contact you directly. Hopefully he contacts you soon. If he does 
> not in the next few weeks, I will ping him for you.
>
> Nick
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Possible Future Talks

[David Mason via talk]

I’d be interested in how one uses the LLVM JIT

../Dave
On May 23, 2022, 5:05 PM -0400, Nicholas Krause via talk , 
wrote:
> Greetings,
> I've managed to get in contact with the maintainer for the LLVM RiscV 
> backend. Not sure if it's of interest to the group.
> But, he can talk about compiler backends in general or related to RiscV. For 
> those who aren't aware as discussed at the
> RiscV talk, this would be the compiler support for the ISA for RiscV. Or if 
> there is another topic from the LLVM side
> I can try to find someone there.
>
> Take care,
> Nick
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] ot: perhaps, headphones?

[David Mason via talk]

On Jan 14, 2022, 11:25 AM -0500, GTALUG Talk , wrote:
> I have to add though, that I won't be buying Bose going forward for other 
> reasons however.
Of course, when you say that everyone (or at least me) wonders what those 
reasons might be. :-)

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] keyboards

[David Mason via talk]

On Oct 23, 2021, 2:11 PM -0400, William Park via talk , wrote:
> Thanks for info. Do you have "clicky" or "tactile" keys?
The Matias keyboard I have is marketed as an Apple replacement, and it has the 
same feel as the newest Apple keyboard, which means that there is some 
resistance to the keypress (and the same resistance throughout the keystroke), 
and there is a bit of a tactile click but it is fairly quiet. There was an 
issue with key feel for apple (I think laptop) keyboards a few years ago, but I 
think most people like the current keys (duplicated on my Matias). They may 
have other key dynamics available, I’m not sure.

Hope that helps.

../Dave
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] keyboards

[David Mason via talk]

I am not the keyboard geek that some of you obviously are, however I would be 
remiss to not point out a Canadian company:

https://matias.ca/ergopro/programmable/

I don’t have their ergo (split) keyboard, but I do have 2 of their keyboards 
and quite like them. I just got a new iMac with the beautiful colour-coded 
keyboard, but immediately miss the backlighting of the Matias keyboard. I am 
not a perfect touch typist, although quite fast. I use my keyboard mostly for 
programming.

This conversation does make me curious about the split keyboard. Is Giles “took 
a month to get used to it” typical?

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Looking for assistance with Firefox

[David Mason via talk]

I have a colleague who has a great workflow. Iff he wants to do research on 
something, he opens a fresh window… opens as many tabs as necessary for the 
project, then closes the window with all its tabs.

I haven’t been able to integrate the discipline yet (my tab/window collection 
sounds a lot like Lennart’s), but it seems like a good model.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] [GTALUG-Announce] Meeting Tomorrow

[David Mason via talk]

I just thought I’d point out that the meeting page gets screwy if you look at 
it with a window wider than about 80% of a 2560 screen (the left navigation bar 
slides over the content block). Same on Firefox, Chrome, and  Safari.

../Dave
On Aug 9, 2021, 10:03 PM -0400, hi--- via talk , wrote:
> https://gtalug.org/meeting/2021-08/
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] e-reader recommendations

[David Mason via talk]

Sorry, just noticed I wrote “it will also read” but I didn’t say what!

It does epub well… doesn’t do mobi.

../Dave
On May 12, 2021, 6:53 AM -0400, o1bigtenor via talk , wrote:
> On Tue, May 11, 2021 at 9:00 PM David Mason via talk  wrote:
> >
> > I have a reMarkable https://remarkable.com/ which is great for PDFs, 
> > writing notes, etc. and it will also read.
> >
> > It’s quite nice. I generally use it instead of paper for notes. You can 
> > even transfer web pages to it (from a Chrome plug-in) to read them in more 
> > comfort.
> >
> An interesting tool!
> For $750 inc taxes - - - - dunno.
> Find the idea that handwriting is wonderful just a little goofy - - -
> - my typing is easily 3x my handwriting speed - - - - accuracy is
> likely far better than any ocr program as well.
>
> The limitations of these devices are quite interesting.
> The whole genre seems like toys for those with lots of toys.
> It would help if technical books were not so expensive and so hard to
> get. (Reader use)
> Oh well - - - not designed for my kind of uses.
>
> Regards
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] e-reader recommendations

[David Mason via talk]

I have a reMarkable https://remarkable.com/ which is great for PDFs, writing 
notes, etc. and it will also read.

It’s quite nice. I generally use it instead of paper for notes. You can even 
transfer web pages to it (from a Chrome plug-in) to read them in more comfort.

../Dave
On May 11, 2021, 6:27 PM -0400, Lennart Sorensen via talk , 
wrote:
> On Tue, May 11, 2021 at 01:40:03PM -0400, Trevor Woerner via talk wrote:
> > As I expected, there's lots of really good feedback :-)
> >
> > I should have clarified that I have lots and lots of tablets and phones and
> > all those sorts of devices, but I've never had an e-reader and I'm curious
> > enough to at least want to try one (mostly for battery life, eye strain,
> > and general impressions). Ideally I could just buy one and it would be
> > great, rather than having to try a bunch of them before finding one I like
> > :-)
>
> Well if you want something with great battery life that is great for
> reading ebooks in daylight, an ereader is great. For the things you
> listed though, they are useless.
>
> So if you want to carry 200 books with you, they are fantastic. They
> remember what page you were on in each book. Very handy for book worms.
>
> They are very much not generic computing devices at all though. They do
> one thing well and that's it. I know sony tried doing mp3 support for
> audio books on early models and dropped it later since it drained the
> battery and was no match for an ipod shuffle for audio books.
>
> --
> Len Sorensen
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Idiomatic programming versus micro-optimizations

[David Mason via talk]

On Apr 9, 2021, 2:55 PM -0400, Aruna Hewapathirane 
, wrote:
> This is simply an attempt to inform and educate ( i am not trying to start 
> any flame wars ). This article I feel describes aptly why C will not go away.
> https://drewdevault.com/2019/03/25/Rust-is-not-a-good-C-replacement.html
> Thanks - Aruna ::

Thanks, Aruna. It’s a good article. I don’t necessarily agree with all the 
premises, nor conclusions - most particularly the “we don’t need no 
parallelism” part - but there is legitimacy in the argument that Rust is more a 
C++ replacement than a C replacement.

I think the lifetime analysis is a game-changer. I’d much rather program in 
Rust, even with the “cruft” than in C (for the record I have written C 
compilers and have well over 100K lines of C code under my belt), because I 
think the safety and concurrency are critically important. Go is interesting, 
but has never caught my interest, partly because it has a garbage collector 
(not that I have a problem with garbage collection…. just if I’m going to use a 
GC’ed language I have a lot better options than Go).

Perhaps the eventual C replacement is a stripped-down Rust - lifetimes, type 
inference, safety + concurrency, but not much more.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Google wins over Oracle in Java API copyright suit

[David Mason via talk]

On Apr 9, 2021, 1:45 PM -0400, Stewart C. Russell via talk , 
wrote:
> After all with a loop you are controlling the
> execution order of the processing.  If done right you usually shouldn't
> need to care.
>
>
> But in document processing, you really really /really/ want the output
> to come out in the same order as the input. Which is why functional
> languages seemed a strange choice for document transformation. The
> absence of side-effects can be handy in document processing, but being
> in the right order is usually what publishing houses get paid the big
> bucks to do.

The key part of what Lennart wrote is “if done right”. How could you imagine 
that the functional program would return the results out of order? 
Compilers/interpreters are allow to make whatever optimizations they want, as 
long as time and memory consumption are the only things that change from the 
original!
> I've had to process utility time series power generation data in XSLT.
> That was horrid. Order matters a lot there, too.
XSLT is a very specific functional language, designed for a very particular 
job, which (from my limited experience) it does somewhere between “adequately" 
and “very well”. But while it may be Turing complete, it is not a general 
purpose language and I would not want to program such problems with it!!! Pure 
lambda-calculus and Turing machines are also Turing complete, but you sure 
wouldn’t want to be programming in those either.

Haskell, Scheme, OCaML, Erlang/Elixir, Scala are much, much friendlier 
languages for general purpose programming. All the ML dialects, as well as 
Scheme and Scala support more imperative programming as well, that can help 
with the transition.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

[GTALUG] Idiomatic programming versus micro-optimizations

[David Mason via talk]

On Apr 9, 2021, 11:51 AM -0400, Len Sorensen via , wrote:
> But using a loop means you are telling the system how to do things,
> rather than telling it what you want done and letting it (usually) do
> a better job at the how. After all with a loop you are controlling the
> excution order of the processing. If done right you usually shouldn't
> need to care.
Another fine example of this is with Rust where you can write very natural, 
idiomatic, safe, code and then trust the compiler to mostly optimize most 
perceived overhead away. In fact there are often optimizations that Rust can do 
that C compilers can’t do.

Here is a very interesting series of articles where someone takes a heavily 
optimized C program, converts it to Rust - with progressively more idiomatic 
versions, and ends up with a Rust program 3% faster than C (clang) and 20% 
faster than C (gcc), then rewrites it using the natural Rust iterators rather 
than indexing loops and it’s now 27% faster than clang!! 
http://cliffle.com/p/dangerust/
> But yes functional languages require a different philosophy. Functional
> languages are not for people that want to micromanage the computer.
The point is, apart from academic exercises, the percentage of people who 
*need* to micromanage the computer is getting vanishingly small. I would say 
measured in the low thousands, but hundreds is probably more accurate. Unless 
you’re writing a compiler, interpreter or for an application where the computer 
has to add no more than $0.25 to the cost of the product, you really shouldn’t 
care.

It really is time for C to go away! But TIOBE doesn’t agree 
https://www.tiobe.com/tiobe-index/ with C back to #1 and assemble up to #14 
(While Rust is #29) Just shows how important my opinion is!!

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Google wins over Oracle in Java API copyright suit

[David Mason via talk]

Haskell, Scheme, Lisp, Elixir (mostly), Elm, Clojure

Static OO is kinda an oxymoron. Smalltalk, Ruby, Python are good OO (in my 
highly opinionated opinion).

../Dave
On Apr 8, 2021, 12:29 PM -0400, William Park via talk , wrote:
> On 4/8/21 11:37 AM, Lennart Sorensen via talk wrote:
> > If I had my way, functional languages would be what is used, and
> > definitely not any that were object oriented, at least not in the way C++
> > and Java are. Multiple inheritance should not exist.
> Can you throw few names of "functional languages" at us non-compsci folks?
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] RISC-V based GPUs?

[David Mason via talk]

Thanks, Ivan.

There’s a really interesting link in there about Larabee and Intel politics: 
https://tomforsyth1000.github.io/blog.wiki.html#%5B%5BWhy%20didn%27t%20Larrabee%20fail%3F%5D%5D

../Dave
On Jan 31, 2021, 8:22 AM -0500, Ivan Avery Frey via talk , 
wrote:
> https://www.tomshardware.com/amp/news/risc-v-open-source-gpu-nvidia-intel-amd-arm-imagination
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] [GTALUG-Announce] On the recent GTALUG Outage

[David Mason via talk]

Hi,

I’m Chair of the Department of Computer Science at Ryerson, and we already 
provide GTALUG with a meeting room (when there’s no pandemic and it’s safe to 
meet in person).

I see no reason we couldn’t provide a virtual machine to run GTALUG’s 
visibility. I can’t imagine there is any significant amount of data or 
bandwidth required. So we would be happy to assist.

../Dave
On Jan 7, 2021, 2:29 PM -0500, Rouben via talk , wrote:
> I work at UofT in IT and can inquire to see if UofT would be interested in 
> hosting GTALUG infrastructure. Would this be of interest to the community? 
> Any reasons (eg. reservations/requirements/bylaws/rules) against this?
>
> Rouben
>
> > On Thu, Jan 7, 2021 at 14:17 D. Hugh Redelmeier via talk  
> > wrote:
> > > | From: Scott Sullivan 
> > >
> > > | We have great thanks to express to Hugh Redelmeier. A steadfast 
> > > volunteer and
> > > | contributor of our community. Uncoordinated with the board, he reached 
> > > out our
> > > | provider and offered to pay the bill. Successful by-passing our
> > > | misunderstandings, hotfixing the situation, and bringing us back 
> > > on-line. He
> > > | will be compensated for an out of pocket costs.
> > >
> > > 1. The fix is temporary.  Linode has reinstated the server for the rest of
> > >    the billing cycle (I don't know how long that is; I've asked).  They
> > >    will get back to me about what I/we can do.
> > >
> > > 2. I offered to pay.  I don't know if they actually charged me.  If so,
> > >    I'm happy to contribute that to GTALUG.  Thanks for the several offers
> > >    of financial help.  Don't worry about it (unless this becomes
> > >    permanent).
> > >
> > > 3. I explicitly did not represent myself as authorized to linode.  I have
> > >    no admin access or anything like it.  GTALUG needs admin access to the
> > >    server or it needs to replace the server.  In other words: we still
> > >    have a problem.  This just buys GTALUG a bit of time.
> > >
> > > | Speaking personally, these are hard times. I as I've returned to help 
> > > bridge
> > > | the gap from Chris's passing, I've found to many functions and 
> > > knowledge and
> > > | pool in to few people and was not being effectively taught to new
> > > | contributors.
> > >
> > > This is a real problem with community groups.  It can be addressed but it
> > > isn't easy.  GTALUG has done amazingly well.
> > >
> > > | In the past, the board was just the voting head of a rotating group of 
> > > some
> > > | 10-15 individuals that contributed and attended the executive meetings.
> > >
> > > That seems like a very healthy number, based on my experience with other
> > > groups.
> > >
> > > ---
> > > Post to this mailing list talk@gtalug.org
> > > Unsubscribe from this mailing list 
> > > https://gtalug.org/mailman/listinfo/talk
> --
> Rouben
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Right to Repair Article in NYT

[David Mason via talk]

On Oct 29, 2020, 4:19 PM -0400, D. Hugh Redelmeier , wrote:
>
> The distinction in the case of excrement is just a matter of idiomatic
> English. "Cowshit" is just not a common expression. I've
> never heard "steershit" or "heifershit".

The closest I’ve ever heard is “off in a cloud of heifer dust” (which, in a 
corral, is dirt mixed with desiccated and pulverized cow manure). It’s also the 
name of a book: 
https://burnstownpublishing.com/product/off-in-a-cloud-of-heifer-dust-%E2%94%80-some-ottawa-valley-yarns/
 and a spice-mix.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Dice Keys

[David Mason via talk]

A riff off: https://en.wikipedia.org/wiki/Diceware

../Dave
On Aug 28, 2020, 11:15 AM -0400, Christopher Browne via talk , 
wrote:
> Here's a cool thing I saw recently...
>
> https://www.schneier.com/blog/archives/2020/08/dicekeys.html
>
> The intention of this parallels the various Bitcoin "Solid Steel Passphrase 
> Wallet" items that were popular a year or so ago
> (See https://www.toughgadget.com/bitcoin-crypto-metal-recovery-seed-wallets/, 
> https://www.buybitcoinworldwide.com/wallets/steel/ )
>
> It's a case for a set of 25 dice that looks like a Boggle game set; it will 
> generate and "record" what ought to be a Sooper Seekrut key as would be used 
> for things like:
>  - master key for password manager
>  - U2F key for 2 Factor Authentication
>  - Secret key for cryptocurrency wallet
>
> By being a set of dice with a nice plastic box to hold them securely, this is 
> not vulnerable to various threats common to electronic devices:
>  - EMP (for those highly worried about nuclear devices)
>  - Water damage
>
> Of course, if all your disk drives get toasted, there might not be any data 
> left to decrypt or systems to connect to.  And plastic will melt away or burn 
> when exposed to fire...
>
> But it's pretty cool, I'm tempted to grab a set.
>
> There's a web app: https://dicekeys.app/
>
> It appears that this application, embedded in a single JavaScript file, runs 
> locally, inside your browser, so that usual criticisms about it being a giant 
> security vulnerability of sharing your key with their web site seems like it 
> mightn't apply.  How to confirm in an authoritative way that nothing is 
> *actually* shared seems like the fun security question.
> --
> When confronted by a difficult problem, solve it by reducing it to the
> question, "How would the Lone Ranger handle this?"
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] web hosting question

[David Mason via talk]

I use https://www.ovh.com for hosting, although 
https://www.canadianwebhosting.com/ looks good, too.

I use https://easydns.com/ for DNS registry and DNS hosting.

I’m very happy with both easydns and ovh!

../Dave
On Jul 29, 2020, 1:10 PM -0400, o1bigtenor via talk , wrote:
>
>
> > On Wed, Jul 29, 2020 at 11:43 AM Alvin Starr via talk  
> > wrote:
> > > On 7/29/20 12:19 PM, Don Tai via talk wrote:
> > > > A friend on here uses Planethoster, a Canuck company based in Montreal
> > > > but also have servers in France. He has had good experiences so far,
> > > > but you might need to request English once in a while. Why do you want
> > > > to have a Canuck host provider? This severely limits your choices of
> > > > host provider. I use A2 in Michigan.
> > > >
> > >
> > > The US hosting rules are such that the EU does not allow things like
> > > medical and certain kinds of personal data to reside in the US.
> > > Currently Canada is still considered safe byt the EU.
> > >
> > > Not completely sure how that relates to US companies that have Canadian
> > > centers.
> >
> > I will admit to being more than somewhat paranoid over bureaucratic
> > oversight and intervention. OUr Canuckistani versions are problematic enough
> > so I'd rather not add even more 'fun'.
> > >
> > > A couple of general rules to go by when hosting your data somewhere.
> > > 1) your data may be monetized without your knowledge.
> > > 2) whoever hosts your data owns your data.
> > >
> > That's why I'm a little hesitant.
> > At a certain point a website is a reasonable idea (as a business decision).
> > I don't think I really want to run a website out my personal systems along
> > with all the other stuff I have to cover for the business. The day tends to
> > already be far too short!
> >
> > Thanks for the reminder!
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] (question) GPU + Data center = ?

[David Mason via talk]

Thanks, Hugh! Nicely explained.

../Dave
On Jul 14, 2020, 10:15 AM -0400, D. Hugh Redelmeier via talk , 
wrote:
> | From: David Mason via talk 
>
> | The short answer is: Machine Learning (and other data-mining-like 
> applications)
>
> A much LONGER answer:
>
> There has been a field of Computing on GPUs for perhaps a dozen years.
> GPUs have evolved into having a LOT of Floating Point units that can
> act simultaneously, mostly in lock-step.
>
> They are nasty to program: conventional high-level languages and
> programmers aren't very good at exploiting GPUs.
>
> NVidia's Cuda (dominant) and the industry standard OpenCL (struggling)
> are used to program the combination of the host CPU and the GPU.
>
> Generally, a set of subroutines is written to exploit a GPU and those
> subroutines get called by conventional programs. Examples of such a
> library: TensorFlow, PyTorch, OpenBLAS. The first two are for machine
> learning.
>
> Some challenges GPU programmers face:
>
> - GPUs cannot do everything that programmers are used to. A program
> using a GPU must be composed of a Host CPU program and a GPU
> program. (Some languages let you do the split within a single
> program, but there still is a split.)
>
> - GPU programming requires a lot effort designing how data gets
> shuffled in and out of the GPU's dedicated memory. Without care,
> the time eaten by this can easily overwhelm the time saved by using a
> GPU instead of just the host CPU.
>
> Like any performance problem, one needs to measure to get an
> accurate understanding. The result might easily suggest massive
> changes to a program.
>
> - Each GPU links its ALUs into fixed-size groups. Problems must be
> mapped onto these groups, even if that isn't natural. A typical size
> is 64 ALUs. Each ALU in a group is either executing the same
> instruction, or is idled.
>
> OpenCL and Cuda help the programmer create doubly-nested loops that
> map well onto this hardware.
>
> Lots of compute-intensive algorithms are not easy to break down into this
> structure.
>
> - GPUs are not very good at conventional control-flow. And it is
> different from what most programmers expect. For example, when an
> "if" is executed, all compute elements in a group are tied up, even
> if they are not active. Think how this applies to loops.
>
> - each GPU is kind of different, it is hard to program generically.
> This is made worse by the fact that Cuda, the most popular language,
> is proprietary to NVidia. Lots of politics here.
>
> - GPUs are not easily safe to share amongst multiple processes. This
> is slowly improving.
>
> - New GPUs are getting better, so one should perhaps revisit existing
> programs regularly.
>
> - GPU memories are not virtual. If you hit the limit of memory on a
> card, you've got to change your program.
>
> Worse: there is a three or more level hierarchy of fixed-size
> memories within the GPU that needs to be explicitly managed.
>
> - GPU software is oriented to performance. Compile times are long.
> Debugging is hard and different.
>
> Setting up the hardware and software for GPU computing is stupidly
> challenging. Alex gave a talk to GTALUG (video available) about his
> playing with this. Here's what I remember:
>
> - AMD is mostly open source but not part of most distros (why???).
> You need to use select distros plus out-of-distro software. Support
> for APUs (AMD processor chips with built-in GPUs) is still missing
> (dumb).
>
> - NVidia is closed source. Alex found it easier to get going. Still
> work. Still requires out-of-distro software.
>
> - He didn't try Intel. Ubiquitous but not popular for GPU computing
> since all units are integrated and thus limited in crunch.
>
> Intel, being behind, is the nicest player.
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] (question) GPU + Data center = ?

[David Mason via talk]

The short answer is: Machine Learning (and other data-mining-like applications)

../Dave
On Jul 13, 2020, 11:19 PM -0400, William Park via talk , wrote:
> Hi all,
>
> Perhaps off topics... I keep coming across "GPU in Data Centers", ie.
> Nvidia. Exactly what does GPU have ANYTHING do with data centers? I
> can't imagine people setting up a server in the cloud, and then using
> that to process graphics or play games.
> --
> William Park 
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Bash does-directory-exist question

[David Mason via talk]

(base) : ~/foo ; [ -w .. ] && echo true
true
(base) : ~/foo ; /bin/pwd
pwd: .: No such file or directory
(base) : ~/foo ; pwd
/Users/dmason/foo
(base) : ~/foo ; [ -w $PWD ] && echo true
(base) : ~/foo ;

So, /bin/pwd fails and [ -w $PWD ] also fails, as John hypothesized

../Dave
On Jul 10, 2020, 11:01 AM -0400, John Sellens via talk , wrote:
> On Fri, 2020/07/10 09:38:48AM -0400, Giles Orr via talk  
> wrote:
> | This gives immediate visual feedback on the write-status of the
> | current directory. But test's '-w' and '-d' both claim that you're
> | still in a valid directory under the above circumstances. Does anyone
> | know of a simple way to find out if the directory you're currently in
> | actually exists?
>
> The directory "." will still exist while you have it open (your current
> directory), but will be unreachable, as you observed with stat(1) and
> the number of links.
>
> Would checking for "test -d $PWD" work? I think $PWD is the full path
> and so if it's no longer reachable, the test should fail?
>
> Hope that helps
>
> John
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

[GTALUG] I’m obviously way behind in my reading: IBM owns Redhat

[David Mason via talk]

This is not news… except to me, but I thought the article might be interesting 
to someone.

https://www.popularmechanics.com/technology/a24397193/ibm-red-hat/

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] [GTALUG-Announce] Meeting Tomorrow

[David Mason via talk]

You might want to consider Jitsi in the future… it’s open source and seems to 
work very well (I was an attendee to a Jitsi meeting last night) and looks like 
easy to install https://www.youtube.com/watch?v=IQRwtUamHQU

../Dave
On May 11, 2020, 1:38 PM -0400, hi--- via talk , wrote:
> 
>
> # Lightning talks and Round Table
>
> This is GTALUG's version of an un-conference, a loosely structured short 
> talks emphasizing the informal exchange of information and ideas between 
> participants, rather than following a conventionally structured GTALUG 
> meetings.
>
> If you already have a topic in mind please send an email speak...@gtalug.org 
> to be added to the list of scheduled talks.
>
> A Round Table Q Session is where we take questions from the audience, and 
> the audience discusses and attempts to answer those questions from their own 
> knowledge. It's a great way to meet fellow members of our community and 
> discover the skill sets we each bring to the table.
>
> Zoom meeting URL: 
> https://us02web.zoom.us/j/82232483670?pwd=SEp6eU1FWUNCTDJncllqWFYzN292dz09
> Meeting ID: 822 3248 3670
> Password: 093063
>
> One tap mobile:
>
> * tel:+14388097799,,82232483670
> * tel:+15873281099,,82232483670
>
> Dial by your location(Canada):
>
> * tel:+1-438-809-7799
> * tel:+1-587-328-1099
> * tel:+1-647-374-4685
> * tel:+1-647-558-0588
> * tel:+1-778-907-2071
>
>
> Find your local number: https://zoom.us/u/acqLA3Cd7V
>
>
> ## Schedule
>
> * 7:30 pm - Meeting and presentation.
>
> # Code of Conduct
>
> We want a productive happy community that can welcome new ideas, improve every
> process every year, and foster collaboration between individuals with 
> differing
> needs, interests and skills.
>
> We gain strength from diversity, and actively seek participation from those 
> who
> enhance it. This code of conduct exists to ensure that diverse groups
> collaborate to mutual advantage and enjoyment. We will challenge prejudice 
> that
> could jeopardise the participation of any person in the community.
>
> The Code of Conduct governs how we behave in public or in private whenever the
> Linux community will be judged by our actions. We expect it to be honoured by
> everyone who represents the community officially or informally, claims
> affiliation or participates directly. It applies to activities online or
> offline.
>
> We invite anybody to participate. Our community is open.
>
> Please read more about the GTALUG Code of Conduct here:
> .
>
> If you have any questions, comments, or concerns about the GTALUG Code of
> Conduct please contact the GTALUG Board @ .
> ---
> GTALUG Announce mailing list
> annou...@gtalug.org
> https://gtalug.org/mailman/listinfo/announce
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] On the subject of backups.

[David Mason via talk]

ZFS is another option. And it handles delta-backups very easily.

../Dave
On May 5, 2020, 11:27 PM -0400, Lennart Sorensen via talk , 
wrote:
> On Mon, May 04, 2020 at 10:42:25PM -0400, Alvin Starr via talk wrote:
> > The files are generally a few hundred KB each. They may run into a few MB
> > but that's about it.
> > I use to use ReiserFS back in the days of ext2/3 but it kind of fell out of
> > favor after the lead developer got sent away for murder.
> > Reiser was much faster and more reliable than ext at the time.
> > It would actually be interesting to see if running a reiserfs or btrfs
> > filesystem would actually make a significant difference but in the long run
> > I am kind of stuck with Centos/RH supported file systems and reiser and
> > btrfs are not part of that mix anymore.
>
> ReiserFS was not reliable. I certainly stopped using it long before
> the developer issues happened. The silent file content loss was just
> unacceptable. And it wasn't a rare occurance. I saw it on many systems
> many times. ext2 and 3 you could at least trust with your data even if
> they were quite a bit slower. Certainly these days RHEL supports ext2/3/4
> and XFS (their default and preferred). I use ext4 because it works well.
> GlusterFS defaults to XFS and while technically it can use other
> filesystems (and many people do run ext4 on it apparently) I don't
> believe they support that setup.
>
> > I am not sure how much I can get by tweaking the filesystem.
> > I would need to get a 50x -100x improvement to make backups complete in a
> > few hours.
> > Most stuff I have read comparing various filesystems and performance are
> > talking about percentage differences that is much less than 100%.
> >
> > I have a feeling that the only answer will be something like Veeam where
> > only changed blocks are backed up.
> > A directory tree walk just takes too long.
>
> Well, does the system have enough ram? That is something that often
> isn't hard to increase. XFS has certainly in the past been known to
> require a fair bit of ram to manage well.
>
> --
> Len Sorensen
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Laptop recommendations?

[David Mason via talk]

Way off-topic, but this reminds me of writing “Cold Start” cards for the 
IBM 1130. A card had 80 columns and 12 rows, so you had 80 words of program but 
the instructions needed to be carefully chosen because you only had 12 bits and 
the other 4 bits were filled in by hardware.

The standard one moved itself out of the way and then read the first sector off 
the disk. But someone wrote a “rabbit” card which would reproduce itself on 
every card in the hopper (typically you’d put a cold start card on the front of 
your deck and then hit the Initial Program Load sequence… so the rabbit card 
would destroy the whole deck!! One that I wrote basically made the line printer 
dance around the room (a.k.a. destroy itself if you didn’t stop it quickly).

../Dave
On May 1, 2020, 2:45 PM -0400, Lennart Sorensen via talk , 
wrote:
> On Fri, May 01, 2020 at 12:14:11PM -0400, D. Hugh Redelmeier via talk wrote:
> > Yeah. I only used PDP-8 computers before floppies had been invented.
> >
> > You didn't have to remember the RIM loader for the PDP-8/I, it was
> > silk-screened onto the console.
> >
> > 
> >
> > You didn't often need to use it since its only purpose was to load the
> > BIN loader, and that usually persisted (core memory FTW!).
>
> I wonder if the developers had competitions on who could write the
> shortest working loader program.
>
> --
> Len Sorensen
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Adding all users to the "disk" group: bad idea, or terrible idea?

[David Mason via talk]

On Feb 21, 2020, 10:14 AM -0500, mwilson--- via talk , wrote:
> Would it be as simple to change the permissions on the particular drive,
> and then just let the users have at it?
>
> eg. sudo chmod a+rw /dev/mmcblk0

To be safe, you would have to disallow mounting of filesystems on that drive 
(or any other drive with the same media) because someone could make a 
filesystem on that media, create a SUID program in that filesystem (through raw 
I/O), mount the filesystem, and then run the SUID program. If I recall, there 
is a flag you can set on mount that prevents running of SUID programs on the 
filesystem…. that would also suffice.

../Dave

---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Rust intro

[David Mason via talk]

If you don’t care about safety around the allocations of your particular data 
type, you can certainly have “unsafe” sections of your code that are 
responsible for the allocation/deallocation of your data type. In the unsafe 
code (which can be as little as an expression or as much as a function) you can 
do anything. The point is that the unsafety is limited to sections of code that 
can be identified as unsafe and can be identified by something as crude as grep.

If you haven’t yet, I encourage you to read through the 
http://cliffle.com/p/dangerust/ article. It’s very good, and will demonstrate 
that in Rust, you can be as crude as C or as safe as Haskell, or anywhere in 
between - with very fine granularity.

../Dave
On Jan 2, 2020, 5:16 PM -0500, Nicholas Krause , wrote:
>
>
> On 1/2/20 1:57 PM, David Mason wrote:
> > On Jan 2, 2020, 11:22 AM -0500, Nicholas Krause , 
> > wrote:
> > > Thanks for letting me know. Don't know if it will be solved as that's a 
> > > problem in my view.
> >
> > I was going to say that you could implement a garbage collector in/for 
> > Rust, but first I checked if there was a crate for one, and sure enough, 
> > there is: https://github.com/Manishearth/rust-gc/ for those occasions when 
> > you really need a garbage collector. It’s in development, but looks pretty 
> > usable.
> >
> > ../Dave
> This isn't a issue that can be solved by a garbage collector which are 
> notoriously for having
> these leaks but something like a unsafe mode with something similar to 
> weak_ptr in C++.
>
>
> Not sure what the unsafe Rust implementation would would do but its a 
> outstanding
> issue if you assume that memory has a non circular owner,
>
> Nick
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Rust intro

[David Mason via talk]

On Jan 2, 2020, 11:22 AM -0500, Nicholas Krause , wrote:
> Thanks for letting me know. Don't know if it will be solved as that's a 
> problem in my view.

I was going to say that you could implement a garbage collector in/for Rust, 
but first I checked if there was a crate for one, and sure enough, there is: 
https://github.com/Manishearth/rust-gc/ for those occasions when you really 
need a garbage collector. It’s in development, but looks pretty usable.

../Dave
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Rust intro

[David Mason via talk]

Rust uses a reference-counting collector for allocations that go beyond what 
the borrow-checker can handle. You have to explicitly use the RC allocation. 
Reference counting, as you may know is a more predictable memory allocation 
technique and works well for many data structures, such as trees (binary or 
otherwise). It however has problems with cyclic data structures (doubly-linked 
lists, general graphs, etc.). Since the reference-counted allocations are 
explicit, it is usually not onerous for the programmer to handle the 
de-allocation of these data structures. Having a built-in RC collector is a big 
win over C/C++ - your effective alternatives.

If you aren’t willing to deal with the reference counter, your closest choices 
to Rust are D, Nim, or (less close) Go. But if you look at the benchmark game 
site, you’ll see that garbage-collected languages are often a factor of 3 
slower than Rust/C/C++. If that degree of performance matters to you, I think 
you should use Rust rather than C or C++. If it doesn’t, you have a plethora of 
choices, including Go, Java, C#, Haskell, Python, Lisp, or (my favoured) 
Smalltalk.

../Dave
On Jan 1, 2020, 6:09 PM -0500, Nicholas Krause , wrote:
>
>
> On 1/1/20 11:44 AM, David Mason wrote:
> > Borrowing is entirely a compile-time analysis. There is no runtime impact 
> > (other than the fact that you can get away without a garbage collector - in 
> > a safe way).
> >
> > The Learn Rust the Dangerous way article is very good, by the way! I 
> > heartily endorse it for the C-philes among GTALUG. If you haven’t read it, 
> > one of the things that might convince you is that the leaderboard for this 
> > highly-optimized n-body simulation has Rust in the first-place 
> > https://benchmarksgame-team.pages.debian.net/benchmarksgame/performance/nbody.html
> >  - faster than C, C++, Fortran or Ada. I’ve added it to my list of 
> > resources for Rust: https://cps506.scs.ryerson.ca/Resources/rust.html
> Ownership makes sense its a compiler version of smart pointers. My concerns 
> are still:
>  What about circular references in which the owner depends on data from the 
> child
> but cannot free it due to the knowledge also depending on the parent. Binary 
> trees
> are a problem here. Or you must assume like garbage collectors this  never 
> occurs
> and this is one way to get memory leaks in a lot of garbage collectors fast.
>
> Rust seems fine for a lot of things but this one case does not seem solved at 
> least
> in my knowledge or is  assumed to not be a big issue and I could be wrong but
> from my limited research it appears not,
> Nick
> >
> > ../Dave
> > On Dec 31, 2019, 4:22 PM -0500, Nicholas Krause via talk , 
> > wrote:
> > >
> > >
> > > On 12/31/19 11:57 AM, D. Hugh Redelmeier via talk wrote:
> > > > | From: Tom Low-Shang via talk 
> > > >
> > > > | I'm interested in your thoughts on Rust if you attended the talk.
> > > >
> > > > The talk was mostly a guided creation of a program. So I don't think
> > > > that it answered any of your questions.
> > > >
> > > > | I'm currently learning Rust the old fashioned hacker way (from books 
> > > > and
> > > > | other people's code :)). My biggest mistake was trying to use Rust 
> > > > with
> > > > | SDL2 to display some graphics. My head still hurts from banging it 
> > > > into
> > > > | a wall called 'lifetimes'. :)
> > > >
> > > > The whole idea of borrowing etc. is fundamental to Rust and how it
> > > > ensures safety. Without garbage collection. If you don't like or
> > > > understand this approach, Rust isn't useful.
> > > Hugh,
> > > I've a question about how borrowing is implemented internally as it can 
> > > lead
> > > to a problem, if I allow lots of memory can my program stall because of 
> > > this
> > > at the end of a block. In addition due to this does borrow checking
> > > limit or
> > > not implement something like freelists or caching to get better usage of 
> > > the
> > > CPU cache as that's also a concern.
> > >
> > > Thanks,
> > > Nick
> > > > ---
> > > > Post to this mailing list talk@gtalug.org
> > > > Unsubscribe from this mailing list 
> > > > https://gtalug.org/mailman/listinfo/talk
> > >
> > > ---
> > > Post to this mailing list talk@gtalug.org
> > > Unsubscribe from this mailing list 
> > > https://gtalug.org/mailman/listinfo/talk
>
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Rust intro

[David Mason via talk]

Borrowing is entirely a compile-time analysis. There is no runtime impact 
(other than the fact that you can get away without a garbage collector - in a 
safe way).

The Learn Rust the Dangerous way article is very good, by the way! I heartily 
endorse it for the C-philes among GTALUG. If you haven’t read it, one of the 
things that might convince you is that the leaderboard for this 
highly-optimized n-body simulation has Rust in the first-place 
https://benchmarksgame-team.pages.debian.net/benchmarksgame/performance/nbody.html
 - faster than C, C++, Fortran or Ada. I’ve added it to my list of resources 
for Rust: https://cps506.scs.ryerson.ca/Resources/rust.html

../Dave
On Dec 31, 2019, 4:22 PM -0500, Nicholas Krause via talk , 
wrote:
>
>
> On 12/31/19 11:57 AM, D. Hugh Redelmeier via talk wrote:
> > | From: Tom Low-Shang via talk 
> >
> > | I'm interested in your thoughts on Rust if you attended the talk.
> >
> > The talk was mostly a guided creation of a program. So I don't think
> > that it answered any of your questions.
> >
> > | I'm currently learning Rust the old fashioned hacker way (from books and
> > | other people's code :)). My biggest mistake was trying to use Rust with
> > | SDL2 to display some graphics. My head still hurts from banging it into
> > | a wall called 'lifetimes'. :)
> >
> > The whole idea of borrowing etc. is fundamental to Rust and how it
> > ensures safety. Without garbage collection. If you don't like or
> > understand this approach, Rust isn't useful.
> Hugh,
> I've a question about how borrowing is implemented internally as it can lead
> to a problem, if I allow lots of memory can my program stall because of this
> at the end of a block. In addition due to this does borrow checking
> limit or
> not implement something like freelists or caching to get better usage of the
> CPU cache as that's also a concern.
>
> Thanks,
> Nick
> > ---
> > Post to this mailing list talk@gtalug.org
> > Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
>
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] legal history of university AUPs, was Re: For Chris: Commodore BASIC as a scripting language

[David Mason via talk]

Moral rights are most frequently applied to things like reports that people are 
contracted to write (so that the contracting agency can’t change the sense of 
the report in a press release, for example.

Moral rights cannot be sold/assigned, but can be waived. People hired to write 
reports are sometimes required to waive their moral rights as part of the 
contract to hire them. Which makes one absolutely an expert for hire….

I was unaware of the limited protection of moral rights in the US, but 
according to 
https://library.osu.edu/site/copyright/2017/07/21/moral-rights-in-the-united-states/
 it only applies to visual works - not other forms of copyrighted work (such as 
articles, reports, or code).

../Dave
On Aug 27, 2019, 4:06 PM +0200, D. Hugh Redelmeier via talk , 
wrote:
> | From: D. Joe via talk 
>
> An excellent and informative post.
>
> | Amongst those of us of a certain age in the US, who don't have legal
> | training, I suspect the lore persists that copyright restrictions apply
> | only if the copyright for the work has been registered--this used to be
> | true in living memory.
>
> Are you in the US?
>
> My understanding of US law (shaky!!) is that damages are limited
> if you didn't register the copyright.
>
> The other thing that USAnian (and Canadian) people don't understand is
> the "moral rights" component of copyright. The US copyright law got
> this concept by way of Berne.
>
> I had heard that US residents don't have moral rights under copyright law
> in the US but foreigners do. I don't think that these have been
> asserted often.
>
> Even in Canada, moral rights are pretty rarely invoked. I only
> remember one example: Michael Show vs Eaton Centre vis a vis ribons on
> the Canada Geese sculptures. (Snow won).
>
> I imagine that moral rights could be useful in the open source world.
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Script to show HTTP(S) and TLS details for a website

[David Mason via talk]

Line 130 s/-eq/=/

Otherwise, cool! Thanks!

../Dave
On Aug 10, 2019, 5:32 PM -0400, Jason Shaw via talk , wrote:
> Looks like a handy script to have.  My only real suggestion is to change the 
> shebang to
> #!/usr/bin/env bash
> as it's more portable than the current one.  Still not perfect, but works 
> reliably on more systems.
>
> -jason
>
> > On Sat, Aug 10, 2019 at 11:46 AM Giles Orr via talk  wrote:
> > > This may be seen as self-promotion - that's not totally wrong.  But I 
> > > think this may also be useful to others and (as I acknowledge in the blog 
> > > post) I'm quite pleased with the resultant script.
> > >
> > > Over the past year and a half I've slowly developed a shell script that 
> > > gives a concise summary of the state of TLS and HTTP(S) on a given 
> > > website.  It looks like this:
> > >
> > >     $ tlsdetails google.ca
> > >     Using OpenSSL:  /usr/bin/openssl
> > >     Expiry Date:    Oct 27 17:27:07 2019 GMT (78 days)
> > >     Issuer:         Google Trust Services, CN
> > >     TLS Versions:   tls1_3 tls1_2 tls1_1 tls1  (tried but unavailable: 
> > > ssl3 ssl2 )
> > >     HTTP Version:   2
> > >
> > > I first started work on it after a couple embarrassing certificate 
> > > expiries.  It then grew to check the Issuer, TLS versions, and more 
> > > recently whether or not a site supports HTTP2.
> > >
> > > (The pointer to the OpenSSL version is shown because the script will also 
> > > run on Mac, and their version of 'openssl' is problematic at best.  That 
> > > line is of course easy to remove if you don't like it.)
> > >
> > > If you're interested, you can find the details here:
> > >
> > > https://www.gilesorr.com/blog/tls-https-details.html
> > >
> > > Any suggestions to improve the script would be most welcome.
> > >
> > > --
> > > Giles
> > > https://www.gilesorr.com/
> > > giles...@gmail.com
> > > ---
> > > Post to this mailing list talk@gtalug.org
> > > Unsubscribe from this mailing list 
> > > https://gtalug.org/mailman/listinfo/talk
> ---
> Post to this mailing list talk@gtalug.org
> Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk
---
Post to this mailing list talk@gtalug.org
Unsubscribe from this mailing list https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] A find alternative: fselect

[David Mason via talk]

Good point. You can turn the checks on, but see this discussion about the 
default: https://github.com/rust-lang/rust/issues/47739

../Dave
On Jun 20, 2019, 12:26 AM -0400, D. Hugh Redelmeier via talk , 
wrote:
> | From: D. Hugh Redelmeier via talk 
>
> | Rust does a few things that are relevant (I think -- I've not actually
> | used Rust):
>
> | - unfortunately, I think that Rust only catches integer overflow in
> | debug mode. That's a mistake, but it's probably because checking is
> | considered too expensive.
> | 
> 
>
> A very recent CVE against the Linux kernel exploits integer overflow
> CVE-2019-11477: SACK Panic (Linux >= 2.6.29)
> 
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] A find alternative: fselect

[David Mason via talk]

On Jun 14, 2019, 1:03 AM -0400, Dhaval Giani , wrote:
> > HOWEVER, since Rust code is intrinsically much, much safer than C code, 
> > stability of API is much more legitimate a characterizer of the version 
> > that you want than bug-fixes (and bug-fixes are almost never 
> > security/safety related).
>
> Please please please. Bug fixes, specifically the ones that get
> shipped fairly quickly, are almost always security related. Security
> bugs are just a class of bugs.

I was unclear. I completely agree that in the C/Assembly world, bug-fixes are 
almost always security related. In the Rust world, security bugs are very hard 
to create (not impossible, but hard), so most *Rust* bug-fixes are bugs for 
sure, but not security bugs.

> Also, till I see some way of formally verifying, that rust code is
> safe enough that security issues are not possible, remind me to be
> quite sceptical of the claim that Rust code is secure. In the last two
> years, we have seen classes of bugs believed to be impossible.

Even formal proofs of software correctness would not have caught some of those 
bugs, and the complexity of the hardware and software involved in those bugs 
arguably make formal proof impossible. So we all work in the realm of 
probabilities. I’d just like to improve my odds.

> I will accept your claim that in the hands of an average programmer,
> Rust is probably safer, but we have enough experience with C/Assembly,
> that I am going to claim that with someone experienced, they can
> create as secure/safe programs in C, as experienced folks in Rust.

With 45+ years of programming, dozens of programming languages and several 
hundred thousand lines of code under my belt, I don’t think I’m an average 
programmer (but then nobody thinks they’re average - the Lake Wobegon effect 
https://8hours.org/2012/11/do-you-think-you-re-better-than-average/ - in 
one study 93% of American drivers thought they were better than average). I 
have rarely written code that has security bugs in it, but I very much 
appreciate the fact that the Rust compiler has my back.

What I really appreciate with Rust is that I get better zero-cost abstractions 
than I would with C++, but with vastly greater safety (and of course, C has 
essentially no useful abstractions except functions). Those abstractions mean I 
have a deeper understanding of my code and therefore am less likely to create 
bugs, and I’m more productive. And even with the abstractions, Rust produces as 
good code (within fuzz) as C. 
https://www.apriorit.com/dev-blog/520-rust-vs-c-comparison has a good 
comparison.

So why would an experienced programmer use C instead of a language that has so 
much more? Efficiency? a non-argument. Machismo/bragging-rights? go for it. 
Laziness? understandable, but very short-term. Access to the lowest level of 
code? Rust does that with “unsafe” code sections that are identifiable with as 
low-level tools as grep. As the second link points out, there are tools that 
can do *some* of Rust’s safety analysis for C++ (none for C), but why wouldn’t 
you want that part of every compile?

I’m not sure this thread is worth my contributing to, as there are so many 
people that know better, but I am doing so because both as a 
programming-language professor and researcher, and as a consumer of technology, 
I would like it if more programmers used better and safer tools, going forward! 
Go-lang is cool, but its abstractions aren’t all zero-cost. So for me, Rust and 
Smalltalk are the sweet-spots.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] A find alternative: fselect

[David Mason via talk]

My preface: 40 years of C programming is enough! I don’t plan to write any more 
C; where I need that low-level efficiency, I’ll use Rust. Ditto no C++; if I 
want objects, I’ll do it right with Smalltalk.

With that said, Hugh makes a legitimate point (if we were talking apples and 
apples).
> On Thu, Jun 13, 2019 at 1:12 AM (actually, earlier) D. Hugh Redelmeier wrote:
 |   Rust culture and practice doesn’t seem to like shared libraries.. I do.

Rust does tend to create static code from crates (its packaging system). But C 
does too, for source code that you compile into your code. The difference is 
that, because crates are understood by the compiler it is *much* easier to make 
decisions about what version you want in Rust.

Rust also *does* use dynamic libraries for C/system code. I have a web server 
that uses sqlite3 and the generated executable includes the sqlite3, resolve, 
and system dynamic libraries.

So the static part is the crates, and it is trivial to discard the previously 
frozen versions and compile with the latest versions.

HOWEVER, since Rust code is intrinsically much, much safer than C code, 
stability of API is much more legitimate a characterizer of the version that 
you want than bug-fixes (and bug-fixes are almost never security/safety 
related). Rust uses semantic versioning of crates, so you can specify, e.g. 
version 2.3.* of a crate, to get any bug-fixes without any breaking changes.

All in all, I think Rust is making the right choice here. (And C is making the 
right choice for C code, too - because it’s so unsafe!)

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] war story: fixing an LCD TV

[David Mason via talk]

I have used a projector for “TV” (Netflix/Youtube) watching for the last 5 
years. It has 2 HDMI ports, and is quite nice. It’s “only” 1080p, but when it’s 
a 90” screen, that’s pretty good. I had a BenQ, but the bulb was always a 
problem, so I recently replaced it with a XGIMI, and it’s awesome!

I occasionally plug the HDMI into my laptop, and that works well too. It claims 
to be 4K (speaking of brands) and it will take a 4096x2160 signal, but the 
underlying display is 1920x1080.

I also have a 39” “RCA” that I bought at Loblaws 5 years ago for $199 (I think. 
it *might* have been as much as $299, but I don’t think so)… The colours aren’t 
perfect (mostly meaning that the range is limited), but it’s our bedroom 
screen, and it works pretty well as a monitor for my laptop, too.

../Dave
On May 31, 2019, 1:50 PM -0400, James Knott via talk , wrote:
> On 2019-05-31 01:45 PM, Evan Leibovitch via talk wrote:
> > How viable Is it to use a TV as monitor, ignore the "smart" crap and
> > just plug in the HDMI? Are there features or specs needed for a TV to
> > make it usable for close viewing?
> >
>
> I have used my TVs on occasion as a monitor and worked fine.  On the
> other hand, my monitor is connected to a Rogers box for watching TV.  As
> for close viewing, that would depend entirely on the TV.  Some are
> better than others.  For example, my new Sharp TV seems to have a
> Sharper (sorry ) image than the old Sharp TV.
>
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] of routers and access points

[David Mason via talk]

I have ordered a Ubiquiti EdgeRouter X-SFP and long-range, 802.11ac, PoE, 
access point. Since the EdgeRouter is WRT compatible, I can install that and 
get DNS running there too, which means that I can power off the old Linux 
gateway machine. I could have continued using my existing router as an access 
point, but the new one is longer range, higher speed, and easier to wire.

I looked at the Amplifi and it looks quite nice, comparatively priced with the 
option I went with, and very plug-and-play (with iOS/Android controls), but 
running WRT and DNS on the router was too good an option for me to pass up.

Looks like the EdgeRouter can handle nearly a gigabit WAN throughput with some 
“buffer bloat” (essentially a bit of queuing delay) or about 200Mbit with SQM 
and very minimal delay, 550Mbit without SQM but still minimal bloat. 
https://forum.openwrt.org/t/ubiquiti-edgerouter-x-loading-openwrt-and-performance-numbers/27470/6
 All those numbers are way beyond the bandwidth I’m currently buying, so I’ll 
probably use SQM 
https://openwrt.org/docs/guide-user/network/traffic-shaping/sqm I just ran 
http://www.dslreports.com/speedtest/49045641 and you can see that the delays 
are highly variable and generally suck! I’ll post revised scan after I’ve 
installed the new router.

Thanks for all of your advice.


../Dave
On Apr 26, 2019, 12:52 PM -0400, Jamon Camisso via talk , 
wrote:
> On 4/25/19 4:00 PM, David Mason via talk wrote:
> > Is there an openwrt non-wifi, PoE router? Also the WRT1900ACv2 seems to
> > have limited availability.
>
> Looks like the ERX you were looking at is compatible, TIL!
>
> https://openwrt.org/toh/ubiquiti/ubiquiti_edgerouter_x_er-x_ka
>
> I might try flashing mine sometime.
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] of routers and access points

[David Mason via talk]

Is there an openwrt non-wifi, PoE router? Also the WRT1900ACv2 seems to have 
limited availability.

../Dave
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] of routers and access points

[David Mason via talk]

This (letting the router do all the work) is a possibility, but the router is 
very far from all the other connections, so at the moment I have a long RJ45 
cable to the router. So I’d have to run another long cable back to the switch 
where everything else is plugged in.

Part of my question was whether anyone else is using POE access points, and 
whether it’s worth doing as right now I have an extension cord running to the 
router. If so I’d need a router that supported POE, so I mentioned the Ubiquiti 
EdgeRouter as such a (non-WiFi) router.

To clarify the topology, the modem is just a modem (cable), so some router is 
needed. Also, I don’t think the DLink WiFi router does DNS other than relay 
(because I want to have names for my local servers). It also flakes out for a 
few seconds, so replacing it now seems like a reasonable thing.

Lennart mentions the WRT1900ACv2, but I’d like to get the system so that it’s 
less brittle (i.e. if I’m not home for a week or 2 and bad things happen, i’d 
like others to be able to get things back to functioning) so I’m not sure about 
custom firmware.

../Dave
On Apr 24, 2019, 3:11 PM -0400, Don Tai , wrote:
> Why don't you let the router do all the router work, such as Wifi, routing, 
> DHCP, DNS, etc? I have my router RJ45'd to my Bell router, and let it do all 
> that. The Bell router's wifi is rarely used, but is a backup. Connect all 
> your boxes RJ45 to your router if you wish, then they can all play well 
> together, no? Is this too simple a setup?
>
> > On Wed, 24 Apr 2019 at 11:58, David Mason via talk  wrote:
> > > I had a setup where my modem was connected to my gateway Linux box which 
> > > did routing, DHCP, DNS, etc. and was also connected to a switch which was 
> > > connected to a WiFi router being used as an access point. This is too 
> > > brittle for my tastes, as other members of my family can’t resolve 
> > > problems with it.
> > >
> > > Something went wrong so I have temporarily reconfigured it so the WiFi 
> > > router is connected directly to the modem, but this doesn’t let the WiFi 
> > > clients connect to the Linux boxes or other hard-wired machines or let 
> > > them connect to the internet. (I could connect the other devices to the 
> > > router, but for WiFi propagation reasons, the WiFi router is far away 
> > > from all the other devices, so this is not ideal.)
> > >
> > > So I was thinking of getting a 
> > > https://www.canadacomputers.com/product_info.php?cPath=27_1046_363_id=102961
> > >  router and connecting to that the Linux box (which would still do DHCP 
> > > (possibly) and DNS, the router configured back as an access point (or a 
> > > POE access point such as 
> > > https://www.canadacomputers.com/product_info.php?cPath=27_1056_356_id=031082,
> > >  as well as the other wired devices.
> > >
> > > Thoughts? Thanks.
> > >
> > > ../Dave
> > > ---
> > > Talk Mailing List
> > > talk@gtalug.org
> > > https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

[GTALUG] of routers and access points

[David Mason via talk]

I had a setup where my modem was connected to my gateway Linux box which did 
routing, DHCP, DNS, etc. and was also connected to a switch which was connected 
to a WiFi router being used as an access point. This is too brittle for my 
tastes, as other members of my family can’t resolve problems with it.

Something went wrong so I have temporarily reconfigured it so the WiFi router 
is connected directly to the modem, but this doesn’t let the WiFi clients 
connect to the Linux boxes or other hard-wired machines or let them connect to 
the internet. (I could connect the other devices to the router, but for WiFi 
propagation reasons, the WiFi router is far away from all the other devices, so 
this is not ideal.)

So I was thinking of getting a 
https://www.canadacomputers.com/product_info.php?cPath=27_1046_363_id=102961
 router and connecting to that the Linux box (which would still do DHCP 
(possibly) and DNS, the router configured back as an access point (or a POE 
access point such as 
https://www.canadacomputers.com/product_info.php?cPath=27_1056_356_id=031082,
 as well as the other wired devices.

Thoughts? Thanks.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Help with choosing a WiFi router

[David Mason via talk]

Sorry, I mis-spoke. I actually am just looking for an AP.

I have a (Debian) gateway machine that is connected to the wireless, the 
Internet, and the file server via 3 separate wired ethernets. I don’t like the 
idea of commercial software connected to the Internet.

I was looking at TP-Link Archer C7 AC1750 Dual Band Wireless AC Gigabit Router 
https://www.amazon.ca/dp/B00BUSDVBQ?ref_=ams_ad_dp_ttl for $80 (I have Prime).

Are there just APs that are better price/performance?

Thanks again

../Dave
On Mar 31, 2019, 8:56 AM -0400, James Knott via talk , wrote:
> On 03/30/2019 11:10 PM, D. Hugh Redelmeier via talk wrote:
> > | From: David Mason via talk 
> >
> > | We’ve had a DLINK DIR-628 for quite a while.
> >
> > There's a new standard since then: 802.11ac. Confusing, because there
> > already were 802.11a and 802.11c.
>
> That's why 802.11ax is called WiFi 6.  They've assigned other numbers,
> such as 5 to 802.11ac.
> https://en.wikipedia.org/wiki/IEEE_802.11ax
>
> However, it's getting to the point where the improvements don't benefit
> individual users much.  For example, MIMO improves performance for
> multiple users more than individuals.  The main benefit for individuals
> would be somewhat greater distance, due to beam forming or focusing the
> signal in one direction.
>
> >
> > There's been a race to be able to claim speed in wireless routers.
> > They use multiple antennae to allow multiple devices to be supported
> > at one time (beamforming?).
> >
> > <https://en.wikipedia.org/wiki/IEEE_802.11ac>
> >
> > These advantages depend on the clients supporting 802.11ac.
>
> 802.11ac is becoming common these days.  My Pixel 2 phone, which I
> bought over a year ago supports it.
> 802,11ax supports MIMO in both directions, whereas 802.11ac only on
> transmit.
>
> >
> > Enthusiast routers have gotten more expensive and look more like
> > aggressive alien space ships. Just look at this one:
> >
> > <https://www.newegg.ca/Product/Product.aspx?Item=9SIADGE5R87033>
>
> They are also capable of much more.  Some support mesh, where multiple
> APs can work together to provide seamless coverage over a large area.
> >
> > I like the idea of OpenWRT but don't actually use it. Even though
> > I have two wireless routers, I only use them as APs -- the routing
> > functions are not used. So my advice about consumer wireless routers
> > is pretty theoretical. For gateways (including the routing function),
> > I use little PCs running CentOS or Fedora.
>
> I prefer separate APs and routers.  Many APs support power over Ethernet
> (PoE), so you don't have to worry about having power near where the AP
> works best.  Mine's at the top of the wall in my laundry room, which
> puts it roughly in the middle of my condo.
>
>
> O'Reilly has some good WiFi books, by Michael Gast.  Recommended reading
> if you really want to learn about WiFi.
>
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

[GTALUG] Help with choosing a WiFi router

[David Mason via talk]

We’ve had a DLINK DIR-628 for quite a while. When it works it has been great, 
but it drops out several times a day. Sometimes it come back fairly seamlessly, 
but sometimes you have to turn off the wifi on connected devices and turn it 
back on again. Power-cycling it always resolves the problem, but it’s annoying.

So I guess I’m in the market for a new WiFi router. A little bit of additional 
range would be great, and if there’s more speed available, so much the better, 
but as I said, when it’s working, it’s been great. It’s not uncommon for 3 
video streams being viewed while I’m ssh’ed into remote systems, or downloading 
files, or cloning repositories, and it all works fine.

Price isn’t a major concern… I’d happily pay for reliability and performance. 
To the degree it matters, we’re mostly connected from Apple devices (except for 
the gateway and RAID servers - both of which are Debian - which are wired 
together).

Thanks for any advice!

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] “Safe” Jobs (was Boeing India software engineers)

[David Mason via talk]

This whole thread is redundant. AIs (or more accurately Machine-Learning 
systems) are already eating into many aspects of Law, Tech support, etc. They 
will only expand… and will take over many of the low-value jobs domestic, or 
outsourced.

And, of course, come the Singularity, nothing is safe.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Ontario Bill 72: "Right to Repair"

[David Mason via talk]

No, he dropped out of Reed college - https://en.wikipedia.org/wiki/Steve_Jobs

But he was certainly not an engineer (to his credit).

../Dave
On Mar 7, 2019, 5:15 PM -0500, James Knott via talk , wrote:
> On 03/07/2019 01:20 PM, Gary via talk wrote:
> > Incidentally, some MBA's such as Steve Jobs, do provide real value,
> > instead of relying on marketing hype, but sadly, in this age, many do not.
>
> Jobs was a MBA?
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

[GTALUG] WiFi (un)happiness

[David Mason via talk]

---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Practical Use of GRUB's DSL: With the examples inexplicably left out of the GRUB documentation

[David Mason via talk]

I use LaTeX for everything - most particularly for papers and lecture slides. I 
haven’t used it in a while, but HeVeA (http://hevea.inria.fr/) is the best way 
to convert LaTeX to HTML.  When I last used it, it was quite effective. And it 
is quite focussed on including code in documents.

../Dave
On Oct 16, 2018, 9:16 AM -0400, Giles Orr via talk , wrote:
> I've finally released a document I've been working on for a while:
>
> "Practical Use of GRUB's DSL: With the examples inexplicably left out of the 
> GRUB documentation" ( https://www.gilesorr.com/grubdsl/ )
>
> Daniel did the editing, and set me on this path in the first place with his 
> own interest in GRUB.  The GRUB DSL is somewhat similar to an older version 
> of a Linux shell, but documentation of its functionality online is poor to 
> non-existent.  Don't get me wrong: GRUB has a lot of documentation of 
> individual commands, but how you can assemble them into useful scripts is 
> barely documented at all.  (Yup, its an uncommon application realm.)  My 
> greatest frustration with GRUB's DSL is the lack of redirection and pipes (I 
> understand why they're not there - but it would be nice to have them).
>
> I hope this is helpful to someone!
>
>  -
>
> As a technical side note, I learned the basics of LaTeX in an attempt to 
> create this document in that language (I'd been wanting to learn it for 
> years).  LaTeX is supposed to be able to generate not only PostScript and 
> PDF, but also HTML.  But it turns out the HTML generator isn't nearly as well 
> maintained as the PDF generator, and in the end I could find no way to 
> implement my code examples in such a way that the HTML generator wouldn't 
> fail on them.  I also found the language unnecessarily complex for what I was 
> trying to achieve and ultimately switched to raw HTML (which you see above).  
> LaTeX may have been a poor choice.  :-)
>
> --
> Giles
> https://www.gilesorr.com/
> giles...@gmail.com
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] programming system questions

[David Mason via talk]

Slack is far, far, from a simple text chat, regardless of what you (or I) think 
it should be. I hadn’t checked its size. On my system it’s about 200MB+another 
200MB per workspace. Yeah, that’s outrageous. But Firefox is using about 10GB, 
so I’m not exactly convinced that a web app is going to have a significantly 
smaller footprint than Electron.

The fact that Slack is such a high-quality app is in my mind a huge vote in 
favour of Electron (as is Discord)

../Dave
On Oct 11, 2018, 10:53 AM -0400, Jamon Camisso via talk , 
wrote:
> On 11/10/18 10:37, David Mason via talk wrote:
> > Electron is probably the easiest way to move a web app to an
> > application. See:
> >     https://electronjs.org/
> >
> > It uses html, javascript, css and should be able to connect to your
> > postgresql database.
>
> slack, what should be a simple text chat built on electron is taking
> 873044 kb RSS memory on my desktop.
>
> electron should be the last resort - it is probably the worst thing you
> could do for your users if you care about resource constraints.
>
> I would question the merit of even moving from a web application. I am
> on 400/100kbit DSL in the country, and I don't have any issues with even
> large react or ember based web apps.
>
> Cheers, Jamon
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] programming system questions

[David Mason via talk]

Electron is probably the easiest way to move a web app to an application. See:
    https://electronjs.org/

It uses html, javascript, css and should be able to connect to your postgresql 
database.

../Dave
On Oct 11, 2018, 8:53 AM -0400, o1bigtenor via talk , wrote:
> Greetings
>
> I have been using a web application for one of my business functions.
> Besides the limitation of it being a web application (the web is NOT
> as consistent nor as quick as is considered 'normal' for those that
> like in rural areas) there are to many areas where I need to tweak the
> results. Therefore I would like to re-create this application.
>
> At its base its somewhat like a recipe system. I plan on using
> postgresql for data holding and now what might be a good way of now
> doing the work.
>
> Think that an item has from 5 to 50 of 50 information fields. I am
> adding a number of items together and specifying quantities of each
> item trying to achieve levels that I previously specified.
>
> The present application uses css and javascript and some other things
> that are on the 'home' (program owner's) server.
>
> Hopefully this description is clear enough and gives enough
> information so that suggestions might be forthcoming. (I asked an
> engineer friend and his suggestion was to use a monte carlo type of
> system but that would be several levels more complex that I think
> might be necessary for this application.)
>
> Regards
>
> Dee
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Swappiness

[David Mason via talk]

See https://www.howtoforge.com/tutorial/linux-swappiness/

Setting it to a much lower number (like 20) sounds like it would work better if 
your ram usage is fairly stable. Sound like, “try it and see” is called for (an 
painless).

../Dave
On Sep 5, 2018, 6:52 AM -0400, o1bigtenor via talk , wrote:
> On Tue, Sep 4, 2018 at 11:16 PM, Evan Leibovitch via talk
>  wrote:
> > On Tue, Sep 4, 2018, 10:40 PM John Weintraub, 
> > wrote:
> > >
> > > Is your swap equal to twice the amount of RAM? That's the default; I am
> > > not sure where you came up with the number "60";
> >
> >
> > I was referring to this:
> >
> > https://en.m.wikipedia.org/wiki/Swappiness
> >
> Greetings
>
> Article looks interesting except the systemd provisions are attributes and and
> are not mentioned. Bottom of the article indicates a somewhat recent editing
> date but with no mention of systemd I have further questions. The article
> mentions places to change 'swappiness' and I can find the actual file but 
> there
> is nothing in the sysctl.conf (IIRC the name) file re: swappiness. Somehow I
> would prefer something that did refer to systemd foibles as well for a guide 
> so
> as not to really pooch anything.
>
> I, too, am interested largely because I have lots of ram and wouldn't mind
> speed enhancements.
>
> Regards
>
> Dee
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Which Distro is Best for Running a ZFS-on-Linux Fileserver.

[David Mason via talk]

OK, so I have an 8TB Seagate USB disk and have created a zpool on it called
backup1. My main pool is called tank. I tried:

: ~ ; sudo zfs snapshot -r tank@2018-08-31

: ~ ; sudo zfs list

NAMEUSED  AVAIL  REFER  MOUNTPOINT

backup1 508K  7.14T   136K  /backup1

tank   1.66T   916G   412K  /tank

tank/audio 12.1G   916G  12.1G  /audio

tank/cvs   32.7M   916G  32.7M  /tank/cvs

tank/etc   18.1M   916G  18.1M  /tank/etc

tank/home   531G   916G   531G  /home

: ~ ; sudo zfs list -t snapshot

NAMEUSED  AVAIL  REFER  MOUNTPOINT

tank@2018-08-310  -   412K  -

tank/audio@2018-08-31  0  -  12.1G  -

tank/cvs@2018-08-310  -  32.7M  -

tank/etc@2018-08-310  -  18.1M  -

tank/home@2018-08-31   0  -   531G  -
and now I try (after some research):

: ~ ; sudo zfs send -R tank@2018-08-31 | sudo zfs recv -vd backup1

cannot receive new filesystem stream: destination 'backup1' exists

must specify -F to overwrite it

warning: cannot send 'tank@2018-08-31': Broken pipe

Any quick help?

Thanks  ../Dave

On Tue, 28 Aug 2018 at 17:33, Scott Sullivan via talk 
wrote:

> On 2018-08-27 09:24 AM, Giles Orr via talk wrote:
> > On Sat, 25 Aug 2018 at 14:21, David Mason via talk  > <mailto:talk@gtalug.org>> wrote:
> > This system is <5 years old, and at the time was kind-of leading
> > edge. so I’m not worried about that.
> > It’s a 4.4Tb raidz2 at 64% and has performed flawlessly.
> > Unfortunately I don’t really have the time to do any serious digging
> > right now, either.
> >
> > How do others backup their ZFS systems? Getting a 4T external drive
> > doesn’t seem like the best plan, but maybe there isn’t any other
> choice.
>
> In my case I built a secondary NAS and disk array, and do regular 'zfs
> snapshots' and 'zfs sends'. In recent history I've started using
> zfs-snap-manager to automate that.
>
> https://github.com/khenderick/zfs-snap-manager
>
> It's a rather coarse tool... doesn't support automate snapshots more
> frequent then once a day, but will happily send over any you've made
> manually (via a cron job or alternative method).
>
> Currently the developer has only packaged it for Arch. But I've built an
> rpm spec file for it. Attached.
>
> > Actually, that sounds like a really good plan.  In fact, buy two so you
> > can do rotating backups.  Think about your alternatives - about the only
> > one that occurs to me is a tape drive.  There used to be consumer-grade
> > tape backups, but they don't exist anymore and I'd argue this is no
> > longer a viable solution outside the data centre.
> >
> > Buying external hard drives is a really good idea: they're dirt cheap
> > (at least compared to the alternative - failure of your primary).
>
> I agree with Giles. If you don't want to drop the coin on a second NAS,
> this is a very usable strategy. Get a 6 or even 8TB disk, format it as a
> ZFS pool and turn on zfs's block compression, and set copies to '2'.
>
> zfs set compression=lz4 
> zfs set copies=2 
>
> Setting a number of copies, is normally not useful for a multi-disk
> array, as the copies can end up on the same disk. But on a single disk,
> they are an insurance policy against bad sectors.
>
> Then you just zfs send your snapshots to it. I regularly use this as a
> local backup strategy with my work laptops.
>
> --
> Scott Sullivan
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Which Distro is Best for Running a ZFS-on-Linux Fileserver.

[David Mason via talk]

If you’re building 4TB, I’t recommend raidz2 which doesn’t give you much 
storage (4 x 1TB disks would give you <2TB of user storage). I have 5 x 1TB 
raidz2 which gives me 2.82TB of user storage. This may have changed, but when I 
went to ZFS you couldn’t add more disks and change the structure, you could 
only enlarge the disks - so if you went for a 5-disk raidz2, you were stuck 
with a 5-disk raidz2. I really hope this has changed, but I haven’t researched 
it.

The reason to go with raidz2 as the pool gets larger is that with raidz if a 
disk dies, while it rebuilds on the replacement disk, you have no redundancy, 
so any error on one of the other disks will be replicated.

So I’d recommend you get an additional SATA card, and then you don’t need to do 
ZFS on /, and you could add another drive to get better cost-effectiveness from 
your ZFS. I have / on a moderate-sized SSD and 5 hard drives in a ZFS pool.

One reason to use not-whole-disks for ZFS is that you can migrate to larger 
disks as time goes on.  I started out with 500GB (if I remember correctly) and 
was able to move to 1TB and copy the partitions.

Scott knows much more about this than I, but it may (now) be that (with a 
raidz2) you can enlarge by simply removing a xGB disk and replacing it with a 
2xGB drive, let it rebuild, then replace the next one. The problem with that is 
you are replacing the disks, so it’s not very incremental. So if you can 
enlarge by adding an additional drive, that is substantially better.

../Dave
On Aug 30, 2018, 9:34 AM -0400, Scott Sullivan via talk , 
wrote:
> On 2018-08-29 11:43 PM, Amos H. Weatherill wrote:
> > Scott,
> >
> > My reasoning for / on ZFS is pretty Simple ... the machine that is
> > becoming my first NAS only has 4 SATA Ports, so I can't afford to Waste
> > one on a boot drive.
>
> Recommended best Practice is to use ZFS with whole disks. That said,
> most of the arguments for that are 'because the manual says so',
> 'because zfs datasets are far more flexible then partitions' and
> references to Solairs taking advantage of disk caches. I throw that all
> out the windows in favor of doing at rest encryption, with whole luks
> partitions(*).
>
> My more practical argument is choice of MBR vs GUID partitioning. The
> latter is just cleaner (and the default when ZFS manages the disk), and
> works well with large disks (>2TB).
>
> But if your booting from that disk, you either need to be:
> "BIOS / CSM" + MBR + /boot
> or
> UEFI + GUID + "biosboot (partition)" + /boot
>
> Either of those makes for some lopsided partitioning, compared to the
> remainder of your data disks. A work around is to use a USB drive for
> your /boot. But in general your creating a more complex setup to
> maintain either way.
>
> Not knowing what hardware your using, if you have PCIe slots additional
> sata ports can be had for a low a $10/port.
>
> I've been using the Syba / IOCrest cards for a variety of needs,
> including ZFS arrays without issue.
>
> https://www.newegg.ca/Product/Product.aspx?Item=N82E16816124064
>
> > For Distro, I think I'll go with Fedora, as long as the / on ZFS guide
> > is sufficiently detailed.
>
> Fedora was not one of the ones I listed as having a guide to do rootfs
> on ZFS. If you found one, can you post the link?
>
> I'd also not recommend fedora in general for a NAS. CentOS would be a
> more dependable choice. LTS Ubuntu would be more reasonable as they
> ship(**) ZFS and support rootfs on it.
>
>
> ===
> * Native encryption in ZFS was added after the OpenZFS split from
> Sun/Oracle. So work to re-added it has been happening for a while. We're
> likely to see a stable version in the v0.8.x series.
>
> ** This is due to their adoption of a minority legal opinion about
> compatibility of the CDDL and GPL licenses that has not been tested in
> court.
> https://blog.ubuntu.com/2016/02/18/zfs-licensing-and-linux
> https://sfconservancy.org/blog/2016/feb/25/zfs-and-linux/
>
> --
> Scott Sullivan
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Which Distro is Best for Running a ZFS-on-Linux Fileserver.

[David Mason via talk]

I haven’t done a comparison lately, but used LVM for years, and liked it, but 
ZFS is so much more flexible: resize filesystems, snapshots, raidz2, sending to 
another ZFS system (for seamless backups and redundancy), auto-silvering, 
plug-and-play.

../Dave
On Aug 28, 2018, 9:18 PM -0400, Alvin Starr via talk , wrote:
> At the risk of poking the bear.
>
> Why use ZFS at all?
>
> ext4,XFS+LVM will do most all of the same things and from what I have
> read ZFS is slower than ext4,XFS and BTRFS.
>
>
> On 08/28/2018 05:18 PM, Scott Sullivan via talk wrote:
> > Having read through the thread to date, I'm actually a little
> > disappointed at the number of linux users pushing towards a Solaris or
> > BSD for ZFS.
> >
> > My primary File servers (4 of them) are all using ZFS for their data
> > partitions.
> >
> >
> > Amos,
> >
> > ## Couple of Answers to your questions
> >
> > A) Disto?
> >
> > I regularly run ZFS on CentOS and Fedora on a mix of SSDs and HDDs of
> > both the internal and external varieties. Fedora has some caveats,
> > only in that sometimes the kernel releases get ahead of what the ZFS
> > on linux team will support. And it's just a matter of waiting on a
> > working kernel zfs combination a week or two for them to catch up.
> >
> > But frankly, just pick your favorite distro and follow the relevant
> > getting started guide.
> >
> > https://zfsonlinux.org/
> >
> >
> > B)  Distro with ZFS root support (at install time)?
> >
> > No distro install supports this yet as I've seen. Although the do it
> > yourself ubuntu guide is lengthy, but very well detailed.
> >
> > https://github.com/zfsonlinux/zfs/wiki/Ubuntu-18.04-Root-on-ZFS
> >
> > Arch also support ZFS root, but their installation is all largely
> > manual to begin with.
> >
> > https://wiki.archlinux.org/index.php/Installing_Arch_Linux_on_ZFS
> >
> >
> > ## Couple of my own Questions
> >
> > 1) Why root (/) on ZFS, what is your use case / risk your trying to
> > mitigate?
> >
> >
> > On 2018-08-24 02:26 PM, right.maple.nut via talk wrote:
> > >
> > > Hello All,
> > >
> > > Like the Subject Line says, I'm setting up a ZFS File Server for my
> > > Home Network.
> > >
> > > Given that I will have to go to the trouble of setting up the Distro
> > > and Migrating the Linux Install to ZFS Root, I don't want to have to
> > > do this too many times.
> > >
> > > So, which Distro are the favourite for Running ZFS-on-Linux?
> > >
> > > Also, is there such a thing as a Linux Distro that is smart enough to
> > > give you a choice if you are willing to use non-GPL'ed code in the
> > > Installer, so that I can just Install Directly on a ZFS Pool?
> > >
> > > Thank You in Advance for your Input.
> > >
> > > Regards,
> > > Amos
> > >
> > >
> > >
> > > ---
> > > Talk Mailing List
> > > talk@gtalug.org
> > > https://gtalug.org/mailman/listinfo/talk
> > >
> >
> >
>
> --
> Alvin Starr || land: (905)513-7688
> Netvel Inc. || Cell: (416)806-0133
> al...@netvel.net ||
>
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Which Distro is Best for Running a ZFS-on-Linux Fileserver.

[David Mason via talk]

On Aug 25, 2018, 6:21 AM -0400, D. Hugh Redelmeier via talk , 
wrote:
> | From: David Mason via talk 
>
> | I am interested in this question too. I currently am running on Debian
> | 7.7 but I’m not sure I can upgrade to a more current version, which is
> | frustrating because I want to install Java (to run a Minecraft server)
> | and (when I tried, so while back) I couldn’t get it to install because I
> | had such an old version of Debian.
>
> Back in 2017 March 12 you mentioned a problem "Updating Wheezy to Jessie".
> Did you try Stewart's and Lennart's suggestions?

I put a little time into it, but didn’t get to success. Unfortunately, life 
intervened. It’s still important, but as you say, as you fall behind, it 
becomes more difficult to leap-frog into the present.

> This is exactly the kind of problem that a debian local users group might
> address. Recently I suggested that GTALUG could function as a debian LUG.
> I guess that here we have a test of this idea.

I think it already does. What did you have in mind beyond the mailing list?

> Even for the debian upgrade attempt it might be good to have a backup.

Yes, I know. :-)

> After enough time, the old hardware becomes obsolete too and it makes
> total system rebuild more sensible than update. In my most recent
> example, I jumped about 15 years ahead in hardware.

This system is <5 years old, and at the time was kind-of leading edge. so I’m 
not worried about that.
It’s a 4.4Tb raidz2 at 64% and has performed flawlessly. Unfortunately I don’t 
really have the time to do any serious digging right now, either.

How do others backup their ZFS systems? Getting a 4T external drive doesn’t 
seem like the best plan, but maybe there isn’t any other choice.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Which Distro is Best for Running a ZFS-on-Linux Fileserver.

[David Mason via talk]

I am interested in this question too.  I currently am running on Debian 7.7 but 
I’m not sure I can upgrade to a more current version, which is frustrating 
because I want to install Java (to run a Minecraft server) and (when I tried, 
so while back) I couldn’t get it to install because I had such an old version 
of Debian.

So I was thinking of FreeBSD, but OmniOS looks pretty interesting. I won’t be 
doing any upgrade for several months (because as far as I can see, I need to 
install a new ZFS system and copy it over. I can’t risk loosing this system). 
It would be good to have a backup, anyway. I could try this on a virtual 
system, but I don’t know long it would take to back-up over the network.

../Dave
On Aug 24, 2018, 3:51 PM -0400, D. Hugh Redelmeier via talk , 
wrote:
> | From: right.maple.nut via talk 
>
> | Hello All, Like the Subject Line says, I'm setting up a ZFS File Server
> | for my Home Network. Given that I will have to go to the trouble of
> | setting up the Distro and Migrating the Linux Install to ZFS Root, I
> | don't want to have to do this too many times. So, which Distro are the
> | favourite for Running ZFS-on-Linux? Also, is there such a thing as a
> | Linux Distro that is smart enough to give you a choice if you are
> | willing to use non-GPL'ed code in the Installer, so that I can just
> | Install Directly on a ZFS Pool? Thank You in Advance for your Input.
> | Regards, Amos
>
> Why not run a *BSD?
>
> Or something Solaris-related? Open Indiana
>
> - They more naturally run ZFS.
>
> - there is no clash-of-licenses drama.
>
> - the code is probably more tested.
>
> - there are even out of the box fileserver distros based on *BSD.
>
> - TrueOS, FreeNAS?
>
> Is there something that you don't like about UNIX (as opposed to
> Linux) for a file server?
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] GTATLUG talk about Atmel AVR

[David Mason via talk]

I regret to say that Peter died a week or so ago. He was always curious and 
willing to experiment with all sorts of things.

I miss him already, although I hadn’t seen him for several years.

../Dave
On Aug 15, 2018, 4:26 PM -0400, Kevin Cozens via talk , wrote:
> Greetings, Peter Hiscocks.
>
> I tried sending the following to your Ryerson email address but the message
> bounced so I am trying to get it to you this time via the GTALUG mailing list.
>
> I saw a video of your talk about Atmel AVR parts. Interesting talk. There
> are a few points you didn't make about the AVR microcontrollers.
>
> First, the instruction set is optimized for C. This is one of the reasons I
> chose to use AVR parts instead of PIC. That and no bank switching of memory
> with AVR.
>
> Second, you can use the AVR part without the need for an external crystal in
> many cases unless you are doing something requiring very precise timing. If
> you don't need precise timing you can use the built-in oscillator but it can
> also be calibrated to improve its timing accuracy.
>
> The real "Hello, World" program for hardware (such as microprocessors,
> microcontrollers, CPLDs, and FPGA's) is the blinking light.
>
> The question from the audience "Is the Atmel only on the Arduino?" was
> interesting. A lot of people associate the ATMEL AVR processors with Arduino
> when Arduino is more about the form factor and pinouts on a PCB and the
> software libraries, and programming method than about a given processor. You
> could have an Arduino board that doesn't use an ATMEL part.
>
> --
> Cheers!
>
> Kevin.
>
> http://www.ve3syb.ca/ | "Nerds make the shiny things that
> https://www.patreon.html/KevinCozens | distract the mouth-breathers, and
> | that's why we're powerful"
> Owner of Elecraft K2 #2172 |
> #include  | --Chris Hardwick
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Increasing interest in the Go language

[David Mason via talk]

I'm not a crypto expert. (Hugh?)

Most of the crates seem like they are either rust connectors to other
systems (like OpenSSL) or use https://github.com/briansmith/ring (which
follows BoringSSL, which is Google's fork of OpenSSL). On the other hand,
the links suggests that ring is currently failing tests, so ???

What Rust brings to the table, besides safety and speed, is a type system
that could give you a lot more confidence about the correctness of an
algorithm (though I haven't looked at these crates to see if they are
exploiting this potential).

../Dave

On 3 February 2018 at 23:26, Jamon Camisso via talk <talk@gtalug.org> wrote:

> On 2018-02-03 10:28 AM, David Mason via talk wrote:
> > I'd also comment on Rust being an interesting competitor to Go.
> >
> > Rust has better performance, complete statically determined safety
> > (enforced by the type system), no garbage collection, minimal runtime,
> and
> > an active group targeting WebAssembly (i.e. very high performance browser
> > programs). It's what you should be programming in if you think you need
> C.
> >
> > Go has a simpler type system and good-enough performance for many
> > applications. It might be what you should be programming in if you need a
> > higher-performance Python (but with a lot fewer libraries).
> >
> > They both interop with C and C++; I think Rust has a richer set of
> > libraries (crates they call them).
>
> https://github.com/rust-unofficial/awesome-rust#cryptography is what
> scared me off Rust when I was looking at it for a personal project to do
> with TLS certificates a year or so ago. Specifically, the multitude of
> crates for something that's so easy to get wrong made me wary.
>
> Has the situation changed much in Rust crypto land?
>
> Cheers, Jamon
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Increasing interest in the Go language

[David Mason via talk]

I'd also comment on Rust being an interesting competitor to Go.

Rust has better performance, complete statically determined safety
(enforced by the type system), no garbage collection, minimal runtime, and
an active group targeting WebAssembly (i.e. very high performance browser
programs). It's what you should be programming in if you think you need C.

Go has a simpler type system and good-enough performance for many
applications. It might be what you should be programming in if you need a
higher-performance Python (but with a lot fewer libraries).

They both interop with C and C++; I think Rust has a richer set of
libraries (crates they call them).

I considered both for my comparative programming languages course. I went
with Rust in the end because the other languages I talk about all have
garbage collection, and I wanted the students to see how a sophisticated
type system could give you performance as-good-as or better-than C and C++,
along with safety.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] the layers of things underneath Linux

[David Mason via talk]

Really interesting! Thanks Hugh for the pointer.

On 14 January 2018 at 13:37, D. Hugh Redelmeier via talk 
wrote:

> "Replace Your Exploit-Ridden Firmware with Linux - Ronald Minnich, Google"
>
> I just watched 
> Recommended if you want to know underground the landscape.
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Programming languages (in comparison?) - -was Learn Swift for Apple/iOS. Learn ??? for Google/Android.

[David Mason via talk]

On 12 December 2017 at 08:48, o1bigtenor via talk  wrote:

> A little tidbit for you as you continue your praise of the 'new' ff! (Yes
> I understand
> that chromium isn't ff but they both suffer under the same kind of
> programming
> - - - at least IMO!).
>

Which is exactly what ff57's use of Rust addresses.


> Michael Gilbert 
> 5:03 AM (2 hours ago)
> to debian-securit.
>

Of the 17 Chrome vulnerabilities listed, 7 would never have happened if
they were using Rust instead of C++.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Programming languages (in comparison?) - -was Learn Swift for Apple/iOS. Learn ??? for Google/Android.

[David Mason via talk]

On 11 December 2017 at 07:20, o1bigtenor via talk  wrote:

> I've tried vivaldi, min, chromium (that's chrome on linux), opera and ff.
> They all crap out. FF fails perhaps the best - - - it just gets slow.
> Opera just dies with no warnings at all! I don't like getting tracked ALL
> the time so chromium or variants are out.
>

FF57 is pretty awesome. Rust will help it become even better.

If all you want to do is rant that all browsers crap out under certain
loads, can we say, "so stipulated" and move on?

There's also Brave https://brave.com (with Brendan Eich behind it) which
might meet your needs better.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Learn Swift for Apple/iOS. Learn ??? for Google/Android.

[David Mason via talk]

I'll echo Hugh's recommendation for Rust. No GC (unlike golang), and
preformance equal to C/C++ but safe!​

I don't intent to write any more C.

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

[GTALUG] Raspberry PI hdmi problem

[David Mason via talk]

Evan's post reminded me of a problem I have left on the back-burner for a
year.

I also have a RPI Model B that I bought as a media centre, but for that to
work, it needs to go through my Pioneer receiver to my projector (because I
have other sources and I need to switch among them).  Unfortunately, that
doesn't work, although it works if directly plugged in to the projector.

I tried at the time to email someone at Raspberry, but never got a reply.
When I saw Evan's post, it reminded me that I'd really like to get it
working, and that somebody here might know something.

Thanks  ../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

[GTALUG] Updating Wheezy to Jessie

[David Mason via talk]

I have a server at home that tends to be install and forget.

It's running Wheezy and zfsonlinux.  Unfortunately zfsonlinux no longer
supports Wheezy, so upgrading isn't as straight-forward as I would hope.

I am following instructions at https://www.howtoforge.com/
tutorial/how-to-upgrade-debian-wheezy-to-jessie-stable-release/ and one of
the early steps is to do apt-get update which gets some "No public key
available" for the zfsonlinux/wheezy links. This leads to the following:

root@server:/etc/apt# apt-get upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
The following packages have been kept back:
  default-jdk default-jre default-jre-headless zfsutils
The following packages will be upgraded:
  apt apt-utils base-files bash bind9-host binutils cpio cups cups-client
cups-common cups-filters cups-ppdc dbus debian-archive-keyring dnsutils dpkg
  dpkg-dev e2fslibs e2fsprogs exim4 exim4-base exim4-config
exim4-daemon-light file fontconfig fontconfig-config foomatic-filters
ghostscript
  ghostscript-cups gnupg gpgv grub-common grub-pc grub-pc-bin grub2-common
gstreamer0.10-gconf gstreamer0.10-plugins-bad gstreamer0.10-plugins-base
  gstreamer0.10-plugins-good gstreamer0.10-x host hpijs hplip hplip-data
icedtea-6-jre-cacao icedtea-6-jre-jamvm icedtea-netx icedtea-netx-common
  isc-dhcp-client isc-dhcp-common java-common krb5-locales libapt-inst1.5
libapt-pkg4.12 libass4 libavcodec53 libavformat53 libavutil51 libbind9-80
  libc-bin libc-dev-bin libc6 libc6-dev libcairo-gobject2 libcairo2
libcomerr2 libcups2 libcupscgi1 libcupsdriver1 libcupsfilters1 libcupsimage2
  libcupsmime1 libcupsppdc1 libcurl3-gnutls libdbus-1-3 libdns88
libdpkg-perl libevent-2.0-5 libexpat1 libfontconfig1 libfreetype6 libfuse2
libgc1c2
  libgcrypt11 libgd2-xpm libgdk-pixbuf2.0-0 libgdk-pixbuf2.0-common libgif4
libgme0 libgnutls-openssl27 libgnutls26 libgs9 libgs9-common
libgssapi-krb5-2
  libgssrpc4 libgstreamer-plugins-bad0.10-0 libgstreamer-plugins-base0.10-0
libgtk-3-0 libgtk-3-bin libgtk-3-common libhpmud0 libicu48 libidn11 libisc84
  libisccc80 libisccfg82 libjasper1 libk5crypto3 libkadm5clnt-mit8
libkadm5srv-mit8 libkdb5-6 libkrb5-3 libkrb5support0 liblcms1 liblcms2-2
libldap-2.4-2
  liblwres80 libmagic1 libnspr4 libnss3 libnvpair1 libpcsclite1 libperl5.14
libpixman-1-0 libpng12-0 libpoppler19 libpython2.7 librsvg2-2
librsvg2-common
  libsane-hpaio libslp1 libsqlite3-0 libss2 libssl1.0.0 libsystemd-login0
libtasn1-3 libtiff4 libupnp6 libuutil1 libx11-6 libx11-data libx11-dev
libx11-doc
  libx11-xcb1 libxapian22 libxfixes3 libxi6 libxml2 libxpm4 libxrandr2
libxrender1 libxslt1.1 libxtst6 libxv1 libzfs2 libzpool2
linux-headers-3.2.0-4-amd64
  linux-headers-3.2.0-4-common linux-image-3.2.0-4-amd64 linux-libc-dev
live-tools locales mime-support multiarch-support openjdk-6-jdk
openjdk-6-jre
  openjdk-6-jre-headless openjdk-6-jre-lib openssh-client openssh-server
openssl perl perl-base perl-modules poppler-utils printer-driver-hpcups
  printer-driver-hpijs printer-driver-postscript-hp python-crypto
python-imaging python2.7 python2.7-minimal rpcbind spl spl-dkms ssl-cert
sudo tar tzdata
  tzdata-java unzip vim-common vim-tiny wget wpasupplicant xscreensaver
xscreensaver-data zfs-dkms zfsonlinux
199 upgraded, 0 newly installed, 0 to remove and 4 not upgraded.
Need to get 217 MB of archives.
After this operation, 4,578 kB disk space will be freed.
Do you want to continue [Y/n]?
WARNING: The following packages cannot be authenticated!
  zfs-dkms zfsonlinux libuutil1 libnvpair1 libzpool2 libzfs2 grub-pc
grub-pc-bin grub2-common grub-common spl spl-dkms
Install these packages without verification [y/N]?
E: Some packages could not be authenticated

And nothing happens.  I guess I have to tell it not to try to upgrade those
unauthenticated packages, but I don't know how.

Any advice appreciated.

Thanks  ../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Anyone want an Apple TV?

[David Mason via talk]

I hit reply not reply-all... and my email went to the world!!!  Is
there a reason for this?

I checked... if I do reply-all, it goes to the same 3 places, *plus* the UU
list.

*sigh*

Ohh... maybe it's because I top-posted! :-) :-)

../Dave

On 9 January 2017 at 07:36, David Collier-Brown via talk 
wrote:

> I have the second-last model, which I've tried and didn't like. I'll bring
> it to the meetings if anyone wants it.
>
> --dave
>
> --
> David Collier-Brown, | Always do right. This will gratify
> System Programmer and Author | some people and astonish the 
> restdav...@spamcop.net   |  -- Mark Twain
>
>
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
>
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Anyone want an Apple TV?

[David Mason via talk]

Hi Dave,

I love Apple TVs, and would gladly accept one (they are a bit hard to find).

However I have a terrible cold and am racing to get ready for the semester
that starts on Friday, so will not be able to get to the meeting tomorrow
(I live seriously in the boonies!).

If no-one takes you up on your offer, you could leave it in the tall metal
cabinet at the back of the room (I just went and checked and it's
unlocked), and I could pick it up on Wednesday morning.

BTW, what didn't you like about it?

Thanks  ../Dave

On 9 January 2017 at 07:36, David Collier-Brown via talk 
wrote:

> I have the second-last model, which I've tried and didn't like. I'll bring
> it to the meetings if anyone wants it.
>
> --dave
>
> --
> David Collier-Brown, | Always do right. This will gratify
> System Programmer and Author | some people and astonish the 
> restdav...@spamcop.net   |  -- Mark Twain
>
>
> ---
> Talk Mailing List
> talk@gtalug.org
> https://gtalug.org/mailman/listinfo/talk
>
>
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk

Re: [GTALUG] Kelly Gotlieb

[David Mason via talk]

Wikipedia on proximity fuze: https://en.wikipedia.org/wiki/Proximity_fuze

Article from Baltimore Sun - citing it as entirely American work, of
course:
http://articles.baltimoresun.com/1993-01-11/news/1993011049_1_fuse-proximity-smart-weapons

../Dave
---
Talk Mailing List
talk@gtalug.org
https://gtalug.org/mailman/listinfo/talk