Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
Jaromil wrote: > from what I understand Simon is saying here, he is not being > understood really. What he is suggesting is to add a small > documentation text that, in case one doesn't knows what to choose, > mentions what is the most common choice in case of most common setups > (today is clearly i686) and when/why to choose otherwise, also linking > to simple explanations that may help orienting people to choose their > installation medium according to their target hardware. Yes indeed, that is exactly it. I may have got some of my terms wrong since I've kinda lost track of what image runs on what hardware. As I mentioned in an earlier message, it's only recently that I got relatively new hardware at work and so was stuck on (IIRC) i386 images (for the VMs) as the newest that would run on any of my hosts. It is my understanding (and I stand to be corrected) that "new" processors can all run amd64 images and that this would be the "default" choice of image unless you know you need something else. Alessandro Selli wrote: > Right. This is the reason because a lot of people download x86 images: all > those who still have 32bit hardware, they're installing GNU/Linux on it, > because it is the only current and maintained OS that runs on a platform that > proprietary, mainstream OSes abandoned years ago. > > 32bit Intel compatible platforms are not going to disappear soon as several > embedded, specialized appliances, networking devices, IoT devices, > mediacenters and so forth are still being produced based on 32bit processors. Indeed, hence why no suggestion to remove the images - just to put a note in to point those who do not run "old" hardware to information to help them choose the right image for their kit. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On Thu, 21 Jun 2018 at 08:41:18 +0200 KatolaZ wrote: > I think you probably missed the point here. believe it or not, there > are genuinely thousands of people who use i686 images because they > happen to have i686 machines. The percentage of those in Devuan is not > that different from the percentage seen in Debian. > > I can't see any good reason not to use an i686 image if you have an > i686 CPU... Right. This is the reason because a lot of people download x86 images: all those who still have 32bit hardware, they're installing GNU/Linux on it, because it is the only current and maintained OS that runs on a platform that proprietary, mainstream OSes abandoned years ago. 32bit Intel compatible platforms are not going to disappear soon as several embedded, specialized appliances, networking devices, IoT devices, mediacenters and so forth are still being produced based on 32bit processors. Alessandro ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
dear Katolaz On Thu, 21 Jun 2018, KatolaZ wrote: > On Wed, Jun 20, 2018 at 11:58:56PM +0100, Simon Hobson wrote: > > I think you missed the point. > > > > All I’m suggesting is that on the download page there are the > > AMD64 images first, then a note giving very basic suggestions not > > to use i[3-6]86 images unless you need to AND PROVIDE LINKS TO > > SOME GOOD WRITEUPS. It’s not trying to force people to do anything > > - just give them a pointer to information that (apparently) quite > > a few people aren’t aware of. > > I think you probably missed the point here. believe it or not, there > are genuinely thousands of people who use i686 images because they > happen to have i686 machines. The percentage of those in Devuan is not > that different from the percentage seen in Debian. > > I can't see any good reason not to use an i686 image if you have an > i686 CPU... from what I understand Simon is saying here, he is not being understood really. What he is suggesting is to add a small documentation text that, in case one doesn't knows what to choose, mentions what is the most common choice in case of most common setups (today is clearly i686) and when/why to choose otherwise, also linking to simple explanations that may help orienting people to choose their installation medium according to their target hardware. I believe this can be a useful addition and wish someone preferably native in english language can contribute it. ciao ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On Wed, Jun 20, 2018 at 11:58:56PM +0100, Simon Hobson wrote: > KatolaZ wrote: > > >>> Of course they are, it's all over the Internet. > >> > >> Being pedantic, that’s not the same - and you **should** know that. > >> IMO there’s a choice to be made - do we (collectively) want to be > >> inclusive and support all those who don’t know much about computing but > >> want to try an alternative to Mac/Windows; or do we (collectively) want to > >> stay elitist and show an attitude that “people should know these things” ? > > > > [cut] > > > > Unfortunately, there is no middle-ground here. We have been telling > > people for years that running proprietary software is potentially > > harmful for their privacy and security. The result is that 98% (and > > maybe more) of all the CPUs on this planet run a proprietary operating > > system with proprietary software. > > > I think you missed the point. > > All I’m suggesting is that on the download page there are the AMD64 images > first, then a note giving very basic suggestions not to use i[3-6]86 images > unless you need to AND PROVIDE LINKS TO SOME GOOD WRITEUPS. It’s not trying > to force people to do anything - just give them a pointer to information that > (apparently) quite a few people aren’t aware of. I think you probably missed the point here. believe it or not, there are genuinely thousands of people who use i686 images because they happen to have i686 machines. The percentage of those in Devuan is not that different from the percentage seen in Debian. I can't see any good reason not to use an i686 image if you have an i686 CPU... HND KatolaZ -- [ ~.,_ Enzo Nicosia aka KatolaZ - Devuan -- Freaknet Medialab ] [ "+. katolaz [at] freaknet.org --- katolaz [at] yahoo.it ] [ @) http://kalos.mine.nu --- Devuan GNU + Linux User ] [ @@) http://maths.qmul.ac.uk/~vnicosia -- GPG: 0B5F062F ] [ (@@@) Twitter: @KatolaZ - skype: katolaz -- github: KatolaZ ] signature.asc Description: PGP signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On Wed, 20 Jun 2018 at 14:07:51 +0200 KatolaZ wrote: > And this is not at all elitism: it's just the humble admission that > what is "True" for me might not be "THE Truth" for everybody, and that > what is important, crucial, fundamental for me might be just bullshit > for the rest of the world. And the rest of the world might actually be > right... > > HND > > KatolaZ Where is the "Like" button? :-) Alessandro ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
KatolaZ wrote: >>> Of course they are, it's all over the Internet. >> >> Being pedantic, that’s not the same - and you **should** know that. >> IMO there’s a choice to be made - do we (collectively) want to be inclusive >> and support all those who don’t know much about computing but want to try an >> alternative to Mac/Windows; or do we (collectively) want to stay elitist and >> show an attitude that “people should know these things” ? > > [cut] > > Unfortunately, there is no middle-ground here. We have been telling > people for years that running proprietary software is potentially > harmful for their privacy and security. The result is that 98% (and > maybe more) of all the CPUs on this planet run a proprietary operating > system with proprietary software. I think you missed the point. All I’m suggesting is that on the download page there are the AMD64 images first, then a note giving very basic suggestions not to use i[3-6]86 images unless you need to AND PROVIDE LINKS TO SOME GOOD WRITEUPS. It’s not trying to force people to do anything - just give them a pointer to information that (apparently) quite a few people aren’t aware of. There’s an analogy with a situation where I used to work. The authorities created a new one-way system that turned the road past our office into one of the main ways out of town. Most of us think it was a stupid move, and it certainly made getting out of the office car park a PITA at peak times - not to mention the hazard of trying to cross as a pedestrian. But even after several years, we still got people going the wrong way - quite frequently. A colleague made it his mission and kept petering the local highways department sending them lots of photos, even catching up with some of the drivers (seeing where they went and popping out to speak to them) and finding out the reason for their mistake. At first, the authorities just kept repeating that the signage was there - so it’s all the drivers’ fault for not seeing the signs. They were in the same state as some people on this “tell people or not tell them” discussion - the information is around to be seen, it’s not our responsibility to make any allowance for people not having seen it. The fundamental issue was that the signage was designed by people who knew about the one way system - and shared the same problem shared with direction signs designed by people who know the local area. At one junction, yes there’s a sign, but it not at the right height to be in the driver’s eyeline, and it’s not where they would be looking if their intention is to turn the wrong way - so it didn’t get seen Eventually the highways department relented - made some of the signs bigger and put some “NO ENTRY” painted signs on the road surfaces. Guess what - in spite of the previous signage supposedly being quite adequate, the incidence of people going the wrong way dropped from several per week (sometimes several in a day) to perhaps as little as one/month. Just to prove the point, someone knocked down the lamp post one sign was attached to, and the incidence of people going the wrong way went up again for the very long time (many months) before the lamp post was replaced and the sign re-instated. So people can keep complaining that others are missing information that’s “all there to be seen”, or add a one-liner along the lines of “BTW - had you missed this ?" ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
Quoting Erik Christiansen (dva...@internode.on.net): > On 20.06.18 12:04, Simon Hobson wrote: > > FWIW, even technical users can lack what some may think is “really > > basic knowledge” - I fell that the most important thing I’ve learned > > over the years is just how much I don’t know ! > > Despite using *nix exclusively for three decades now, linux for around > two, matching a linux distro to my cpu requires more chip knowledge than > I have. I remember grubbing about on the net to try to find out what > sort of beast my: > > $ cat /proc/cpuinfo > processor : 0 > vendor_id : CentaurHauls > cpu family : 6 > model : 13 > model name : VIA C7 Processor 1500MHz > ... > clflush size: 64 > cache_alignment : 64 > address sizes : 36 bits physical, 32 bits virtual > > might be. The last line suggested that I needed i386, but line three > hinted vaguely at i686, maybe. The internet was no help when I looked, > some years ago. Taking a stab in the dark, I found that 686-pae runs > fine on it, but how would one know in advance? That's an i386 clone with some 686 features such as PAE. Obviously, this is pretty obscure data, so you would _not_ know in advance. (The VIA C7 'Esther' core was manufactured for C7 by Centaur Technology, thus the vendor_id string.) How to make an optimal kernel for one: https://blog.laczik.org/centos-6-5-kernel-compile-for-via-c7-cpu/ > The devuan ascii I run on my quad-core celeron host is also i686, as I > figure it's new, and 686 has to be better than anything with lower > numbers, right? Possibly you are fully aware of this, but: Code compiled for generic i386 will run on any i386-family processor whatsoever, but sacrificing some performance and other advantages possible if you use a kernel better tuned to the _specific_ i386-family CPU. Complicating this picture, more and more distros that still ship an i386 flavor have been making the judgement call to use an i686-optimised (Pentium Pro-optimised) kernel for installation and default operation, which then prevents installation on literal 386, 486, Pentium, Pentium-MMX, etc. pre-686 microarchitectures.. Fairly lucid clarification: https://www.redhat.com/archives/rhl-list/2006-October/msg03684.html ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On Wed, Jun 20, 2018 at 07:08:36AM -0500, Nate Bargmann wrote:
> At one time there were comments that not all packages were available on
> amd64. Even today that can be true, especially for some third party
> stuff. To wit, a prominent manufacturer of amateur radio equipment
> offers utilities for its hardware for Linux in addition to the other
> two, but the Linux version is 32 bit only and is dynamically linked to
> 32 bit libraries. Requests for amd64/x86_64 versions have been ignored
> as have requests for armhf.
Yeah, but there's many more packages that haven't been ported to 32-bit:
chdist create stretch-amd64
chdist create stretch-i386
$EDITOR ~/.chdist/stretch-{amd64,i386}/etc/apt/sources.list
chdist update stretch-amd64
chdist update stretch-i386
chdist compare-bin-packages stretch-amd64 stretch-i386|grep ' UNAVAIL .'|wc -l
139
chdist compare-bin-packages stretch-amd64 stretch-i386|grep ' UNAVAIL $'|wc -l
383
Alas, there's plenty of packages producing some compat binaries on one of
the archs. Thus, let's see what sources produce at least one binary,
something that chdist can't provide without some extra code:
grep-dctrl . -nsSource:Package
~/.chdist/stretch-amd64/var/lib/apt/lists/*_Packages|
cut -d' ' -f1|sort|uniq >amd64
grep-dctrl . -nsSource:Package
~/.chdist/stretch-i386/var/lib/apt/lists/*_Packages|
cut -d' ' -f1|sort|uniq >i386
diff -u9 amd64 i386 |grep ^+|wc -l
23
diff -u9 amd64 i386 |grep ^-|wc -l
86
i386 only:
atitvout dgen digitools fdflush fenix-plugins gatos gmod
google-android-build-tools-installer longrun lphdisk mig nyquist
pcsx2 pforth pixbros s3switch smlsharp spellcast steamcmd
xserver-xorg-video-geode yforth zsnes
amd64 only:
accelio afio agrep ariba bcal bcftools blasr blimps bwa caffe-contrib
caja-dropbox cen64 cen64-qt cluster3 crac cufflinks darktable drdsl
embassy-phylip eztrace-contrib falcon fdkaac fermi-lite fitgcp fsm-lite
gasic genesisplusgx giira gmap hisat2 hwloc-contrib iausofa-c iva
jellyfish kissplice kpatch libisal libretro-snes9x libsdsl libssw
libvcflib metis-edf mgltools-cmolkit mokutil mrs mssstest nastran nttcp
obs-studio ocaml-fdkaac openstack-debian-images openzwave-controlpanel
parafly pbbam pbbarcode pbdagcon pgcharts powder prctl princeprocessor
r-cran-rjsonio rapmap rna-star rocksdb rsem salmon sga shim sift
snap-aligner snaphu soapdenovo soapdenovo2 spades srst2 starpu-contrib
swarm-cluster tegrarcm tinyows tome tophat uftrace varscan vsearch xsnow
Meow!
--
⢀⣴⠾⠻⢶⣦⠀ There's an easy way to tell toy operating systems from real ones.
⣾⠁⢰⠒⠀⣿⡁ Just look at how their shipped fonts display U+1F52B, this makes
⢿⡄⠘⠷⠚⠋⠀ the intended audience obvious. It's also interesting to see OSes
⠈⠳⣄ go back and forth wrt their intended target.
___
Dng mailing list
Dng@lists.dyne.org
https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On 20.06.18 14:07, KatolaZ wrote: > I will continue telling people what I think they should know, but the > only way out of ignorance is knowledge, awareness, and individual > action. You can't force people to get interested, to learn, to become > responsible, to understand, to agree with you, to embrace your > personal ideal of freedom. And supporting high maintenance unmotivated users would be unachievable. If curiosity and motivation are missing, it's a lost cause. But diversity is good, I think. What would be unbearable would be if everyone had to suffer M$ or systemd chains. > You can just "Act as if the maxims of your action were to become > through your will a universal law of nature". > > And this is not at all elitism: it's just the humble admission that > what is "True" for me might not be "THE Truth" for everybody, and that > what is important, crucial, fundamental for me might be just bullshit > for the rest of the world. And the rest of the world might actually be > right... Well, they are right in that linux is not the turnkey deliverable which so many need. Driverless cars and driverless OSs are the current drip-dry fashion. But who wants to be in the majority, the crowds are too big. Many thanks for the inspired work. Erik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
At one time there were comments that not all packages were available on amd64. Even today that can be true, especially for some third party stuff. To wit, a prominent manufacturer of amateur radio equipment offers utilities for its hardware for Linux in addition to the other two, but the Linux version is 32 bit only and is dynamically linked to 32 bit libraries. Requests for amd64/x86_64 versions have been ignored as have requests for armhf. A recent email thread blamed "Linux" for making this so hard because multi-arch must be enabled. I noted that if the manufacturer would also provide 64 bit versions of their utilities that this pain would be avoided. Curiously, there was no reply to that fact. It should have been noted that choosing a distribution with multi-arch enabled out of the box, such as Ubuntu, would make installation no more painful than on Windows 64. To date the manufacturer's Web site states that 32 bit compatibility libraries are required. Some users may make the choice to simply install a 32 bit distribution as a result. My suggestion, and solution, is to spin up a 32 bit VM to run the utilities for the odd time that I need them. - Nate -- "The optimist proclaims that we live in the best of all possible worlds. The pessimist fears this is true." Web: http://www.n0nb.us GPG key: D55A8819 GitHub: N0NB signature.asc Description: PGP signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On Wed, Jun 20, 2018 at 12:04:32PM +0100, Simon Hobson wrote: > Alessandro Selli wrote: > > > >> The problem is that people are not told why they should away from i386, and > > > > Of course they are, it's all over the Internet. > > Being pedantic, that’s not the same - and you **should** know that. > IMO there’s a choice to be made - do we (collectively) want to be inclusive > and support all those who don’t know much about computing but want to try an > alternative to Mac/Windows; or do we (collectively) want to stay elitist and > show an attitude that “people should know these things” ? [cut] Unfortunately, there is no middle-ground here. We have been telling people for years that running proprietary software is potentially harmful for their privacy and security. The result is that 98% (and maybe more) of all the CPUs on this planet run a proprietary operating system with proprietary software. We have been telling people for years that they should sign all their emails and encrypt important data. The result is that strong encryption is used by an ever smaller fraction of users than 10 years ago. We have been telling people that they should use only open formats to store their data. The result is that closed formats and restrictive protocols have become standards accepted by the W3C. We have been telling people for years that online social platforms are used to massively spy on their users. The result is that today those platforms represent 50%-60% of the overall Internet traffic. We have been telling people that systemd is a fatal risk for the entire Linux ecosystem. The result is that systemd has become default in 90% (and maybe more) of Linux distributions. I will continue telling people what I think they should know, but the only way out of ignorance is knowledge, awareness, and individual action. You can't force people to get interested, to learn, to become responsible, to understand, to agree with you, to embrace your personal ideal of freedom. You can just "Act as if the maxims of your action were to become through your will a universal law of nature". And this is not at all elitism: it's just the humble admission that what is "True" for me might not be "THE Truth" for everybody, and that what is important, crucial, fundamental for me might be just bullshit for the rest of the world. And the rest of the world might actually be right... HND KatolaZ -- [ ~.,_ Enzo Nicosia aka KatolaZ - Devuan -- Freaknet Medialab ] [ "+. katolaz [at] freaknet.org --- katolaz [at] yahoo.it ] [ @) http://kalos.mine.nu --- Devuan GNU + Linux User ] [ @@) http://maths.qmul.ac.uk/~vnicosia -- GPG: 0B5F062F ] [ (@@@) Twitter: @KatolaZ - skype: katolaz -- github: KatolaZ ] signature.asc Description: PGP signature ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
On 20.06.18 12:04, Simon Hobson wrote: > FWIW, even technical users can lack what some may think is “really > basic knowledge” - I fell that the most important thing I’ve learned > over the years is just how much I don’t know ! Despite using *nix exclusively for three decades now, linux for around two, matching a linux distro to my cpu requires more chip knowledge than I have. I remember grubbing about on the net to try to find out what sort of beast my: $ cat /proc/cpuinfo processor : 0 vendor_id : CentaurHauls cpu family : 6 model : 13 model name : VIA C7 Processor 1500MHz ... clflush size: 64 cache_alignment : 64 address sizes : 36 bits physical, 32 bits virtual might be. The last line suggested that I needed i386, but line three hinted vaguely at i686, maybe. The internet was no help when I looked, some years ago. Taking a stab in the dark, I found that 686-pae runs fine on it, but how would one know in advance? The devuan ascii I run on my quad-core celeron host is also i686, as I figure it's new, and 686 has to be better than anything with lower numbers, right? Things were much more defined in my days of sysadminning a whole department's sparc servers and desktops. Erik ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
Re: [DNG] Educating people (Was: One week into Devuan 2.0 ASCII -- Some stats)
Alessandro Selli wrote: > >> The problem is that people are not told why they should away from i386, and > > Of course they are, it's all over the Internet. Being pedantic, that’s not the same - and you **should** know that. IMO there’s a choice to be made - do we (collectively) want to be inclusive and support all those who don’t know much about computing but want to try an alternative to Mac/Windows; or do we (collectively) want to stay elitist and show an attitude that “people should know these things” ? When you make that statement that a possibly newbie user should know such technical details because “it’s all over the internet" then you are falling into the elitist trap of assuming that those with less knowledge than yourself *should* have that knowledge - and more importantly *should* know that they need to gain that knowledge and where to gain it from. The big problem being that these people don’t know what they don’t know, have probably never seen these technical discussions “all over the internet”, and even if they have then they will not have been able to understand what it means in practice. There is a middle ground here. No need to put all the arguments on the DL page, but put a note along the lines of “unless you are certain you need to use i386 then you should use the AMD64 images, for more information see ”. It’s not offensive to those who do need those versions, it’s educational to those who didn’t know, and it doesn’t take up much space on the page. FWIW, even technical users can lack what some may think is “really basic knowledge” - I fell that the most important thing I’ve learned over the years is just how much I don’t know ! At a previous place I only got “hand me down” hardware as the company manglement were (and still are) wedded to a mindset of “if it didn’t come from Redmond then we don’t want to know” and so the services I ran on Debian only got hardware that had been retired from Windoze hosting due to Windows Server requirements. Eg, At one time I had a load of Dell 2850 servers as they’d had to upgrade to 2950 (or better) - I vaguely recall that it was a requirement of HyperV for a better processor. For a long time, ALL my servers were i686 images. This was partly because I had older servers (as my immediate manager (a knowledgable and good chap) put it, 9 years past their refresh date), but partly because it was a long time after it became the case before I found out that newer Intel processors run AMD64 images. Of course, even when I got a host capable of running AMD64 code, I still had to keep my VMs compatible with the older hosts. A couple of years ago I managed to get some AMD64 based machines as hosts - still ancient hand-me-downs but massive upgrades in capability and much reduced power consumption - and started migrating stuff to AMD64. But before I could get very far they made the 2 of us left redundant and got rid of it all - the “I don’t understand it so it’s going” mentality of a particular mangler. I did feel sorry for some of the customers as this mangler produced one screwup after another due entirly to his “change stuff and see what breaks” approach to systems reliability - a five day email outage that I could have fixed in a couple of hours (most of that being time to copy the mail store), customers DNS breaking because he had no clue about how DNS secondaries work when you just remove the master, …. But that’s drifting off the topic - it’s not fair to assume that everyone else knows what you do, and if they don’t then it’s their fault. ___ Dng mailing list Dng@lists.dyne.org https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
