>On Wed, 3 May 2017 13:02:58 -0400
>LM <[email protected]> wrote:

> I was wondering if there's anyone on the list who enjoys modifying
> Open Source code and adding their own patches to projects.

There are some of us, yes, more-or-less. :)

> Does anyone else on this list enjoy modifying code or does anyone
> know of any good forums to discuss this sort of thing?  Would
> appreciate hearing from you.

While not strictly related to modifying open-source, I found Usenet to
be an invaluable source of deep knowledge. In the domain of
electronics, most (almost all?) of the deep knowledge that can be found
on the Web is actually Usenet posts injected into a web forum shell.

Relating to software in general, the group I got the most deep
knowledge from is comp.arch.embedded . That group talks a lot about
writing code that needs to fit into a particular environment, which is
a lot like writing patches. Relating to open source, the most
interesting group is alt.os.linux which would get occasional - but
dependably regular - insightful discussions. I should probably rephrase
that as "rare", but I think this will sell better. xD

> I'm currently working on... (...) Would be very interested to hear
> what others are working on when they modify or add their own features
> to source code.

I had some patches included in polipo, a caching web proxy. When the
project was getting orphaned, I wanted to step in and adopt it but I
had too much going on in life and a caching web proxy simply wasn't
high enough on the list of priorities.

I'm probably going to completely miss the topic in the rest of the
mail, since I'm about to talk about writing whole projects and not
modifying existing ones, but WTH. :) You can skip the entire rest of
the mail if we should focus the conversation to modifying and patching.

Several years ago I wrote an implementation of NetBIOS-over-TCP, the
protocol used under SMB and CIFS (known in open-source-land as the
Samba suite). I first started writing bits and pieces just for fun,
then realized it's a great learning experience and finally later
thought I might get some money by licensing it. I published it on
Gitorious but didn't actually promote it in any real capacity. As far
as I can tell, noone ever heard of my project, even though it is the
most complete implementation of the protocol. Says I. :D It even
supports the client-side name service over TCP! Too bad there are no
servers anywhere - no other software supports any kind of name service
over TCP. :)

After Gitorious went under, Nbworks, as the project is named, was not
on the Internet for a while, but I've since uploaded it to Github. It's
now on https://github.com/akuktin/nbworks . It has lots of manpages.

Latter years went on without me being a notable presence on the
Internet, until about a year ago when I got serious about implementing
an open-source computer. As in "hardware".

See, the reason I got into LFS in the first place was that I wanted to
have a transparent and non-bloated computer. After finishing a LFS that
I was happy with, and living with it for a few years, I wanted to
expand my influence to the underlying hardware. UEFI and EFI firmware
doomed the desire and I started looking. My first few attempts were
revolving about taking smartphones and installing some variant of LFS
on them. There are even some mails in the archive of this very list
from that period of my life. They detail some of the attempts and the
many failures I experienced.

As time went on, I came to understand this was a losing proposition,
for two interlocking reasons. First of those is that no public
open-computer project that exists or existed had support from the
hardware manufacturers (OEM-s). Therefore, running software on chips
first requires those chips to be reverse-engineered. This is time and
labour-consuming. It results in software that is tightly tied to the
chips - and this is the second reason - and can be invalidated by the
OEM simply shipping modified hardware. Even if the OEM was kind and
didn't make changes just to invalidate reverse-engineering, the simple
market forces, business/operation cycles and innovation would
eventually invalidate all reverse-engineering efforts.

We would have to spend - literally - the rest of our days just
reverse-engineering chips to run software on.

So I took the next exit off the rat race highway. I would
short-circuit the problem and build my own computer. (Insert the
Bender-from-Futurama-with-blackjack-and-hookers meme here. :) )

It's called Special Snowflake, works in the simulator and is currently
in labour pains, as I try to bring it into physical reality.
Specifically, the current design doesn't fit into the FPGA chip it is
meant to fit in. I'm working on it when I can (since the New Year the
project almost stalled) but in the meantime you can examine the code on
https://github.com/akuktin/special_snowflake . I'm going to help anyone
who wants to play with it - there's a LOT of new stuff to learn if
you never did this before. :)

-- 
Svi moji e-mailovi su kriptografski potpisani. Proverite ih.
All of my e-mails are cryptographically signed. Verify them.
--
You don't need an AI for a robot uprising.
Humans will do just fine.
--

Attachment: signature.asc
Description: PGP signature

-- 
http://lists.linuxfromscratch.org/listinfo/lfs-chat
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page

Reply via email to