Re: [Libcdio-devel] RFC - move libcdio from savannah.gnu.org to github?

[Pete Batard]

On 2024.09.07 02:10, E Shattow wrote:
> In 20-30 years' time all of these contributions and interactions will
> be a matter of public record and not hidden behind some walled garden
> that faded away into an acquisition of intellectual property rights.

You know that the one thing that Microsoft has been really good at, for 
these last 20-30 years, is both preservation of their existing data and, 
because it's their bread and butter, not trying to prevent access to 
developers?

They do have some history of doing things that actually cost them to 
help developers (MSDN documentation, Visual Studio Community), and in 
the process provide top class non garden-walled developer resources.

So, based on past history, I *currently* can not share your pessimism 
with regards to what the future *might* hold.

And as I stated right from the start, I think it is foolish to deprive 
us of a top of the line hosting framework because one person thinks, 
without proof, that a framework we can move to *might* become restricted 
one day, whereas (and this is currently true of GitHub with his numerous 
APIs that allow the retrieval of pretty much all of the data associated 
with a project hosted there) we should be able to take all the content 
with us and move to a different host if that ever happens.

Therefore, I can only reiterate that we should not make a decision based 
on distrust of some hypothetical future happenings, but instead we 
should make one based on the current stewardship that Microsoft/GitHub 
has provided to developers, and, so far, I do not see anything in that 
stewardship that either gives me pause, or makes me fear for a future 
where Microsoft may close the door, without warning, and take part of 
the projects contributions with them.

If anything, outside of the FSF which is motivated only by its 
altruistic goals and therefore difficult to best... at the price, of 
course, of not having the financial capabilities to provide the kind of 
top of the line project hosting facilities we seek, I have a lot more 
confidence in Microsoft not doing the kind of drastic reversal you 
alleged it might do, on account that they have the ability to divert 
resources from elsewhere if needed, as opposed to what a smaller source 
hosting entity like SourceForge recently tried to do, in trying to in 
ads into project downloads without the project owner's approval, because 
they were haemorrhaging money...

That's also part of the reason I see a move to GitHub as safer than 
anything else, because it does have the resources, and it has grown so 
big that any major restrictive move will be met with a huge public outcry.

> Do you believe this if we stray from the code hosting resources FSF
> directly supports?

I believe that the FSF is lacking, and will always be lacking the 
resources to meet the requirements we seek. That's unfortunate, and as 
an FSF supported I certainly wish it was otherwise, but I pretty much 
see it as a permanent fact.

> The way that "kids today write code" is Continuous Integration first

The way competent developers write code today is to use resources like 
CI, because they help detect *some* issues early, and help avoid 
degraded user experience, and user experience should be the prime 
motivation for everything a developer does (be it from using a GPL 
license, to how they choose to design their software).

If you think CI is something that only kids should be invested into, I 
think you may want to reconsider the way you develop software.

> then ask ChatGPT all the questions, submit pull requests and gamify
> everything in the comments section with likes and follows and emoji
> code-switching.

It seems to me like you want to bring the discussion back to these 
specific topics that you are unhappy with, and that I see as irrelevant 
to the current discussion.

I've been maintaining a very popular Open Source project on GitHub for 
more than 12 years now. I've had about 2500 issues opened against it and 
countless PRs. But I cannot say I have seen ANY of what you are reporting.

> I'm not a code contributor so I will just reinforce
> both the positions that yes it is important to have all the fancy
> development environment that young GitHub-only developers are
> comfortable existing within (or you will not have any contributors
> eventually);  and that it's probably a bad idea for social
> responsibility reasons in the world software ecosystem to chase after
> these contributors and lower the barrier for code contributions.

As both a code contributor and maintainer, it is my view that it's the 
project maintainer, and the project maintainer alone, that sets the 
standard for code contributions, and that being hosted on this or that 
environment will not somehow "lower" the quality of said contribution, 
unless the project maintainers chooses to.

> We're not waiting for Microsoft to ban or unban this github user or
> that one at the moment,

And we are not speculating at what Microsoft *might* do when we can 
balance both positive and negative positions that Microsoft has taken 
towards developers (some of the positive ones I highlighted above).

I'm sorry but I cannot follow you on a "Microsoft will surely screw this 
up" until Microsoft *actually* attempts to screw this up (and no, I 
don't think it will be "too late" by then).

As far as its GitHub stewardship is concerned, and even if I have loads 
of misgivings in other areas that Microsoft is involved (since Microsoft 
is a large company with loads of different teams working in different 
directions), I will continue to weigh in to an "innocent until proven 
guilty" stance.

> so what would be worthwhile to place this
> important software in that restricted position? What could very
> reasonably be done is to mirror the existing repo exported to GitHub
> (for example, GitLab, et.al.) whatever platform is popular at the
> moment, sure, here is what you want in the place that you expect it!

Except, and I have seen this on multiple projects, failing to declare a 
host as the official repo leads to user confusion and diminished 
contributions. Plus, the whole point, IMO, is to be able to bring things 
like CI to libcdio, which means that the host where we run CI has to 
become the official repo.

> But we can ignore the dogfooding of social media distractions. People
> are free to set up CI and making a mirror of the existing git repo

Except that, if it is not part of the official repo, then it becomes 
utterly pointless. You *will* alienate contributors because they will 
get the direct message that you don't think that their contributions are 
worthwhile.

> allows these "new style" coders to reference the github instance of
> the repository to do these things. There's no downside to meeting
> people halfway and not getting vacuumed up into the data mining casino
> and time-wasting trash bin that is the comments section of a pull
> request or bug tracker ticket not be preserved in the next decade
> onward.

As the project maintainer of a GitHub repository that has been preserved 
into the next decade onwards, and that has never felt "vacuumed up into 
the data mining casino and time-wasting trash bin that is the comments 
section of a pull request" (because it is really not that difficult to 
set boundaries and GitHub does offer you tool to avoid abuse if needed), 
I once again have to disagree to what looks to me like an uneducated 
stance on the matter. My current experience is in direct contradiction 
of what you state and I have never had the feeling that being hosted on 
GitHub took any of the freedom I wanted, as a project maintainer, to 
conduct the project in the way I saw fit, and with a strong adherence to 
the philosophy of the FSF's goals.

> Bit of a rant but really you want this to get a good full discussion I
> think it's a question we should pose to the FSF, what is it we want
> that is not already there?

Users. Loads of them. So that you cannot pull up the kind of 180 you 
assume GitHub is poised to pull, without triggering a massive uproar. As 
well as resources. Loads of them. So that you can assign full time 
developers to take care of maintaining and improving your hosting and 
add new features. And CI that is compatible with GitHub Actions, because 
it is top class (even if it has some drawbacks) and developers are tired 
of the constant NIH they have to wade through when CI is *precisely* not 
something they should have to rewrite when/if they switch host.

As much as I do my best to help the FSF, and am partial to their goals, 
I don't see any of the 3 points above ever happening with their source 
hosting... which is precisely why we are having the current discussion.

Regards,

/Pete