Live data from Hacker News

Give up GitHub: The time has come

sfconservancy.org

151–160 of 585 posts

Re: Give up GitHub: The time has come

#151
post #104

I publish my public FOSS work on a self-hosted Gitea. I don't allow account creation, and people can send me pull requests by email. That said, I think one thing (other than interface and brand loyalty) that keeps FOSS projects on GitHub is network effects. You can reasonably expect to search it and find the projects you're looking for, and your account lets you use the issue tracker and pull requests on other projec…

Gitea (and a few others) are working on federation for pull requests, which would allow someone to fork your project to their own server, and send a pull request offering you to merge from their server into yours. It also builds on top of ActivityPub which is supposed to allow federation with the greater ActivityPub ecosystem ("fediverse"). I guess this would allow people to like or comment your issue or pull request…

Hot damn, that sounds awesome. Easy PRs would be a huge step to getting me off GitHub.

Re: Give up GitHub: The time has come

#152
post #56

Earlier quoted context omitted.

> refusal to use Git[1] due to its complexity I kind of like this friction. If a software developer finds git too complex, that’s an important signal that I should minimize my exposure to them. I’d never work in a team that dumb, and I wouldn’t want any dependencies on their code. Of course there are other reasons not to use git, but complexity is not one.

My take is completely opposite from yours. I think it's great when someone "gets" that git is an extremely powerful and capable tool with a horribly unintuitive user interface. Whether you call that "complexity" or just "terrible UX", to me that's a sign that someone has good instincts regarding the risks associated with this kind of "complexity". Whether I agree that a decision to shun git for that reason is justifi…

^ this. The comment's arrogance and narrow-mindedness is preposterous, and would be a strong signal to me to stay away.

I find Git's CLI and workflow model complex because I had the "misfortune" of using other DVCS products (Mercurial, Bitkeeper, Bazaar NG, Darcs, etc) before it. All did a far superior job of presenting roughly the same conceptual model in their command-line tools. Git's is a step backwards.

Git didn't invent distributed version control, it didn't perfect it, it wasn't particularly superior to the others [it does have some benefits in terms of performance tho, yes], it was merely at the right place (the Linux kernel when Linus got sick of Bitkeeper's business model) at the right time (when people finally got sick of CVS and SVN garbage.)

And the rise of GitHub was certainly part of the rise of Git's prominence.

Slightly different circumstances and it could have been any of the other open source distributed revision control systems instead.

Re: Give up GitHub: The time has come

#153
> What case law, if any, did you rely on in Microsoft & GitHub's public claim, stated by GitHub's (then) CEO, that: “(1) training ML systems on public data is fair use, (2) the output belongs to the operator, just like with a compiler”? In the interest of transparency and respect to the FOSS community, please also provide the community with your full legal analysis on why you believe that these statements are true.

> If it is, as you claim, permissible to train the model (and allow users to generate code based on that model) on any code whatsoever and not be bound by any licensing terms, why did you choose to only train Copilot's model on FOSS? For example, why are your Microsoft Windows and Office codebases not in your training set?

I'm not sure I buy this argument. In the first point, the authors state that Copilot was trained on public data. In the very next point, they slightly tweak it by saying training was done on "any" code which loses the distinction between public and private code. Obviously Windows and Office are not public code.

I also interpreted "public data" to mean they trained on codebases that explicitly specified, say, MIT licenses or other permissible licenses. That seems like fair use to me. Those licenses don't explicitly restrict training AI models on their codebases do they? It's ironic if these licenses started banning AI training now though. That would effectively mean Copilot would be sole trained AI model.

I'm happy to be proven wrong though. In general I have a distrust of Copilot. I fear it would make individuals worse programmers in the end at the cost of productivity

Re: Give up GitHub: The time has come

#154

> If it is, as you claim, permissible to train the model (and allow users to generate code based on that model) on any code whatsoever and not be bound by any licensing terms, why did you choose to only train Copilot's model on FOSS? For example, why are your Microsoft Windows and Office codebases not in your training set? This is my favorite question about Copilot ever.

My favorite answer is: if you included Microsoft source code the quality of suggestions drops below viable product.

Re: Give up GitHub: The time has come

#155

Earlier quoted context omitted.

Depends on the uni. Mine focused on fundamentals, almost to a fault; we learned algorithmic complexity and relational normal-forms but never touched eg. JavaScript, Python, or - ironically - git. Practical skills were very much not the entrypoint; I had to learn most of that myself.

Yeah I had to learn Assembly in college. We also had to physically print out our C++ code and turn it in to the professor. Did it make me a better programmer? Maybe but to this day I still haven't had to use it. Most of my practical skills I picked up on the job. P.S. I've seen the professor grading the printed out programs and he'd do it by flipping to the last page which was supposed to have the result output, and…

Did we go to the same school? Halfway through the semester he gave up on even that pretense and started throwing them out in front of us and giving everyone a C-

Re: Give up GitHub: The time has come

#156

> If it is, as you claim, permissible to train the model (and allow users to generate code based on that model) on any code whatsoever and not be bound by any licensing terms, why did you choose to only train Copilot's model on FOSS? For example, why are your Microsoft Windows and Office codebases not in your training set? This is my favorite question about Copilot ever.

The boring answer is probably something along the lines of “copilot was trained by employees of OpenAI who aren’t technically MS employees”. When I worked at MS you had to jump through all sorts of hoops to get access to code from other orgs. I can’t imagine what BS you’d need to do to give access to a vendor.

Re: Give up GitHub: The time has come

#157

> For its part, Git was designed specifically to make software development distributed without a centralized site. Just yesterday I had to explain the basic premise/history of Git to a young intern. I had asked him if he was using Git to manage his little pet project the company gave him to play with. “No”, he replied, he didn’t know what the company’s policy was to posting code in public on GitHub. As I explained to…

Going through security certification procedure, I was asked to list all the third party SAAS things we use. I didn't include github, gitlab, or anything else, because we don't use it. The auditor was going off on a tirade about how lack of version control is not okay at all, so convinced they were that 'no github or gitlab' must therefore mean 'no version control'. The mind boggles. He barely believed me when I showe…

maybe this? https://www.gerritcodereview.com/

Re: Give up GitHub: The time has come

#158
post #56

Earlier quoted context omitted.

> refusal to use Git[1] due to its complexity I kind of like this friction. If a software developer finds git too complex, that’s an important signal that I should minimize my exposure to them. I’d never work in a team that dumb, and I wouldn’t want any dependencies on their code. Of course there are other reasons not to use git, but complexity is not one.

Does a typical enterprise need a distributed source control system? It makes a lot of sense for open source. Doesn't the uptake in Github which centralizes this distributed system kind of invalidate its main tenant? I haven't been a hundred percent sure the overhead was ever worth it at most other types of paid gigs over the years. Adding complexity without value is a mistake imo. Maybe it's my own fault I haven't se…

> Does a typical enterprise need a distributed source control system?

Yes, unless you are working on a tiny number of source files. The value is that you can work on entire copies of a code base at a time instead of single files. Git does a much better job than making lots of local copies of source code, passing around big diff files and is a lot less hassle than older centralized source control systems.

> Doesn't the uptake in Github which centralizes this distributed system kind of invalidate its main tenant?

No. Not at all. Github is actually a peer with an a few integrated extras for managing tickets and requesting your code be merged into github's repo branches. Most of the centralization is around access control and automation (i.e. continuous delivery, unit tests, etc...).

> Adding complexity without value is a mistake imo.

The main value of GitHub is reducing some operational complexity. For a small 1-4 person team, it may be of little value, but for larger teams, access control, issues, and automation can have a lot of value. For really small teams, fossil is actually quite nice. That said, there are a lot of great alternatives to GitHub that give you similar features.

> Maybe it's my own fault I haven't seen the value...

Probably not. I did the first 20 years of my career without source control and was able to build some pretty big applications. I do think dvcs was a big improvement, and I'm glad it exists now.

Re: Give up GitHub: The time has come

#159

Earlier quoted context omitted.

I should not have to opt out. GitHub should have to respect my license. I already said they can use my code, as long as they keep an attribution intact (via a BSD license, for example) GitHub is taking my code and ignoring the license. I don’t understand why anyone would think that is ok.

No, you're ignoring what you agreed to when you accepted the terms of service. GitHub can display your code, and YOU granted them that license by accepting their terms. I find the only people making these OSS claims haven't used copilot and tend to lack any real contributions to OSS. What you're describing is just simply not the case for 99.9 percent of the code snippets being produced/generated based on data from Gi…

Are you saying the FSF accepted github's terms of service when someone mirrored Emacs into github?

Re: Give up GitHub: The time has come

#160
post #51

GitHub is a business and currently provides free storage and a pretty nice interface to it. It’s easy to say “our rights are being stripped away” but the view that businesses should operate like non profits or government services with the common good in mind is ludicrous!

These are the immediate "products" GitHub provides, however, I would argue it provides a lot more. GitHub provides a place for people to easily collaborate on FOSS software. It has helped millions of people getting into software development or into their first FOSS project by lowering the barrier of entry significantly. Can you imagine how many people would start contributing to FOSS early in their career if they had…

You are right about the benefits GitHub provides to the FOSS community but I don't think that was ever the stated goal of this company. Their goal is to make profit, not foster FOSS, so we shouldn't be surprised when they make decisions that benefit their bottom line.
Post reply on HN