Meh. I'm slowly moving away from GitHub myself but for none of the reasons mentioned in this article (which reads more like an ad, regardless of whether or not it is one).
Really missed an opportunity to discuss the censorship of both whole countries and individual speech, the merging into a huge conglomerate (Microsoft), the fact GitHub itself open sources an ironically small amount of its own code, etc.
I do like sr.ht but I find it to be a bit tricky to reason about. I'd like to see it evolve for sure. The developer is super friendly and responsive too.
I've been using a private instance of Gitea though. However, it's riddled with bugs :/ Oh well.
I use GitHub (and GitLab) because it is a convenient way of backing up random coding projects I have sitting around, not because I have any expectation it will be useful to someone. It is just versioned file storage at the end of the day.
I've only dabbled in their CI platforms, and as far as I recall, no one has ever made comments on any of my projects. I would be surprised if most people actually think that if they put something on GitHub that a task force of hundreds of dedicated open source volunteers will suddenly descend on the project and start contributing.
The benefit of people throwing their code up on github is that it now can also be searched through. So if there is a configuration variable or a function you are wondering about - you can now find what others are doing. (And https://grep.app/ has a fast UI for this)
Came here to say this. I mean, just trying to figure out how to use a particular Emacs function, I can search GitHub for that function and find hundreds of examples of people using it. So incredibly useful.
Can I just rant about how atrociously terrible and primitive GitHub's search features are?
It's quite common for me to search for something and get tens of thousands of identical results which GitHub forces me to scroll through one page at a time.
It would be so much better if GitHub just showed me one single result for all those thousand which are exactly the same, and let me expand that result if I wanted to see one of the other identical matches.
Also helpful would be a way to filter out filenames (or, even better patterns.. better still regexes) which I don't want.
That would make their search results actually useful.
As it is, unless you're searching for something exceptionally rare, they're next to useless.
What is this article? It seems like someone who doesn't fundamentally understand human behavior, or how businesses work. Please joining me in flagging it.
I dont understand the point of this CI argument. Is he arguing CI is bad in principle? For environmentalism reasons or something?
I dont use Github, but I still run CI for all my projects even if the build is a single line. Its a requirement. Its the only way to get past 'works on my machine'.
Honestly I'm not even sure what the author is suggesting here.
Yeah it seems like they seem to be opposed to CI for some reason.
I mean, I guess most people that use CI are glad they have it (because it fixes exactly that "works on my machine" problem that you mentioned), but who knows.
He's arguing that you only need CI after a lot of other things already went wrong.
If your code is portable, clean, readable, easily maintainable and contains tests, the benefits from CI over running the build locally are negligible. So CI distracts from fixing the issues that people face when building locally.
I don't have GitHub account (and friendly reminder that Github IS Microsoft!), but majority of my clients wants it so I make plenty of accounts for various projects, though 99% are private (not that will stop Microsoft to snoop around anyway)
bUt "Ci iS A HaRmFuL DiStRaCtIoN"; you'd expect that from someone that is really-really old-school or that has never worked in a project with more than a few hundred lines of code.
I have my stuff on github, and I'm very careful to make sure I'm not locked in there. Yes I do use issue tracking and the wiki, but my projects are relatively small and it would be trivial to write a scraper to get that out of there.
> A mostly-imaginary threshold. It's in your head. Think about it.. those with skills & time, who can wholesomely contribute, will highly likely to already have email and perhaps XMPP (Movim) or IRC. It is not the facilities stopping them, it's either skill or will to dedicate time.
I believe this is simply not true. Not every project is the Kernel and not every contributor is planning to become a long term contributor. Drive-by PRs are often of questionable quality, yes, but I'm not sure they're always worthless.
Where this really has value is the sort of project where someone writes something for themselves and then decides "might as well push it to github, maybe someone else can use it too". Then someone else comes along finds this and is like, oh cool, using this as a starting point saves me a few hours, and then decides "might as well open a PR in case they're interested in how i extended it".
These are the cases where I do believe that githubs low friction does make a difference. Not everything is a proper project.
> Where this really has value is the sort of project where someone writes something for themselves and then decides "might as well push it to github, maybe someone else can use it too".
Over the years I have been getting better and better results when looking for solutions when I add the 'github' keyword to the search. If your code isn't on Github and there is nothing really specific I can search for I will not find your code. This might be more the fault of search engines, but still. Sometimes 'gitlab' or even 'sourceforge' will result in something useful if it really isn't on Github.
I don't need GitHub, true, but I want it because I like it.
I don't care that no-one uses my projects: I use them myself and if someone wants to use it too that's awesome! The option is there, it's not like I check at the amount of stars of my projects everyday
>If one wishes to participate in a project, really, truly, wishing to participate, the buttons will not be a barrier.
Why do you want only the most motivated users? Users that casually file a bug report are valuable too.
I think the author tries to make the point that he works on such awesome projects that casual contributions are not necessary and that raising the barrier is not an issue.
You know what? Please also delete your `requirements.txt` and your `package.json`. Contributors who truly want to participate should figure your dependencies out!
It feels like every day people are complaining about more and more trifling things.
What is the point of this? If you don’t want to use GitHub because you feel it’s influencing the reasons you’re coding (e.g. caring about stars, etc.) then that feels like a very personal problem and not one that applies to the vast majority of people.
GitHub is a great place to version control your software. If you’re open source minded it’s a great place to share and find other software. If you want a lightweight CI, they have it, but nobody is making you to use it.
Everyone just feels the need to complain about the miserable state of affairs of everything. There are a lot of bad things in the world worth exploring, this isn’t one of them. Not to mention that GitHub has been one of the big benefactors of the software world.
This kind of anonymous, yet visible positive feedback is what keeps developers motivated.
We are not machines and having a simple yet effective way of expressing appreciation is a good thing.
Sure, it's less personal than say a postcard or an e-mail, but it's a token of appreciation nonetheless, a virtual pat on the back that even the most socially inept users know how to use.
I use some random piece of OSS software or a website and I notice an issue. I feel like I know how to fix it. If I can find their github, I go and offer a PR with a fix. If I can't find their github, or they don't have a github, I move on.
It's not a question of "one wishes to participate in a project, really, truly, wishing to participate" — I don't, it's just how many hoops I need to jump through to submit a 20 LOC patch.
39 comments
[ 3.1 ms ] story [ 56.9 ms ] threadReally missed an opportunity to discuss the censorship of both whole countries and individual speech, the merging into a huge conglomerate (Microsoft), the fact GitHub itself open sources an ironically small amount of its own code, etc.
I've been using a private instance of Gitea though. However, it's riddled with bugs :/ Oh well.
I understand why others would not though.
I've only dabbled in their CI platforms, and as far as I recall, no one has ever made comments on any of my projects. I would be surprised if most people actually think that if they put something on GitHub that a task force of hundreds of dedicated open source volunteers will suddenly descend on the project and start contributing.
It's quite common for me to search for something and get tens of thousands of identical results which GitHub forces me to scroll through one page at a time.
It would be so much better if GitHub just showed me one single result for all those thousand which are exactly the same, and let me expand that result if I wanted to see one of the other identical matches.
Also helpful would be a way to filter out filenames (or, even better patterns.. better still regexes) which I don't want.
That would make their search results actually useful.
As it is, unless you're searching for something exceptionally rare, they're next to useless.
I dont use Github, but I still run CI for all my projects even if the build is a single line. Its a requirement. Its the only way to get past 'works on my machine'.
Honestly I'm not even sure what the author is suggesting here.
I mean, I guess most people that use CI are glad they have it (because it fixes exactly that "works on my machine" problem that you mentioned), but who knows.
Did you miss the point about writing "good & efficient code"TM? Just write good code and you won't every need to think about CI again!
Or will you?
If your code is portable, clean, readable, easily maintainable and contains tests, the benefits from CI over running the build locally are negligible. So CI distracts from fixing the issues that people face when building locally.
https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks
My comment was aimed at:
> I forget to run tests locally all the time.
[1] https://github.com/typicode/husky
Yet how do you know for certain if your code is portable if you only build locally?
> A mostly-imaginary threshold. It's in your head. Think about it.. those with skills & time, who can wholesomely contribute, will highly likely to already have email and perhaps XMPP (Movim) or IRC. It is not the facilities stopping them, it's either skill or will to dedicate time.
I believe this is simply not true. Not every project is the Kernel and not every contributor is planning to become a long term contributor. Drive-by PRs are often of questionable quality, yes, but I'm not sure they're always worthless.
Where this really has value is the sort of project where someone writes something for themselves and then decides "might as well push it to github, maybe someone else can use it too". Then someone else comes along finds this and is like, oh cool, using this as a starting point saves me a few hours, and then decides "might as well open a PR in case they're interested in how i extended it".
These are the cases where I do believe that githubs low friction does make a difference. Not everything is a proper project.
Over the years I have been getting better and better results when looking for solutions when I add the 'github' keyword to the search. If your code isn't on Github and there is nothing really specific I can search for I will not find your code. This might be more the fault of search engines, but still. Sometimes 'gitlab' or even 'sourceforge' will result in something useful if it really isn't on Github.
I don't care that no-one uses my projects: I use them myself and if someone wants to use it too that's awesome! The option is there, it's not like I check at the amount of stars of my projects everyday
>If one wishes to participate in a project, really, truly, wishing to participate, the buttons will not be a barrier.
Why do you want only the most motivated users? Users that casually file a bug report are valuable too.
Get real, there are casual users too
You know what? Please also delete your `requirements.txt` and your `package.json`. Contributors who truly want to participate should figure your dependencies out!
What is the point of this? If you don’t want to use GitHub because you feel it’s influencing the reasons you’re coding (e.g. caring about stars, etc.) then that feels like a very personal problem and not one that applies to the vast majority of people.
GitHub is a great place to version control your software. If you’re open source minded it’s a great place to share and find other software. If you want a lightweight CI, they have it, but nobody is making you to use it.
Everyone just feels the need to complain about the miserable state of affairs of everything. There are a lot of bad things in the world worth exploring, this isn’t one of them. Not to mention that GitHub has been one of the big benefactors of the software world.
We really need someone to solve decentralized likes/reactions.
This kind of anonymous, yet visible positive feedback is what keeps developers motivated.
We are not machines and having a simple yet effective way of expressing appreciation is a good thing.
Sure, it's less personal than say a postcard or an e-mail, but it's a token of appreciation nonetheless, a virtual pat on the back that even the most socially inept users know how to use.
IMO, though, all decentralized solutions suffer the same barriers: network effects and easy of use.
It's not a question of "one wishes to participate in a project, really, truly, wishing to participate" — I don't, it's just how many hoops I need to jump through to submit a 20 LOC patch.
This article is dumb.