Linus Torvalds won't do github pull requests (opens in new tab)

That is the spirit!

I remembered he once said talking angrily is to scare idiots off.

What a man.

Well...

"github is a total ghetto of crap"

"make a real pull request, not the braindamaged crap that github does"

Either way, there are many decisions that people make that comes down to style preference and sure the 'purists' might think it is a watering down of programming....but it makes it more appealing to new developers. I can see Torvalds ranting on Ruby as not a 'real' language....I don't know if he has done so...but I wouldn't be surprised if he did.

With a bigger microphone comes greater responsibility.

I am not taking Github's side on this particular issue - but this issue struck a bigger tone (for me)...where I have seen those with a larger microphone (by virtue of who they are) seem to be complaining louder (publicly) in recent history. It could just be that their microphone has gotten louder - but if it has, to a large extent they have a responsibility to be more conscious.

Imagine if Obama complained about the food at a bakery down the street that he went to - that would effectively kill that establishment. As a patron, does he have a right to complain about disgusting food? Absolutely...as someone with the largest pulpit in the world? Definitely not. There are other means he could use to get his message across.

Then again, I am not comparing Obama's pulpit to that of Torvalds, but I think it makes the point sufficiently.

I guess it was off-putting to me because of the ephemeral effects of him being who he is. Anybody could (and probably would) complain about the way Github has implemented various features - but they wouldn't get the attention that he has gotten. Maybe that's why I perceive it as him being 'cranky'.

So maybe my criticism was unfair because I wouldn't have criticized DHH for doing the same thing - or who knows, maybe I would.

Not sure, but now I see what you mean.

How many projects have had as many people involved as the kernel? Not just software, any project. With people coming and going and working on whatever they want. For 20 years.

People have the right to be offended, they however DON'T have the right to not be offended.

Linus is a nice guy, he works well with others and has built some of the most amazing software in the world today. I work on the kernel and other system code on a day to day basis and have participated in threads with Linus before. When you are wrong he tells you so, in a very clear manner that outlines exactly WHY you are wrong.

Sure, he isn't always concise and succinct but that isn't really the point. If he wants to interject a few choice "swear words" to describe how badly you did something that's fine by me. As long has he describes exactly what is wrong and how to fix it.

People need to get a lot less wrapped up in HOW something is said and focus on WHAT is said. The essence of Linus's comments on this pull request can be summarized easily. GitHub's pull/request features are inadequate for kernel goverernance - and as the creator of git and designer of the inbuilt pull-request functionality he thinks this is subjectively retarded.

If you read past any of that and start saying crap like "he could have been more polite" to him then you have obviously missed the point. It's not his (or anyones) job to cater to peoples inability to grow a skin.

And the side effect is that he gets bored to listen at stuff that he thinks are silly. But see it from a different perspective, Linus is really, culturally an hacker, and not an academia professor. He is out there on github writing flames, arguing with people that have a github account completely clean and never wrote possibly a line of code at all. Exposing himself in the process to insults and so forth.

I really admire him.

And a few messages down, he proceeds to say "You are a moron".

Not exactly a role model of civility.

Somehow, I don't know or have read about that many (if any) world-changing people, who also happen to be "so nice and understanding of others opinions" or "agrees to disagree".

I'm sorry, but there are indeed few ways to design a BMW and many ways to design a glob of metal that can't even move.

In response to the pull request he did simply say "I don't do github pull requests." and explained why, and basically said he would if github supported commit messages in the appropriate style. He never insulted the pull requester. He called whoever is this "Joseph" person a moron, and now we have no idea why.

Actually, I think Linus was factually correct. The persons involved were morons. :-)

Thank you for epitomizing what makes Linus worth emulating.

Rhetorical force. Coooool.

     Git is a notoriously user-hostile piece
     of software interface-wise

Also, its interface is hostile because it requires the user to understand a little about its internals. You may disagree that this is good design, however it gives you unprecedented control over the repository, which you need often when the shit hits the fan. Also, Git is freakishly fast and efficient. Much like C, an interface done with taste, but that doesn't appeal to those with weak hearts. And the fact that such projects as Linux are maintained with Git, projects which have a huge number of contributers, it's a true testament to Git's awesomeness.

And btw, considering how version control is one of the most important tools at your disposal, if not the most important in big teams, stop reading those SVN 2 Git tutorials and go read a freaking manual. It's worth it, because you'd be amazed at how much power it has under the hood.

Probably true, but when you've changed the technology world twice before age 40 people tend to cut you some additional slack.

Then you probably didn't use any SCCS before git. Because CVS is even uglier, Bitkeeper is the reason Linus wrote git and every other I tried was more "user-hostile" than git.

> pretty much the only thing it has going for it on the usability front is Github.

You sound like a Visual Basic programmer who just saw C for the first time.

Edit: Bitkeeper, not bucket!

That said I don't think this is significant enough to add up to a compelling reason either way.

Projects like Git and Linux take their contributions in as patches that are sent to the projects mailing list. As it is not merged in to the master tree at the time of publishing, it would not make sense to use past tense for describing what it should do, if merged.

Thanks to Git's great history-graph altering functions like rebase and cherry-picking, it would make even less sense. You can always take some commit and apply it to new branches or even repositories.

It drives me nuts that I never know exactly when Github will break my one-line descriptions when there appears to be plenty of room left to display the characters. Like this one from Linus:

https://github.com/torvalds/subsurface/commit/bea6637c03a90f...

It's also simpler and promotes less convoluted phrasing.

It is direct, concise, consistent and tends to active voice rather than passive voice. Further, not everyone writing or reading the messages will be a native English speaker. Avoiding past-tense forms just makes it easier.

http://git.kernel.org/?p=git/git.git;a=blob;f=Documentation/...

Linus using git for the linux kernel, the bitkeeper drama, and doing talks about why git is better and everything is broken made git popular.

Github has done a lot to push it, but even without github, git would still be huge.

Linus' use case for a pull request would probably be to much for a popularized and casual use case that is likely typical for GitHub.

And anyway it's Linus' prerogative to dictate his requirements for contributing to his project. The description he gives doesn't make it seem overly difficult (use what's already built-in).

Github wouldn't exist without Linus' git. I am really surprised they didn't take Linus' advice (reasonable, and no difficulty at all) in the first place.

pirtlj "I did not realizes that Linus' shit does not stink." torvalds "you are a moron"

Seems fair to me.

I note that GitHub actually did a blog entry sometime ago about good commit practice, and it seems to be the same as what Torvald's requires. Not sure I'm understanding what th issue is... https://github.com/blog/926-shiny-new-commit-styles

It's not really. If any other maintainer of a project finds they're cool with pull requests from github, then it's a fine tool for those projects. This guy runs a project which is an outlier in it's size, and which is also a far older project than most of what github hosts.

Now if his criticism has useful facts, that MIGHT be useful, and it does: The github format has fewer bits of info. Could some of those bits of info be important? Maybe. I didn't sense a genuine, measured analysis, but instead a rant from a maintainer about something that's not working for his project. Nor did I see a counterpoint from github why their tool does it the way it does. There may be real design decisions there, ones that might be better answers than what Torvalds has, especially for projects far smaller than Torvalds.

Don't let Torvald's status as the maintainer of the Linux kernel (and creator of git, to do that task) make you forget he's also paid attention to because he's famous as well as a useful contributor. Fame doesn't make you more right, just more adored. His ideas still require critical analysis from the rest of us.

Github is not Git. It's a separate product which happens to support Git, but is not Git, strictly speaking. It's basically a social network for building software. Far removed from the revision control software it depends on.

Complaining that the commit log isn't formatted in his preferred manner is similar to an engine designer complaining that the car's ODB2 reports aren't easy enough to read on his hand-held scanner. The engine designer can dislike it all he wants, but if I was the engineering team who designed the whole car, I wouldn't lose sleep over his dislikes.