Email or username:

Password:

Forgot your password?
Top-level
Dana Fried

@williampietri even OSS projects have a central repo with the "canonical" source; I don't see the difference between hosting code on github or equivalent vs having it on a hosted SVN server.

13 comments
William Pietri

@tess Many OSS projects happen to have a central repo. But that's neither necessary or enforced as it is with closed-source software.

An example I was just looking at:

github.com/timothycrosley/stre

This relatively niche project has 194 known forks. One of them exists because somebody sees the original project as "no longer maintained", so they're picking up the ball and running:

github.com/streamdeck-linux-gu

Which is the true one? They both are.

Or we could look at another one I'm familiar with: github.com/CybersecurityForDem

They are researchers who were studying Facebook ads. They shared their code for collecting data and encourage other researchers to fork it and do their own thing. There is no real center, just shared history.

@tess Many OSS projects happen to have a central repo. But that's neither necessary or enforced as it is with closed-source software.

An example I was just looking at:

github.com/timothycrosley/stre

This relatively niche project has 194 known forks. One of them exists because somebody sees the original project as "no longer maintained", so they're picking up the ball and running:

Vrimj replied to William

@williampietri @tess

This reads to me like a reason git is terrible.

It is the same reason that the ECF, the federal courts efiling system is terrible.

It has burned the simplest problems on the alter of the most complex ones.

William Pietri replied to Vrimj

@Vrimj @tess

I agree git focused too much on the most complex uses. But I also think git could be way better for the common use cases while still supporting less common ones. And I think and approach of "who cares about common open-source use cases" is not going to lead to much improvement in open-source version-control systems.

Vrimj replied to William

@williampietri @tess

Is there anything that does this and is usable by people who are not?

William Pietri replied to Vrimj

@Vrimj Sorry, I don't understand the question.

Vrimj replied to William

@williampietri

Is there a package or distribution that does care about the niche needs of the FOSS community that is also routinely and happily used by people who are not members of that community?

Earnest question because I can't think of anything, even Linux repos mostly avoided until there were non FOSS-centric distributions and I remember how controversial that was back when

Virtue signal ๐Ÿ’‰๐Ÿ’‰๐Ÿ’‰๐Ÿ‡บ๐Ÿ‡ธ ๐Ÿ‡บ๐Ÿ‡ฆ replied to Vrimj

@Vrimj @williampietri

I've always thought Mercurial was nice whenever I encountered it. Haven't used it extensively, though.

AndreaTvilling๐Ÿ™ƒ๐Ÿณ๏ธโ€โšง๏ธ๐ŸŒˆ๐ŸŒป๐Ÿ‡บ๐Ÿ‡ฆ replied to Vrimj

@Vrimj @williampietri

Debian distro has few off-shoots, which are more niche, like edu, multimedia and geodata FOSS communities. AFAIK they have non-FOSS shadow sisters in corp/govt world.

meejah

@tess @williampietri Having done that, I can tell you that using Git as a Subversion frontend was way better than using svn as a Subversion front-end (and yes, I agree git is terrible .. but also "worse is better" etc)

Dana Fried replied to meejah

@meejah @williampietri "worse is better" is actually an argument not to use git tho? Like, in the original sense of the phrase?

meejah replied to Dana

@tess @williampietri Maybe?
I take at least part of the original argument to be about being "okay, but actually released" versus "perfect but not released", approximately?

(I'm not trying to defend git's UX here .. but also I've used RCS a little, CVS and Subversion a lot and it's 100x better than those, IMO)

meejah replied to meejah

@tess @williampietri I certainly think the history of Git's development -- or at least, my recollection of it -- matches the "worse is better" notions: it was "just better enough" than sending around literal tarballs that it gained traction.
Well, that and "official kernel approval" etc.
...and so even if better systems exist now, it's too late; Git got popular, then grew lots more features.

AndreaTvilling๐Ÿ™ƒ๐Ÿณ๏ธโ€โšง๏ธ๐ŸŒˆ๐ŸŒป๐Ÿ‡บ๐Ÿ‡ฆ

@tess

I followed more expanded on why git _looks_ wrong. And I have another person saying git was hard. Combined, I think I see why it _looks_ wrong. Would you be interested to discuss it in more safe way, without blowing flames and more like service design folks do?

Go Up