19 comments

[ 3.2 ms ] story [ 53.3 ms ] thread
"... Numerous people have written diff and merge systems for wikis; TWiki even uses RCS. If they used git instead, the repository would be tiny, and you could make a personal copy of the entire wiki to take on the plane with you, then sync your changes back when you're done. ..."

Now that is a compelling idea. How portable is git? ~ en.wikipedia.org/wiki/Git_(software)

My friend looked into porting git to Win32 natively and abandoned the effort because Git uses a ton of UNIX-isms, from fork() to sh scripts.
Not really surprising, given the author...
I read that git was written to intentionally use every trick available in Linux for performance. Mercurial has similar features, but it's much more agreeable to cross-platform work. Mozilla uses it.
check out ikiwiki - http://ikiwiki.info/ - it backends to both git (and svn) to do this - git allows offline work.
"... check out ikiwiki ..."

Did, pretty smart & is what I was interested in. Do you know of any python language bindings? Doesn't seem to be any around, at least as far as I can see.

It's also the next Unix in that it's kind of hard to use for newbies, and shits all over your data when you do something that seems like a routine action (git revert anyone?), but in fact will tell it to self-destruct. svn is emphatically not this way, which, when you aren't doing anything distributedly and don't have very many branches, is happy.
Agreed. SVN is much more user friendly than GIT (command line interface).

The major reason I haven't made the switch from CVS/SVN yet (and am unlikely to).

I'm struggling with this though I can tell it will be well worth it.

Try the peepcode git tutorial, which includes info on migrating from svn.

Git is good at what it was designed to do (manage and distribute changes to source code) but that's not what most people need. Non-programmers are a lot better off with simpler things like OSX Time Machine and iDisk.

That said there seems to be a lot of effort duplicated at the file-system level and at the version-control level. I hope file systems of the future will expose nicer api's (e.g. change logging) to help version-control systems become simpler and more efficient.

One could also develop a separate kernel API for version control. Filesystems without version control would just have null operations in this API, while filesystems with version control would root around in their database for the change information.
If this is true, then webapps should use git instead of databases.
For sure in some cases - yet I don't see any metadata and query capabilities. Are there?
"git is so powerful, that when you put light in it, it can't escape"
If git has five files and you have five files then git has more files than you.

Git can kill two stones with one bird.

This guy is mistaken. It's not good enough to track large files. So I think it's very unlikely that it's what one needs to deal with all of our data needs. It works for small files and big trees of source, but personal data is not like that. The trees are shallow and the files huge. How many people here like to keep their home directory in source control system.
Call me skeptical, but the author seems to do a pretty good job describing the features with existing features in Unix.