| |||||||||||||
A proposed Subversion vision and roadmapA proposed Subversion vision and roadmapPosted Apr 3, 2010 18:58 UTC (Sat) by fhuberts (subscriber, #64683)In reply to: A proposed Subversion vision and roadmap by zorro Parent article: A proposed Subversion vision and roadmap
Well...
I switched our site of 50+ developers from TFS to ...... Git!
It has worked out for use extremely well, even the M$ oriented devs are very very pleased.
I'd say SVN has been overtaken by lessons learned for itself and others :-)
(Log in to post comments)
A proposed Subversion vision and roadmap Posted Apr 3, 2010 19:02 UTC (Sat) by fhuberts (subscriber, #64683) [Link]
BTW
TFS is barely better than SourceSafe, which was _the_ worst vcs out there.
A proposed Subversion vision and roadmap Posted Apr 4, 2010 11:23 UTC (Sun) by nix (subscriber, #2304) [Link]
Hah. My workplace is still using multigigabyte SCCS repositories. Beat
*that* for archaic.
A proposed Subversion vision and roadmap Posted Apr 4, 2010 11:54 UTC (Sun) by tialaramex (subscriber, #21167) [Link]
SCCS is archaic, but VSS is much /worse/. It famously loses data, for a start (hence "Visual Source Unsafe"), and is fiendishly complicated despite being less powerful than CVS. The problem apparently is that all the important Microsoft projects did not use VSS, so there was no-one competent eating the dogfood to complain about the taste. As with most really awful Microsoft products it was bought from a 3rd party to fill a perceived gap in the product line, rather than being developed in-house by smart people who understood version control.
Inadvertently Microsoft actually created a scenario in which Windows developers were less likely to even have source control, because they might try VSS figuring it's the "obvious" choice to go with their other Microsoft tools, and after losing a week's work in one click they give up altogether.
A proposed Subversion vision and roadmap Posted Apr 4, 2010 20:26 UTC (Sun) by nix (subscriber, #2304) [Link]
OK, colour me (un)impressed. Worse than SCCS. Worse, it sounds, than
*nothing*.
A proposed Subversion vision and roadmap Posted Apr 5, 2010 13:43 UTC (Mon) by sorpigal (subscriber, #36106) [Link]
This exact scenario happened to me, including the week of code and swearing off VSS forever. That was two years ago and every time since then when I've agitated for version control management has taken the "But we have VSS" viewpoint. Try explaining that not all version control systems are created equal to non-programmers, try explaining that the Microsoft product is not good enough...
A proposed Subversion vision and roadmap Posted Apr 6, 2010 8:07 UTC (Tue) by BlueLightning (subscriber, #38978) [Link]
Surely "even Microsoft themselves don't use it" would be a compelling enough argument?
CSSC Posted Apr 5, 2010 15:10 UTC (Mon) by cry_regarder (subscriber, #50545) [Link]
I still use CSSC (SCCS work-alike) for quick and dirty one-off tasks like writing a paper in .tex or a small program. It doesn't need a server and doesn't need configurin'. Just works and is so easy for the easy things :-)
It isn't in fedora but I got packages...
Cry
CSSC Posted Apr 5, 2010 16:32 UTC (Mon) by iabervon (subscriber, #722) [Link]
Somewhat surprisingly, git is actually really easy for the really easy things, particularly if you've configured it with your username already for other projects, and are only worried about making bad changes (rather than preserving the data from disks dying or an "rm -rf ."). You just need "git init" the first time and "git add .; git commit" whenever you want to keep a state. You only need a server (or even filesystem storage outside of the directory) if you want to publicize your projects, and you only get anything complicated going on if you have parallel development.
CSSC Posted Apr 5, 2010 21:26 UTC (Mon) by marcH (subscriber, #57642) [Link]
RCS is in Fedora and in any other distribution. There is nothing more convenient for single files (like for instance /etc files). Just type:
ci -t- -l /etc/hosts
and start editing.
CSSC Posted Apr 7, 2010 17:07 UTC (Wed) by intgr (subscriber, #39733) [Link] Having to learn yet another tool that works nothing like any other VCS is quite inconvenient.
CSSC Posted Apr 12, 2010 10:22 UTC (Mon) by marcH (subscriber, #57642) [Link]
There are only three commands to learn (check-in, check-out and rcsdiff) and they are just the very same operations that sit at the core of every other VC.
It is worth it is because RCS is available in every single Linux distribution: even the oldest ones, even the most stripped-down ones. (Remember that I was answering to someone carrying his CSSC packages around).
A proposed Subversion vision and roadmap Posted Apr 6, 2010 13:42 UTC (Tue) by leonid (guest, #4891) [Link]
> and anything that works for projects like the Linux kernel will work for _any_ corporate environment
Two things:
1. You are wrong.
Linux kernel environment consists largely of self-motivated, enthusiastic, competent developers. Corporates often hire the rest of the world.
A proposed Subversion vision and roadmap Posted Apr 6, 2010 17:24 UTC (Tue) by fhuberts (subscriber, #64683) [Link]
> Two things:
> > 1. You are wrong.
maybe, maybe not :-)
> 2. I am jealous of your corporate experience. :)
I was talking about the mechanics of version control, integrating code, etc.
Version control is never just about the tooling. It is so much more, like process, release management, etc.
Good soundbite Posted Apr 7, 2010 20:28 UTC (Wed) by man_ls (subscriber, #15091) [Link] I really like your last sentence. May I quote it? :D
A proposed Subversion vision and roadmap Posted Apr 6, 2010 18:08 UTC (Tue) by vonbrand (subscriber, #4458) [Link] Your comment on Linux development implies that git is somehow only for "genius developers". Yes, it used to be, but it got a major facelift in the 1.5 series, and is rather easy to use now. The current release series is 1.7, so...
Obligatory disclosure: I'm a former
A proposed Subversion vision and roadmap Posted Apr 7, 2010 10:57 UTC (Wed) by leonid (guest, #4891) [Link]
I've seen developers who needed about 6 month to grasp the concept of version control. And that is in 21 century with Google and everything. I suspect that a concept of distributed version control will cause their brain to explode. Seriously. :)
A proposed Subversion vision and roadmap Posted Apr 7, 2010 16:44 UTC (Wed) by nye (guest, #51576) [Link]
Do you happen to understand why that is? The concept of version control seems as obvious to me as the sun, and I can't really imagine it being any other way.
Or do you really mean just that these developers were reticent about using the tools available to them? That makes more sense - laziness is something I can identify with.
A proposed Subversion vision and roadmap Posted Apr 7, 2010 16:51 UTC (Wed) by fhuberts (subscriber, #64683) [Link]
I can answer that for you:
mostly it is because of the strong drive of people _against_ change. most people are not open to the idea that the way they are doing things is no longer the preferred/most optimal way of doing things.
A proposed Subversion vision and roadmap Posted Apr 12, 2010 14:02 UTC (Mon) by nye (guest, #51576) [Link]
Sure, but that's different to being unable to grasp the *concept*, which is what the OP said. I'm guessing they probably didn't mean it literally.
A proposed Subversion vision and roadmap Posted Apr 7, 2010 20:16 UTC (Wed) by vonbrand (subscriber, #4458) [Link] Every one I've met did "version control" of sorts (keep directories with weekly snapshots, whatever) before they were introduced to VCS. The idea is easy to grasp, the how is quite a bit harder to wrap the mind around... and then you have to learn when to commit a change. For me, DVCS was (almost) natural when I met BitKeeper for the first time (after using RCS for personal projects, and never getting around to get fond enough of CVS/SVN to go to the rigmarole of installing and using them in earnest). I've seen no cerebral explosions...
|
|||||||||||||