On Thu, August 16, 2007 8:32 pm, Tracy R Reed wrote: > Steven E. Harris wrote: >> Andrew Lentvorski <[EMAIL PROTECTED]> writes: >>> Or use git, mercurial, darcs, perforce, etc. rather than subversion. >> I know you wrote "etc.", but add monotone¹ to the list. > > I have wondered the past few years: Why were we stuck with only CVS for > so long but then shortly after 2000 a whole flock of revision/version > control systems came along? Did some particular technology mature? Some > idea finally come of age? It somehow became fashionable to write one? > Did we all finally get fed up with CVS at the exact same time? >
If you followed it, disgust for the CVS core team was long standing and fully justified. They treated every suggestion for a feature upgrade with a contemptuous dismissal: "that's the way CVS works and it's just fine." The Subversion team was writing in direct opposition to this attute. Unfortunately, once they released svn, they adopted the same attitude for themselves. They rely on their (excellent) copy algorithm for everything -- branching, tagging -- and it's just not the same. They belittle every suggestion to add these features. As for the proliferation of choice, once the database model, as opposed to the file based model, occured to everybody, it was clear sailing for the development of several competitors at the same time. Commercially, Perforce seems to be the darling right now, and it's pretty good, but if you want to enforce process in p4, you have to script it in yourself, with all the headaches that come with that. More complete commercial tools like Dimensions and Clearcase have process built in, and that's great if it happens to coincide with _your_ processes, but plain hell if they don't. I suspect that we'll see another generation of tools, commercial and otherwise, that come with boilerplate process blocks that are independent and just plug in. I also thing SVN is headed for the ash heap of history precisely because they stonewall on making changes that would set them apart. As Osler pointed out (I'm paraphrasing) when there are a lot of medicines for the same disease, probably none of them is completely effective. -- Lan Barnes SCM Analyst Linux Guy Tcl/Tk Enthusiast Biodiesel Brewer -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
