Hi all, Thanks for an interesting discussion ... and I had forgotten the paper by Parnas/Clements, and the analogy with writing a novel is an interesting one.
> As for the case of removing "illegal" insertions, I think it is far better > to have the real history saying "we had these from this date to that date, > as you can see, but you can also see that they were removed at a > particular time and not used thereafter". This follows the accounting > principle that incorrect entries have to be reversed out, not made > invisible. Without this it is not possible to demonstrate how some > situation arose from an action taken while the repository was incorrect. The comment about an accounting principle is interesting too ... I had not come across this before. My instinct is to side with Eric's view. Yes, mistakes will be made, yes, real life is messy but nevertheless that is what happened. So the real history needs to be recorded somewhere. However Ge Weijers has a point about making the maintenance or further development easier by not making all of a developers' mistakes or dead-end prototypes visible to everyone. Coming from a centralized SCM world I would be looking for a way to hide branches or 'obsolete' them so that they are still present and can be found if necessary, but are not visible to developers who are not interested in them. Tidying up takes time (equals costs money) so there must be some tangible benefit of doing it (probably easier development afterwards?). Of course if you are using this in your spare time then the choice is between tidying something that you just want to deliver or getting on with something new. I guess it comes down to why you need to retain the history. Who uses that history, and what do they use it for? So this comment by Gé grabbed my attention: > I work in a heavily regulated industry, and legal concerns have been keeping > us from implementing any and all DVCSes until now. > I wonder what has changed? Is it a greater understanding of some of the issues discussed in the mail? Dave _______________________________________________ fossil-users mailing list fossil-users@lists.fossil-scm.org http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
