I'm a SubVersion geek, why I should consider or not consider Mercurial or Git or any other DRCS?
- by Pierre 303
I tried to understand the benefits of DRCS. I must recognize I still doesn't get it.
Here are my current beliefs. I'm ready to destroy them thanks to your expertise. I know I'm probably resisting to change. I just want to evaluate how much that change will cost me.
Merging hell can be solved by just applying good practices such as continuous integration.
There is no such good practice than having a private branch for a few days when you are in a self managing team with real collaboration. I use branching for that for very rare cases, and I keep a branch for every major version, in which I fix bugs merged from the trunk.
I see the value of committing offline then pushing online. But continuous integration can help on this too.
I work on very large projects, and I never noticed SubVersion to be slow even when the server is 5000km away on the internet and my small connection (less than 1024D/128U).
Harddisk space is cheap, so having a copy of source code locally doesn't look like a problem to me.
I already have a full copy of the last version on my disk. I don't understand the distributed thing there (maybe THIS IS the key to my understanding?)
I not new in the industry, and judging by my difficulty to understand, I don't think DRCS are easier to understand than SubVersion like. If fact, I don't understand...
Doctor, give me your diagnostic.