Recently, I had the opportunity to reinforce the need for solid sourcecode control within any software project to a friend who was looking for a new coder (PHP).
Previously, his sourcecode control policy centred around having good backups.
Needless to say, good backups are essential, but only the first and most minor step into sourcecode control. Solid revision control and management is essential to any modern software project.
So, I was heartened to hear that upon interviewing his candidate, he said that the first thing out of the applicants mouth was a question regarding sourcecode control.
Great says I, good sign….
‘Yeah, he has written his own sourcecode control system and wants to use that. Good huh?’
WHAT?! His own?
Well, since then I have had time to mull, to digest, and reflect.
Ok, a couple of questions…
1) Is this sourcecode control system freely available and well supported by anyone but this guy?
2) What happens when he walks away from the project?
3) Is it supported by other tools in the software development stack? Capistrano etc?
4) Whats wrong with subversion that this guy thinks he needs to spend time and effort inventing his own?
All of these questions paint a picture that some bespoke sourecode control system is probably a very bad idea. In fact, there would have to be MANY factors going the other way to balance out the negative aspects of this decision.