I don't know if this can really be considered a programming problem, but I was wondering how you organize your source trees.

Right now, my entire source directory is a mess. At one point, I had it split into categories, like games, libraries, utilities. Then, it seems I decided to split it with languages like C#, php, C++. Throw in a few 10 rogue directories for quick tests and demos and I am left with a huge mess of spaghetti with no real idea of how to logically organize it or how to find anything inside it. Couple that with the fact that much of my code is duplicated several times in multiple locations with slight variations to each version.

So I'm looking for ideas before I take on the chore of cleaning it up. My primary concern is that some code can be categorized into more than one folder. Perhaps I should just do one root folder with everything and create a .NET or php program for browsing them by keywords.

I know the feeling. My dev folder is huge and has a complete mess of projects, tests, restarted projects, etc, etc.

I file by project name, the language is irrelevant. I also have a _lib folder (for my code that is re-used in many projects) and a _completed folder (things that are finshed and will not change again). The underscores put them at the top of the file list to make finding them easy.

One thing that does help is to load up rarely used VC++ projects and hit the 'Clean' command (you'll need to do it for both Debug and Release builds). That strips all the intermediate junk out of the folders and reduces the storage size of the project by quite a bit. The stuff that is removed will be built again automagically next time you compile the project.

I am toying with the idea of using Git or some other CVS to try and get all my source under control, but I think the dev folder will always be there 'just in case'.

_________________
10 PRINT "Bad Monkey ";
20 GOTO 10

I use SVN for some projects, but I still hit a wall with the organization scheme. SVN versions commits and not files, so I would have to do each project under its own trunk. It could be nice to have a working source tree for whichever project I am currently on, then just delete the local copy once it is complete or shelved.

Suppose all it would take then is an indexing app to grab/parse the first comment block from the source files - that's usually where I list all of the descriptions, bugs, and todo's for each module and project.

Also, most of my code is doxygen formatted. Does that support dynamic documentation or would I need to manually create each document tree for each project?