Nuts about Subversion!

OK, I know... geekfest! whatever... it's just fun.

Subversion SVN (Version control software) is making my life so much easier.

I use multiple CAD systems (pro/ENGINEER Wildfire (2, 3, and 4 - pre-production, which is another story I'll tell you about later... dang PTC, THANKS-A-LOT GRRRR) anywyay... but I also use Maya for rendering, and organic modelling of things that are, more difficult to model efficiently in pro/E.

Never mind the fact that I have to manage documents, and images, and Illustrator files... etc.

When I worked as an Industrial Designer, in a corporate setting, the engineers and I used Intralink. GREAT for proe, not so great for anything else. While it's possible to manage other types of documents in Intralink, it's REALLY set up to manage versioned pro/E files.

so... back to my excitement... While Subversion SVN is not as dialed in to manage proe files as Intralink, and doesn't have ANY of the renaming capabilities of Intralink, it DOES give me a way to manage a large set of project files in a controlled way.

For the pro/E files, what I do is, I'll be working along, saving etc. When I get to what I deem to be a checkpoint, I'll save, go into my current working directory and do a purge. This step clears me down to only tha latest version of the files I'm working on. However, since the checked out versions are older, some of them get purged from my working directory when I do this. So I just check in the few files that have changed, and then "update" my working directory. Now I've got a much more sparse view of the historical files, and much less data to backup when the project is done.

Another way I'm using Subversion SVN with pro/E is to create actual "tags" (which is, in software terms, like a version release, or even a minor or "dot" release) Since a tag is a "cheap" copy, or rather a link to the versioned file at a particular version, it doesn't take up a bunch of extra space but gives me a snapshot of the model at a moment in time. So I can check out, ONLY the files included in that snapshot.

How I create the snapshot:

Check out the working files, purge the directory (using pro/E's purge command from a cmd window, or DOS box) then the files remaning, as I mentioned before, are just the latest working files of the model I'm building. I'll select all these files, and branch/tag them as "version 1.0.1" or "as_sent_6-6-09" or some way that makes sense as a release schema. Then commit the workspace, and all the files I tagged are now considered part of that release, and I don't have to check out the entire trunk, or any extraneous files.

Using a release schema like this allows me to pull a model as it was at a particular date and time. If there's ever a question about a set of files that was sent somewhere on some date, I can pull THAT exact version, without having to save a zip or rar file with EVERYTHING each time I send a set. (imagine if you're changing a small part like a PCB and each time you send the assembly, you had to backup EVERYHING, including that 20 MB plastic shell that is final!... this way, it's the same file, and it's referenced instead of copied!) very clean, very nice!

anyway... digg diggin duggit!

I've also seen that people are using Subversion with Maya files, and I see no reason it couldn't be used with just about anything.

I must include a word of warning for those who have never used version control. Although version control is an amazing tool to keep organized, it's NOT without it's quirks, and it's POSSIBLE to corrupt your data. So Make backups, and read up on how to and NOT to use Subversion before you go cataloging all your projects into a repository. ;) Just make sure your butt is covered until you REALLY know what you're doing.

Good luck! cheers.

enjoy!