Monday, January 26, 2009

DSpace, FreeBSD, and Gentoo

My name is Ana, and I work at the Southwest Collection at Texas Tech University. I was hired into my position six months ago, and my role is to build and test digitization software and servers for archival digitization purposes.

My first open source endeavor was with FreeBSD, back in 2005. I determined to learn only the command-line interface for the simple reason that the command-line interface provides a very thorough understanding of how an open source, unix-based operating system functions. As I learned more about FreeBSD, I became enamored with its flexibility, the ease of control, and very strong security features.

I built my first production FreeBSD server to serve Plone and Zope for the Landscape Architecture program at Texas Tech University. Zope itself was a little clumsy, but working with the Plone CMS proved extremely easy, and I was able to give professors and student organizations access to build quickly their own classroom and organizational content management systems without deterioration of skill over time. (You can view one example of a departmental announcement CMS served from this box here: http://cms.larc.ttu.edu/TTULarc .)

I am now testing with a product called DSpace, an open source digitization management system with an extremely well-designed search feature. (http://www.dspace.org). I'm impressed with how easy Dspace is to operate, although with my own minimal database experience, I'm going to have to figure out how to replace and manage databases in Postgresql. This uses Apache Tomcat, which is very similar to the regular Apache web server but incorporates a few more utilities. I tried to use DSpace, Tomcat, and Postgresql on FreeBSD, but I couldn't get it working. Not only could I not get it working, I also couldn't locate one thing about Dspace and FreeBSD. (Give it a shot. Google the following, exactly as I type it here: "Dspace"+"FreeBSD"+install) If you look at the results, although there are a ton of websites that appear, you'll see that although the sites look promising, they actually only offer more problems than they solve. For instance there is a very helpful Japanese page that comes up second or third on the list, but after using Google to translate the page, I quickly discovered that the translation wasn't very accurate. I think this page could be very helpful, but I am still not sure.

So, since I could find little help on installing DSpace on FreeBSD (that dspace installation diary for freebsd is the biggest letdown), I determined to try another open source OS, but I decided to try a Linux brew.

The Linux brew I went with was Gentoo, seems like a solid, stable OS that is as flexible as FreeBSD, but with a less elitist group of developers. (I won't go further with this. I love FreeBSD. I hated moving to Linux myself!)

Either way, I found a really well-written Dspace install how-to for Gentoo. I tried to use DSpace 1.5 (the newest version) with these instructions, but it just didn't work. The DSpace installation instructions themselves are fairly shoddy, and I couldn't waste any more time trying to work around installing version 1.5. Instead, I went ahead and moved a notch down to the older DSpace 1.4.2. I'd rather use the newer relase, but I figured as long as I got it put together quickly, I'd be willing to work the kinks out later.

Anyway, here is the link to that website:
http://sentra.ischool.utexas.edu/~i312co/dspace/

I keep meaning to track this guy down and thank him. He did a great job, really helpful. I had to modify the instructions a little bit because I used a newer version of Postgresql and of Tomcat (Tomcat-6), but it worked out fine. I will soon try to update to 1.5, just to find out how complicated the update process is.

The more I think about it, the more it nags me that there is such poor material out there about installing Dspace on a FreeBSD box. I think I'll go ahead and try again on my FreeBSD box, perhaps at my house instead of at work so that I'm not wasting work time doing it, and write up some instructions. This way, there will at least be something out there. (I even tried sending a question out to the FreeBSD mailing list, but there were no responses.) I guess both Dspace and FreeBSD are each individually mildly obscure, and the two combined is simply an aberration. Ha, world, no longer!

The other thing that bothers me is how shoddy the DSpace installation instructions are. The documention goes on for thirty pages (or so, maybe less) about how to create metadata fields for tagging, but there are three pages for installation, and the product will not work when following those instructions. Granted, I've always been better at learning by doing rather than learning from a set of instructions. I know my FreeBSD install experience will improve now that I actually got Dspace working with Gentoo, so perhaps soon I can practice enough to where I can help write better documentation for the Dspace project itself.

That's what the open source community is all about: contributing what you can, when you can, in hopes of helping someone else who might want to someday contribute.

That's all for now. Tune in later--at the same bat time, same bat place!