Last summer I decided to teach myself about relational databases, so I read Kevin's books on SQL and PHP and created a little catalog of my CD collection. I spent a little bit of time creating the database structure, a longer bit of time coding for it, and eons entering data. And it was spiffy, and it was completely useless, because there was absolutely no motive for anyone besides me to use this thing. (And it is also now gone, because of the server's HDD failing, but that's all right.)
Other than being useless, it was actually a nice little database, especially for a first try. Some of the database structure I had has been replicated in our group project for LIS 540 this quarter, but I have also become much more aware of the problems of my old design. Not only did the database structure require me to write all sorts of unwieldy code, but the interface was awful. I realized it wasn't great, but I had no clue about how to improve it.
Some of my interface problem was lack of experience with user interface design, but some of it was also a complete lack of knowledge about information retrieval. It just hit me a couple of weeks ago that what I made this summer was a data retrieval system, not an information retrieval system. The distinction is kind of fine but important. It's the difference between using Perl regular expressions to match values in a certain field and using Boolean logic to combine result sets. I still don't know how to tie in an information retrieval system to a database, but I think we're going to cover that next week, and I'm really excited about it.
I'm planning on recreating my CD database or maybe even expanding the system to include books and other items. You see, I have an Amazon wishlist, because it's fairly easy, but I would rather have a web-accessible wishlist that doesn't favor any specific vendor and which has a lot more features than theirs. For example, I would like to be able to group books by series and to add generic items like socks
without specifying any particular quantity. The vague idea is to have a kind of mega-system that lets me add items to my wishlist and then, when someone buys them, adds them to my catalog of CDs or books or whatever. Actually, I'd like my catalog
to include books that I have read or want to read, even if I don't own them, because that's also a guide to what people might want to get me.
This 540 project is both wonderfully exciting and horribly frustrating, but the process of writing a specification for an information system is proving to be an important learning experience for me. Not only am I getting a taste of what it's like to work on a pseudo-professional team project, I'm realizing the importance of recognizing and modeling user needs before designing a system and designing before implementing anything. This is why I've been thinking about my New Improved Wishlist/Catalog for six weeks and haven't done anything on it yet. But since I had a Great Idea this morning, it's getting more tempting. I hope I can wait until the quarter is over, at least!
Post a comment