Damn, progress on my site has nearly come to a complete halt. I'm pretty tired from work at the moment, so week nights are a write-off, and I like to get out a bit at the weekend. Basketball and my significant other in particular.

Having said that, I've removed the tables I was using to layout content on my site, switching over to CSS completely. There were a few things that weren't totally intuitive, but generally, it wasn't too bad. Mind you, it probably looks shit in NN4! :-)

The Perl scripts are pulling in all the data I want regularly, and without glitches, although I still need to refine the scoring script for the predictions game. Anyway, I'm tired, so it's bedtime for me.

The OO stuff in PHP has gone really well. I've been able to use some template files as the basis for my site, and then created and extended various page object that manipulate these templates to generate the appropriate type of dynamic page. It's made the code considerable more re-usable, and reduced the amount of kludges I've used (which is normally pretty significant!)

The second stage of my development is kicking in. I need to capture a wider range of data from the web now, for some added functionality on the site. Re-visiting my scheduled Perl scripts, they're pretty messy, and not terribly efficient either. To be honest, they smack of amateurish code. I haven't used functions that could have short-cutted a lot of the tedious code, and I've rolled my own functions in circumstances where it hasn't been necessary.

My dilemma now is whether I spend the time to re-write these scripts (for no visible difference, since they work right now), or do I just hack in the additional stuff (which I built last night), and then get on with the web changes, which are considerably more visible? I'm not sure which is going to be the most satisfying...

Just a quick note. I'm using the opportunity to re-design the functional structure of my site, whilst I do the cosmetic overhaul. I'm trying to use the (admittedly limited) object-oriented functions in PHP to create the site. Transitioning to servlets/JSP should be easier then, since I'll have a basic pardigm for the site all sorted.

That's about all for this small update. More soon!

Well, I've decided to use my journal at slashdot to talk to myself about some of the programming stuff I'm doing (in my free time). I'm going to try to be good about making entries in my blog too, but not much of interest happens to me :-(

The thing that bugs me most about my website is graphic design. I don't have an artistic bone in my body, so every page design I come up with is really, really derivative. Not necessarily a terrible thing, but it bugs me nonetheless.

I've gone for XHTML 1.0 Transitional (standards-compliant, but I'm not evangelical enough to go for "Strict"), with as much CSS for the layout as possible. I've also decided not to cater for people with really old, non-compliant browsers (eg NN 4). Since it's up to me, and it isn't a job, I can take those kind of decisions! I'm redoing the design (again!), and I'm happier it's looking a lot better than previously. I need more pictures/photos, but they're not easy to source from the web. Hopefully the new arrival of my digital camera will help out here.

Most of my attention recently is on the dynamic parts of my site, however. I'm constantly in a state of flux as to what language to use. Currently it's PHP, but I'm swaying towards servlets/JSP (as a good learning experience), with some tendencies towards ASP too (is it a more marketable skill? I'm sure switching from PHP can't be that hard, and I've already got some VB experience). Perhaps I'll continue with PHP, and attempt to build versions in the other languages as I go along, in parallel. Of course, I could always try to use mod_perl too... decisions, decisions :-)

The automation thing is nagging away at me too. My cronjob'd perl scripts are working nicely, to retrieve the Chronicle headlines, and the box scores, but it all feels like a bit of a hack. Ideally, I'd parse the Chronicle headlines into an RSS file at the same time as storing them in the database, and I want to create an XML file format for the box scores, rather than using CSV as an intermediate stage. In fact, I'd rather not have an intermediate stage at all. In the long term, I'd like to make box scores available in XML format to anyone who's interested (not too many people, I'm sure).

Perhaps Java is the way to go for my cronjob stuff, if I'm going to be creating XML files. Not too sure yet. At the very least, I'd like the Perl scripts that do the work at the moment to use HTML::Parser. All the regexps I'm using currently seem a bit, well, flaky. It's not that they have failed (in fact they've worked perfectly, the only problems have been inaccuracies in the source data) but they are very vulnerable to changes in the formatting of the source HTML.

And finally, I need to revise the scoring script for the predictions game from the ClutchCity BBS. Since the scoring system is a *little* complex in certain circumstances, there have been three games so far when I've had to manually correct the scores. It's not a difficult job, it just going to be a bit tedious, for not too much reward. I'll get around to it eventually.

That about wraps things up for now, I'm sure I'll post again soon.