PatCavit.com

I end up writing a lot of code to avoid having to do in-depth searching for an already-built solution. I just wrote a basic cookied user authentication system for FeedFilter (and the rest of Tivac.com, I guess) because I didn’t want to go trawling through the MOUNTAINS of PHP code online that do just that.

I think in the end I made the right decision, as it was a pretty small problem and solving it myself means that I know the code inside and out and it does exactly what I need it to. There’s no extraneous code involved, although I will fess up to commenting more than usual. The reason for adding a user authentication system to FeedFilter is pretty obvious. When I first psoted it onto my blog someone went in and deleted my two filters to add some of their own. While I’m all for sharing information and letting other people see what I’m filtering I was a little put out by that. I was using those filters!

So all filters are now tied to a user, and the feeds they output are also uniquely indentified per user. So two users could name their filters the same thing and it wouldn’t be a problem. I’m still not entirely happy about the output arrangement I have. Currently all filter rss feeds are output to one folder that allows browsing. While I’m not particularly paranoid about this, someone else who may use this service (Unlikely though that may be) might not like that so much. Maybe I should do some sort of Google Calendar-esque solution. Maybe later. For right now there’s all of two people using this so unless it bothers one of us it isn’t likely to be changed.

My internship this summer is a paid one, and while I’m making sure to save most of it having money brings along a certain temptation to spend it. To facilitate that I’ve been collecting RSS feeds of sites like fatwallet, techbargains, and the SA Coupons forum so that I can monitor them for good deals on the stuff I want. What I didn’t think about was just how much movement there is on these feeds. I set up a folder in Bloglines for all my “shopping” feeds and it quickly galloped to 500 unread items. I love reading feeds in Bloglines, but I don’t have time to even scan through that much just to find the one or two items I want!

So I’m going to do something about it. There’s a high probability something like this exists already, but it’s a fun side project to keep me from getting too rusty. I don’t get much DB interaction at the internship so doing some design and implementation work to stay sharp is a good thing. I’m working on a flexible RSS filtering system. You tell it the feeds you want it to watch, what you want it to watch for, and what you want to call it. Using a shared RSS DB/updater (I have a few other projects that deal with RSS I’m kicking around) this system will scan the feeds you told it to and build a feed of items that match the filter(s) you’ve set up for them. This output can them be consumed in your feed reader of choice, without turning a firehose of coupons and deals onto your face. Thinking about it, I actually already know something like this exists. Since I have a few issues with the implementation, I’m going to create my own less flexible version! The danger of stubborn people that can bang on a keyboard and have it do things…

If anyone’s interested in this I’ll let them in on the URL once I have something working. Right now it’s a DB schema and some forms that aren’t hooked up to the backend in any meaningful way. I had to get the central RSS DB up and running first though, and now that it’s running and updating all it’s feeds via Magpie I can continue with the filtering bit. Centralizing the RSS updating/storage is a definite Good Thing, as it means I only need one feed parser setup and won’t be grabbing duplicated feeds. Definitely gonna toss some YUI libraries into the mix for some oh-so-delicious JS, and I’ll probably try working on this using Aptana because it is shiny. SFTPDrive rules for this kind of thing by the way.