RasterWeb!

We continue to hack away at our aggregator, originally based on Feed on Feeds, it doesn’t quite resemble it anymore and my changes aren’t exactly neat little patches. (I’m considering calling it “thehomer” but we’ll see.)

So these are some of the features it has today, or I’ve been thinking about trying to add:

• Per-feed poll frequency, variable depending on the feed itself
• Per-feed purge times, some items I want to save forever, some I want just a few days
• Ratings, similar to iTunes, so if I was in a hurry I could just choose to read the feeds/items designated as 5 *’s
• Thumbs up/down rating system for items
• IM notifications, alert me via IM (Jabber or AIM perhaps) about certain events/things/etc…
• XFN, so I could do interesting things based off of the feeds/items from people I have some relationship with
• del.icio.us integration, currently I can add an item to del.icio.us with one click. There’s more to do here of course, like doing an md5 on the url, and hitting del.icio.us to see who else links to it…
• Technorati integration, currently I can do a Technorati Cosmos search with one click
• Google integration, currently I can do a Google “realted” search and find items Google thinks are related to an item. Eventually I’d like to do something a bit more advanced with the Google API)
• Click counts, which keep track of how many times I click on something leading to another site.
• Temporary subscriptions, so when you add a feed you can choose how long you’ll be subscribed to it. (Great for comment feeds!)
• blo.gs integration, I’m currently using Phil’s PHP Blogroll to pull in data, as well as providing a live link to blo.gs update list. (Still contemplating tighter integration here, possibly for marking things as read after visiting a site)

I’ve read through some of the ideas that l.m. had (see "Info Freako, or who’s already past arguing about syndication formats?") as well as stuff Rogers had (see "Wanted: Gluttonous RSS Feeders") and looked at the feature sets of a number of existing aggregators. There’s a lot of good ideas out there!

I’ve also grabbed SimpleAggregator, Andrew’s Feed Reader and Auto-Blogroll, and should read up on Temboz while I’m at it too!

I guess when it comes down to it, the aggregator is becoming a Total Information Portal, to some degree, and maybe Knowledge Management is more where things are heading. I’m really not sure, but I find it quite fascinating.

I’m finding the aggregator space more interesting nowadays than the weblog software space. Years ago a lot of people wrote their own weblog software, and people would come up with ideas, and others would implement the ideas in their home-brewed weblog software. There wasn’t really the commercial end of it that there is now. It was a feature-race, but with a friendly face attached to it. Hmmm, maybe it’s just me…

Anyway, the aggregator landscape seems to be changing on a weekly, if not daily, schedule. I’ve been using aggregators since 2000 or earlier. Not counting the Netscape and UserLand offerings (or my own aggregator named poorly enough “channels”) I’ve toyed with AmphetaDesk, NetNewsWire, Bloglines, a few others I’ve forgot the names of, and more recently Feed on Feeds.

As mentioned previously, I’ve been hacking at Feed on Feeds a bit, molding it to what I want for an aggregator. I’ve been taking ideas from the aggregators I’ve used in the past, and some ideas from other places (like AmphetaOutlines) and my own crazy ideas.

I’ve already made a bunch of UI and functional changes to Feed on Feeds, so that it’s not quite Feed on Feeds anymore, and really needs a new name.

I’ve added clickcounts, so it’ll keep track of how often I follow links to the original site. Of course some of the feeds are full entries, and some are just abstracts, still, this might show how often I follow a link to a site.

I’m working on creating a per-feed update schedule, since some feeds I want updated once an hour (favorite weblogs) and some I don’t care if they get updated once per day (like search results for keywords, top new stories, etc.) I could even allow some to update every X minutes (5, 10, whatever) if it’s coming from my own systems and frequent polling is an OK thing.

I’m interested in using the data from Technorati, Feedster, blo.gs, and del.icio.us in interesting ways. I might try to use Phil’s PHP blo.gs blogroll to see how I could tie blo.gs’s site update polling data into things. I’m more than happy to find other people’s code to glue together, as this has also become a PHP learning project for me.

I think what I’ve gotten out of this so far is that aggregators still have a long way to go, and that’s not to say they aren’t very good yet, that’s just to say that I see incredible possibilities in this space. Good Luck, aggregators of data!