A Serious Question for Feedster Users ... What Is Current ???

Now here's a serious question for people who like Feedster  --  what is current?  Right now Feedster has been running for between 2 and 3 weeks so far and searches to this point are done against "everything".  All along its been my intention segment the search scope into Everything and Current where current is defined as "if it isn't in the user's RSS feed then it isn't current" and then initially search Current letting the user expand their search to Everything.  This will nicely avoid the many issues that arise when you're searching a massive archive.  And people seem to really love the aspect of Feedster's currency.  I get that.

So its now time for me to start defining "Current" and my initial definition of "Current" just died a gory, bloody, fetid death.  Why?  Well in the very back of what I call a brain, I remembered a vague little whisper that said "mark only syndicates one day".  The Mark in question is, of course, Mark Pilgrim of Dive Into Mark fame who keeps only a single day's history in his RSS file.

The way I think of "Current" is that its what I can basically keep in my head at one time.  In my brain I sometimes think of Feedster as the "where did I see that search engine" since this encompasses the currency in Feedster's approach.  That's different from the Google approach of "where can I find this".  Or the Wayback Machine's "did this ever exist" approach.

Some options are:

1. To heck with Mark!  Let current be defined as if its in their RSS file.
2. Ah ... No.  Mark's important and his reason for putting only a single day into his RSS is valid (bandwidth).  Use a two week span as current where two weeks is calculated as backwards from today.
3. Ah ... No.  Mark's important but current is greater than 2 weeks.  Make it the past 30 days.
4. Give user's infinite flexibility in scoping date queries ad nauseum.  Definitely possible but leads to huge UI issues or a much better query language parser than is current written.
5. Do nothing -- leave it as searching all of indexed blog space and let the user choose to sort it how they want.

Thoughts?  I could really use some input here.  On this one I don't want to do the wrong thing.  And yeah I could make it configurable but an intelligent default has to be set and I'm not sure what that's supposed to be.  Am I wrong?  When you're using Feedster, what do you think?  Is the right answer #5 ?

Note: When this is turned on, it will be disclosed since it makes a big difference.

1:11:12 PM      Google It!   comment []    IM Me About This   

Miscellaneous Roundup

Its now been about three years since I was responsible for running a 24x7 web service (now Feedster, then the late, unlamented Mascot Network) and I'd forgotten just how **draining** it is.  Now, mind you, I'm not bitching or whining at all.  Well perhaps a bit but I'm utterly thrilled by how well Feedster is doing (we crossed the 200,000 indexed posts benchmark last night around 3 am or so). 

So with that said, since I just plain don't have the energy to write original content, here are a serious of miscellaneous tidbits that caught my eye today.

• Chris shows how to setup Feedster with Internet Explorer so you can just type RSS: query and have it search Feedster.  Andy points out how Mozilla is easier in this way.
• Via Doc -- Cory finds Danger's attitude towards their customer's contemptous.  Ouch!  But I'd agree based on what he wrote.
• Kasia is taking Google into account when restructuring her blog's archives.  Her rationale is very similar to what Feedster does -- look at posts on a granular basis.  Unfortunately if she does what she says then she's going to majorly break all her inbound links from Google.  Maybe a lot of mod_rewrite ...
• John figured out how to WiFi an old iMac even though he needed a PC to do it.  Wonder if Cisco buying Linksys will make this better or worse?
• Adrian makes some really good comments about how to make registration forms so that people don't cheat. 
• Mark is talking about HTTP error code 410 and how to make pure CSS tabs.  Both are excellent and recommended.
• Mkelley is adding tables back into his CSS!  Hooray!  I'm almost certain to do this to Feedster in next few weeks.  No matter what I do, I can't get the CSS that works on Mozilla on the Mac, IE on the PC, Mozilla on the PC, Opera to work for IE on the Mac.  Now that I see someone else doing this, I can do it without feeling like an utter lame-oid.
• PHP-Con is coming!  Ack.  Need to write my presentation!
• Adrants reveals the truth of Tivo.  No surprise here.
• CNet is talking about the Longhorn file system.  As much as I like the concept, the idea of switching to a brand new file system on a brand new version of Windows scares the crap out of me.  The advice I'll be giving to my clients?  "Wait for at least the first two service packs or beware".
• Andy shows us how Nortel resizes
• A whole bunch of frothy Mozilla goodness is written up by StudioID.
• If you're a Mozilla user then run, don't walk to get Hippocampus.  Way cool. (And yes Andy, I added Blozom News to Feedster).
• I just plain love Simon's PHP and Javascript spelling checker.  Very cool.
• For something different from the norm, Textism tells us how to make soup. 
• Via Evhead: More problems with the Danger over at SixLog.  The worst customer service of the day award has to be awarded to Danger and the Sidekick.
• Interesting thoughts on content targeted ads in search engines that talks about how Google does it.
• Peterme is back and has switched to a Mac.  I'm now using my iBook nightly to work on Feedster when I sit on the couch but I **strongly dislike** it due to the keyboard, lack of delete and backspace keys and alt+tab model.  Ah well... Its damn pretty and the Forest screen saver just plain rocks.
• Not sure if I blogged this before or not but The FurryGoat (love that name) has good thoughts on the whole ISV issue I mentioned earlier that Dave raised.
• Thoughts on Commando Development from Sterling Hughes.
• If you use PHP and don't get offended easily then look at this.
• Jeremy on the unintended consequences of contextual advertising.
• Marco is thinking about why PHP doesn't appeal to businesses and on why SPEWS just plain sucks.  I'd agree with him here.  SPEWS does suck.

9:48:46 AM      Google It!   comment []    IM Me About This