Anyone Want a Feedster Bookmarklet ?

I was thinking it would be really useful to just highlight some text on a webpage and click a bookmarklet in your browser link bar and have it do a feedster search.  Matt Mower gets all the credit for this idea.  And, well of course, Bottom Feeder who already has this working.

Anyway ... I just update to IE 6 which seems to have disabled bookmarklets.  I can drag it to the links bar andeven though I say Yes its ok to add it, it seems to just disappear into the void.  Anyone know anything?

So I have most of the code for this but I can't test the pesky thing.  Arrggh!

10:48:56 PM      Google It!   comment []    IM Me About This   

Is this Real or How Are You Going to Make Money from Feedster?

On a more serious note, I've been getting some questions as to "This seems to be real ... Ah how are you paying for it?".  Good question.  Having suffered through a particularly business challenged DOT com, suffice it to say that I don't expect that to happen.  And bear in mind that I've done this stuff before.  That's not a guarantee of success but it is hopefully at least a small indicator in that direction.

We've got some fairly good ideas of how to make a business out of this that will roll out over the coming months.  I can tell you this:

• No banner ads
• No popups
• No selling of personal information

Beyond that I'm really not ready to say.  And we're already seeing additional interest in our bread and butter consulting services so that's great.  But as they used to say "we will sell no wine before its time" ...  Well I think you get the idea.  Its early yet so while we know how we think we can make money, we need to fine tune the product, work with users, etc.  You know the drill.

On Leaving Us Feedback

If you leave us feedback, leave us an email address if you want a reply.  That's pretty simple.  And be clear if you can.  Just give us something to work with.  Every single comments is a) logged to a database b) assigned a status (open, done, closed, etc) and almost everything is responded to.

10:35:46 PM      Google It!   comment []    IM Me About This   

Feedster Update

New hardware got setup today but is not yet part of the system.   That won't happen until the weekend.  Lots of good ideas and thoughts flowing in from all over today.  Thank you to everyone.  I'd enumerate some specific thanks but that's going to have to wait until tomorrow.  Some non-work issues took over part of my day and then I did the usual "senior geek" thing of dealing with personal stuff by diving neck deep into code and basically ignoring things until I felt better.  After 8 hours immersed in our RSS / RDF parser loader, I managed to a) deal with my issues (or at least run away from them) and b) significantly improved the overall loader, date handing, RDF elements, etc. 

Oh and somehow either the date for the server move got changed or I blogged the wrong date.  Its now Thursday night 9 pm pst / midnight EST.  We shouldn't be offline for more than a few minutes but I just wanted to let you know.

Oh and heck, here are some of the people that tossed a hand our way today (no sense in being lazy, there would just be more tomorrow):

• www.etationmedia.com for tossing some good pointers our way on some very basic web issues we just hadn't had time to deal with and for pointing out that bug I fixed
• John over at John's Jottings for helping with some important edits and for pointing out this really good CVS basics stuff from Aaron
• Trinity for doing a bang up job on looking at our results list and pointing out some very real human factors issues.  You rock!
• Rob Kaper / The Amazing US for kicking in help on implementing our results list and for getting me to fix date sorted results
• Simon for teaching me about a whole new HTML thing that's a) cool and b) useful as all heck.
• Cathleen Moore of Infoworld who's U.S. article appeared in ComputerWorld Australia of all places and brought instant messages to me.  Go figure.

10:32:47 PM      Google It!   comment []    IM Me About This   

Random PHP Oddity

Here's an interesting one.  I was using strtotime() and I noticed it failing (it returns -1 when it fails).  So here was my input:

$pubdate = "March 18, 2003, 06:20:15 PM";

When I tweaked it until it worked, I ended up with this:

$pubdate = "March 18, 2003 06:20:15 PM";

Yup.  That's right -- I removed 1 comma.  I wonder if this is a) php bug b) gnu time bug or c) by design.  And so it goes.

Oh and a quick regex to fix it and Voila!  Dates that parse.

8:05:12 PM      Google It!   comment []    IM Me About This   

Feedster -- DeDuped !

Ok.  Based on really strong user feedback, I'm going to take this live earlier than I had expected.  What I'm referencing is the experimental de-duping I blogged about last night. 

Here's the issue:

You're a blogger.  You publish content at both micro and macro levels.  In particular you have multiple RSS feeds syndicated out for different content views.  For a publisher and for a reader, that's great.  Extreme microtargeting of the audience w/ little more effort than checking a box in Radio (categories) or Movable type (can't remember the feature name).

Here's the problem from a search perspective:

Identical posts but with different urls clutter up the search results.  You know it feels wrong but you may not even be aware of why.

The solution

Smart De-duping.  As part of the indexing process, we're calculating a post level "signature" and then filtering by that rather than by the url.  Here's a picture of before and after side by side.

One thing ... I haven't (yet) put in a Google like prompt "We eliminated some duplicate results, click here to redo the search".  That'll come in the next day or two but I want to make some improvements in the overall result set display so that it fits in better.  If anyone desperately needs the duped results, let me know and I'll restore it at a separate url.

1:21:48 PM      Google It!   comment []    IM Me About This   

Remember BottomFeeder ?  It Just Got Better

I mentioned BottomFeeder yesterday?  sunday?  saturday?  The time frame gets a bit "fuzzy" lately and while I could Feedster for it ....   Anyway I wanted to point out that James, the author, has improved it dramatically by making it contextual i.e.

So I'm reading Ted Neward's post on AOP. I highlight SOP, and hit the Feedster it! menu item. Bam, I get a list of other blog entries on AOP. Very cool. If I was writing an entry of my own on AOP, I'd have all the necessary material immediately at hand, in one tool.   See James' Blog

Or from the email he just sent me:

I just extended that support by making it contextual.  In BottomFeeder, a user can select some text, and then do a feedster search or a google search on the selection - the selection being some text from an RSS feed.

I'm finding Feedster to be invaluable from a blogging standpoint - great work!

*Big Smile*.  Thanks James.  Nice work.  And I know that at least a few of you are looking for a gui aggregator on Linux / *nix  platforms.  Definitely give this one a shot.  I mean where else are you going to get a News Aggregator for AIX ?  (Most other Unices supported; your mileage may vary)

1:10:13 PM      Google It!   comment []    IM Me About This   

Infoworld on Weblogs

Wow.  I had no idea this was going online yesterday just a few hours after I spoke to her (Cathleen is the new Infoworld editor for blogs for Infoworld):

Unlike typical Web search tools such as Google that index Web content at the page level, RSS-focused search engines index RSS feeds and do so more frequently and at a finer level of granularity. "There is a huge need for this capability in blogs. Blogs are heavily nested and interlinked. We take that data stream and go all the way with it. We are getting the whole information feed, storing it, and indexing it," Johnson said.  more...

Interestingly their own search engine doesn't even have this indexed.  Link / Picture.  So I went and looked for their RSS feed using Syndic8 and went to add it to Feedster and .... it was already in our database.  I went and checked using our "Drill" function to look at all posts from InfoWorld and there it was.  Their search results, unfortunately, aren't very good though since their RSS is just article titles, not full content.  Hrm....

Thanks to Sentence & Solas who blogged this before I did.  Much appreciated.

4:41:52 AM      Google It!   comment []    IM Me About This