Scott the Webmaster Returns to Scott the Search Engine Engineer

Well most of today was spent on basic webmastering, content writing and CSS stuff (and I've munged a lot of CSS -- I know it won't validate any longer).  But I just caught my breath and put in a requested feature: RSS Filtering.

Here's What a Filter Does

If you've used Roogle at all then you're starting to see that your own posts show up a lot.  That's bad (well sometimes) so I implemented a new Advanced tab which has the advanced search form with the Filter option.

Let's just say that you don't like my blog.  Then you'd enter http://radio.weblogs.com/0103807/rss.xml in the filter field and do the search.  And this is what you'd get:

• Before
• After

Other

Can there be more than one filter?  Not right now.

Does it remember my filter?  Not yet.

Can I filter from the search results?  Good idea (I just thought of it).  I'll get on that.

Who thought up this in the 1st place?  Well the filter concept credit goes to, well I can't remember but thank you for the idea.  Leave me a note here and I'll give you the credit and a cross link.

11:16:38 PM      Google It!   comment []    IM Me About This   

Being Slashdotted

Demitrious, the contract Sys Admin (he's awesome, just awesome) who works with me on stuff, tells what he did to help the server stand up today.  [_Go_]

8:37:31 PM      Google It!   comment []    IM Me About This   

Best IM Conversation of Today

The worst part is he's close to right*...

kjartanmannes: so whats next for Mr Johnson?
fuzzygroup: in what context ?
kjartanmannes: well, you've been slashdotted so what is your new goal in life?

My sincere thanks to all the messages of encouragement, nice feedback and other comments.

8:27:40 PM      Google It!   comment []    IM Me About This   

Roogle Changes

A couple of new things and some metrics:

• Approximately 1000 odd feeds now
• Reindexed the database
• I wrote an OPML loader so if you want me to load an OPML feed from your aggregator, email it to me and I'll do so.  I'll add a webpage to do that too but that'll take a bit.
• Credits are greatly expanded here and all the logos people sent me are shown as well.
• 43,000 plus queries.
• 13,000 plus distinct ip addresses logged.
• 219,604 plus hits so far. 
• The real full text search engine is partially working -- I have all the code but its a matter of finishing it while doing everything else.

6:46:32 PM      Google It!   comment []    IM Me About This   

Less Google At Roogle

Whole new look and feel (that's the "Less Google", about page, contact info, incremental database index in progress for some of the updates and the absolute fastest webmastering I've ever done.  *Wipes Brow*

3:50:02 PM      Google It!   comment []    IM Me About This   

Oh God I've Been Slashdotted

Ack!  Foo!  Run in circles, scream and shout!

NOTE

This was done as a research prototype to gauge interest.  Clearly there is interest.  Now that I know that, it'll get a lot better.  The search index is by no means comprehensive and the search logic is being fine tuned.  True full text search will be available either later today or tomorrow.

12:50:22 PM      Google It!   comment []    IM Me About This   

Beauty in Computing

A little random but I'm thinking over stuff as I ponder next features (and I also needed a post to verify my rss expiration function).  I'm always amazed by the prettiness that Linux is capable of when the right theme is developed.  A.Sleep's stuff is a good example.  [_Go_]

What made me think of this is Brent's rant on Windows XP and the playskool interface.  I wonder if he'd like A.Sleep's stuff better.  [_Go_]

10:26:03 AM      Google It!   comment []    IM Me About This   

Roogle Now Groups Search Results by Title

Ok.  Thanks to Jason's suggestion, I just implement blog grouping by title i.e. blog titles are now reported on the result list above the blog posting itself.  And if you want to jump directly to a blog's home page, that's now done too.  [_Example_]

Database was rebuilt and re-indexed.  Another 1447 postings as of an hour or so ago.  If you're wondering why I'm doing this all manually for now rather than as a scheduled job, its the normal cautiousness with a new system.

The award for "Most Cogent, Well Thought Out Email I Ever Received on a Saturday Night" goes to Mike of EraBlog.  I'm not a .NET guy but if I was, I'd be checking out Mike's stuff like asap (EraBlog is a .NET way to blog).  Mike -- I'm thinking strongly about your points.  Thank you.

8:37:28 AM      Google It!   comment []    IM Me About This   

Desperately Seeking ... Algorithms !

I know, I know.  Single guy on a Sunday morning shouldn't be searching for algorithms.  Such is the nature of a dedicated geek though.  Here's the request, appropriately enough written as a personals ad:

You: Small, petite, memory shy algorithm able to take a few hundred bytes of text and return to me the correct natural langauge codes i.e. give me "jp" or "il" or anything more correct than what's normally in the <language> element.

Me: Aspiring RSS search engine looking to broaden my horizons, experience new urls and (gasp) boldly recognize languages correctly.

Other: Special points given for being written in PHP.  Extra points given to red heads (oops -- wrong context; scratch that).  

Goal: Long term embedded relationship but will date before marriage.

I know this exists.  I can even remember sitting in an office in Albany, NY one day talking with John Munson (whose email address I no longer have) and discussing it.  I cannot, for the life of me, remember how it worked or its name.  And I'm googling poorly this fine morning.

Thoughts?  Anyone out there got any code to toss my way? 

Example of Why I need It: Here's a blog and here's its rss feed.  Now here's its language element: <dc:language>en-us</dc:language>.  And there's the problem -- this isn't english by a long shot.  But I don't think the problem is to require everyone out there to set this properly.  As they say "sh*" happens and computers are supposed to be smart enough to recognize this. 

Note to hlb -- I'm not singling you out here guy, you're just one of the hundreds if not thousands of blogs with a mis-set language field and you're just the example I happened to grab at random.  This posting also ensures I can find my test case when I need it again so at least by posting this, you know that I'm going to try and get at least your case fixed.

7:12:30 AM      Google It!   comment []    IM Me About This   

Roogle :: It's a Whole New Blog Just for Roogle

For those who don't care a whit for any of my other antics but feel that Roogle is interesting, I just set up a new blog for it.  [_Go_]

And here is its own RSS feed even. 

You may notice that the location is below the FuzzyGroup domain not below the Roogle directory.  Why?  Because I know the name won't stay Roogle for long and I don't know where its going so this at least gives it a permanent home.

I know the UI needs to be synchronized with Roogle itself and that will happen but it needs to wait for the UI on Roogle to get updated.

6:00:51 AM      Google It!   comment []    IM Me About This