Genii Weblog

Possible solution to a Google(per)plexing problem

Wed 3 Sep 2003, 09:29 PM



by Ben Langhinrichs
There have been several posts floating around the Domino blogosphere about Google and other search engines indexing information badly (e.g., Ed here and Colin here), so searches that make little sense find a blog which has mentioned a number of keywords in different contexts.   

A related problem comes with frequently updated information  which is not re-indexed quickly enough by Google or the other search engines.  A search may show your blog, and the search may even be a perfectly legitimate one, but the article to which it refers may have scrolled off into the archives.  This is annoying for the searcher, who can't see why a search for rich turtle geometry results in first place hit for GeniiBlog since the relevant article has been archived.  (This example is bad, because the index has now re-indexed and the archive is showing, but it was the case for a while).

A possible solution to both problems is suggested by Wolfgang Flamme here.   The gist is, when a search comes from a search engine, rather than just showing the webpage, the server could convert the search terms shown in the referrer from any of the well known search engines into a Domino search string.  Thus, the results shown would be  appropriate.

I was wondering if we could take it one step further.  My pal Andrew Pollack has a cool product called NCT Search which is pretty powerful and pretty flexible.  I was wondering if it would be possible to do a two step process such that a search would be done on the search terms, but if no results were found, a new page would display that would indicate why they were being sent here.  In other words, in this search that Ed Brill mentions, a Domino search would find no hits, but it would be worse than useless to put up a screen that said "No documents found".  It might also be useful to find out first if the current weblog content would be a hit, and show that instead of search results.  Smart searching of this nature would lead to a much better interaction with those searching, and better exposure to appropriate content, which serves our needs as well.

So, Andrew, what do you think?

Copyright 2003 Genii Software Ltd.

What has been said:


45.1. Andrew Pollack
(09/03/2003 09:01 PM)

Ohhhh, I like it. I think first thing tomorrow I'll create a bit of code that captures any inbound "google" referrals and passes those search terms to NCT Search, which will then search the entire site, and produce a page that says "here's the page your search found, but these also may be of interest". I like it! It should be VERY easy to do. It would need to be a script library referenced in the WQO for all documents, forms, views, etc. for a site though.

Another way might be some javascript on the web page itself that looks for the same referal code, and if it finds it, adds text to document with "other possible hits". That way, the javascript could be included in the web site html envelope, and not a design change on every design element. hmm.....

Thanks, Ben.


45.2. Rob McDonagh
(09/04/2003 08:22 AM)

Yeah, that would be sweet. But would Andrew be willing to grant a restricted license to 'blog template developers (based on the requirement that when using this great idea a prominent credit would be given to NCTSearch)? 'Cause otherwise Andrew's the only one who'll be able to make use of it. The rest of us could use our own search routines in a similar fashion, but I don't think I'm stretching the truth to say we'd be missing out on quite a bit of functionality.


45.3. Ben Langhinrichs
(09/04/2003 09:54 AM)

We really need to find a way to make blogs a profit center so that bloggers have some cash available for these tools. <grin>


45.4. Julian Robichaux
(09/06/2003 01:32 PM)

Amen to that Ben. If you figure that one out, please let me know. ;-)

- Julian