site search engine

On Thu, 1 Apr 2004 17:53:00 +0200 "Jo" <sliacloujb.TakeThisOut@jetable.com> broke
off from drinking a cup of tea at Club-Internet / T-Online France to
write:

 >I am looking for a site search engine not requiring an outside link to
 >operate. In other words totally independent, working with php and mysql. I
 >have about 600 pages to scan, still growing.
 >Can anyone suggest a good standalone shareware or professional program?
 >Thanks for your help
 >Jo.
 >
 >

No. Which is why I wrote my own! <g>

Before anyone accuses me of spamming, I don't sell it, I created it
strictly for in-house use.

All the available search engines were returning a few lines with a
"relevance" mark and a link to the actual page/article.

I used Perl to scan the pages for the entered text (word or phrase),
further complicated by adding an index of article titles in a flat
file. The basic principle could be built upon to allow boolean
searches or even intelligent searches (though these require so much
processing time as to be unviable, at least real ones do).

The format of such a search engine will depend greatly upon your
website.

If you have a product catalogue web site, with a separate page for
each product, writing a search engine becomes trivial, more so if each
product page follows a convention, for example having the product
title and code within the <title></title> tags. In this case one could
simply loop each file in the directory checking the <title> line for
an entered product name or code and if found return the relevant page
and if not a "not found" type of page.

I don't know PHP, but this is just what PERL was created for and as
such has a very powerful regular expression parser which allows
location of strings, substrings and the like.

You may also need to consider processor overhead, particularly if you
are hosting on a shared server.

Matt

--
A massive matrix of concise, interlinked encyclopaedia information.
For when you just want to know, quickly and easily.
<a style='text-decoration: underline;' href="http://www.probertencyclopaedia.com" target="_blank">http://www.probertencyclopaedia.com</a><!-- ~MESSAGE_AFTER~ -->