Friday, January 25, 2002
A very rare request
I got an email today from someone looking for a tutorial or resources on writing a metasearch engine. I got the email because I've written three of them over the past six years or so.
I told the person that I know of no tutorials or references about writing metasearch engines since when I started, there were about two in existance that I knew of and that to me, it was pretty straightforward what you need to do in order to write a metasearch engine: you get query (using CGI most likely), reformat the queries for each engine you support and make the request like a browser would (which means you need to support HTTP and CGI from the client side) and then process the pages you get back (so you need to parse HTML) and display output like any CGI script can.
Okay, maybe not that easy as there are some nagging details you only find out about by doing an actual implementation (like certain IIS servers will send out two complete header sections, or that IIS doesn't follow the HTTP redirect specification at all, and what exactly are you supposed to do if you get a redirect on a POST?) but if you take it piece by piece it's not that overwhelming.
Or is it just me?