Re: WAIS

Martijn Koster <m.koster@nexor.co.uk>
Message-id: <9307191554.AA04223@dxmint.cern.ch>
To: warnock@hypatia.gsfc.nasa.gov (Archie Warnock)
Cc: www-talk@nxoc01.cern.ch
Subject: Re: WAIS
In-reply-to: Your message of "Mon, 19 Jul 1993 10:12:35 EDT." <9307191412.AA13345@Hypatia.gsfc.nasa.gov>
Date: Mon, 19 Jul 1993 16:54:00 +0100
From: Martijn Koster <m.koster@nexor.co.uk>
Status: RO


Archie Warnock writes:

> It's efficient because you don't have to retrieve an entire document
> using WAIS - only a URL.  If you let the WWW client handle the document
> retrieval from the URL, you always have the documents in context - i.e.,
> you always retrieve from the HTTP server, so things like relative links
> work correctly - they don't if you use WAIS to retrieve the document.
> WAIS handles the searches, so your starting point in the web is a
> document that's relevant to your topic of interest.

I agree completely. I don't want to use WAIS for document retrieval.

> It's easy to implement already, and getting easier.  

I am not quite sure what you are referring to here... 
implementation of WAIS in a WWW client?

What I am after is the best way an information provider can provide a
searcheable index of web-accesible information into the web.
I want the performance and the algorithm of WAIS, but I don't really want to 
have to run a WAIS server, and I'd like to have some HTML control of how the 
results are displayed. Have a look at the searches in my Mac Archive to see
an example <http://web.nexor.co.uk/mac-archive/welcome.html>. Another
application would be the UU Meta library.

If you generalise this I want to do WWW searches for keywords in a file
with three fields, like:

	title	keywords	URLs(or any HTML)

and get a list like:

	<H1>Results</H1>
	<H2>A title</H2>(Score: 1000)
	...the HTML, with a pointer to some document...
	<H2>Another title</H2>...

This would give a general framework to publish indexes, especially if I
can just run a program on this file, and dump the result somewhere under
a HTTP server that would know what todo with it.

> > Can you not achieve the same thing by having waisindex generate an index
> > of a file with titles, keywords and URL's, and have a command line WAIS
> > client called from the HTTP server, or compiled into the HTTPserver do 
> > the lookups? This would eliminate the need for the WAIS protocol, and in
> 
> You probably could, but this seems to require less code hacking.  It's
> not a hard thing to do - it's a clever thing ;-)

If there is an easy way to do the above I'd love to hear it. 

-- Martijn
__________
Internet: m.koster@nexor.co.uk
X-400: C=GB; A=Mark400; P=Nexor; O=Nexor; S=koster; I=M
X-500: c=GB@o=NeXor Ltd@cn=Martijn Koster
WWW: http://web.nexor.co.uk/mak/mak.html