Print

Print


** Reply to note from Majella McCarron <[log in to unmask]>         Mon, 5 Feb 2001 09:30:12 +0000


> Here are three questions which I would ask when evaluating a search
> engine:
>
> 1. Does it index multiple document types - not just txt and html but also
> word,excel,pdf etc? I've had a bit of a read at the Google documentation but so
> far I haven't been able to find any reference to this topic.

If it has drivers to read them :-)

I believe M$ own whatever is called has drivers.
Verity (whoever owns it or whoever it owns now) used to have PDF drivers.

I thought HT-Dig had drivers.

>
>  2  Does it generate query frequency ? Site maintainers want to get inside the
> heads of their visitors to see what they're looking for from their sites. Good
> for both planning content and determining the metainformation to be included in
> pages.

You could probably extract this from your ordinary logs if it does not have its
own logging and from there on you do your analysis.

In a non-statistical manner:
- Most people search for people :-)
- Then most people search for anything that comes to their head thinking a local
engine indexes all the web.
- Then the search for pages which no longer exist or have been moved (ie. from
results of 404 errors).

>
> 3. How easy is it to set up restrained searches?
>
> There are more questions than these of course but for us at Queen's we would want a Yes
> on all three for starters.

As a general comment commercial engines may have drivers to read proprietary
file formats, but for example with M$, making every newer version of
its files backward incompatible those who write drivers would have a non-stop
job. Thus you must expect to pay appropriately.

Wish you luck.
Charles

==============================================
Charles Christacopoulos, Secretary's Office, University of Dundee,
Dundee DD1 4HN, Scotland, United Kingdom.
Tel: +44+(0)1382-344891. Fax: +44+(0)1382-201604.
http://somis.ais.dundee.ac.uk/
Scottish Search Maestro http://somis2.ais.dundee.ac.uk/