According to Malcolm Austen: > On Thu, 21 Feb 2002, Gabriele Bartolini wrote: > + Sorry if this question may be silly, but I didn't know about this > + before. I know I should have looked at the sources first, but I think that > + users should perhaps know about this. Here is my question: by default, > + ht://Dig gives different weights to words depending on the tag they are in. > + What are the default weights? > > I collected up various parts of the 3.1.5 documentation into a web page: > > http://wwwsearch.ox.ac.uk/scores.html > > + header1 i:0 l:734 w:266 > + header2 i:0 l:800 w:200 > + header3 i:0 l:865 w:135 > > Those weights nicely show the undocumented (and unconfigurable) > 'location' weight mentioned in my page above.
Back in July we discussed possible ways of configuring this, but none of that ever made it into 3.1.6. It occurred to me that we could have a "location_factor" which would default to 1000, and words would be ranked from that factor at the start of the document, down to 1 at the end, multiplied by the appropriate scoring factor for the type of word in question. To make location irrelevant, you'd just set location_factor to 1 and all word scores would be multiplied by 1 regardless of location, so only the other scoring factors would have any effect. I think this would be more intuitive and manageable than the last_word_factor we discussed back in July, which wouldn't change the builtin factor of 1000 for first words. Maybe you can nag me about this if/when Geoff and I commit to a 3.1.7 release. Meanwhile, I'll add a link to your web page in the FAQ. Thanks. -- Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]> Spinal Cord Research Centre WWW: http://www.scrc.umanitoba.ca/~grdetil Dept. Physiology, U. of Manitoba Phone: (204)789-3766 Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930 _______________________________________________ htdig-general mailing list <[EMAIL PROTECTED]> To unsubscribe, send a message to <[EMAIL PROTECTED]> with a subject of unsubscribe FAQ: http://htdig.sourceforge.net/FAQ.html
