Skip to main content.
home | support | download

Back to List Archive

Re: meta names not included in swishdefault?

From: Bill Moseley <moseley(at)not-real.hank.org>
Date: Tue Feb 26 2002 - 20:38:54 GMT
At 12:19 PM 02/26/02 -0800, Fred Toth wrote:
>Bill,
>
>>So you can see how the words are nested.  Everything has the "META" 
>>structure flag set since everything in within <html>.
>
>Uh. Maybe. I wonder if you could explain this output a little:

Oh, I guess it's a bit cryptic!

 Adding:[1:html(10)]   'titletext'   Pos:2  Stuct:0x87 ( META HEAD TITLE
FILE )

Ok, adding to file 1, meta "html" (metaID number = 10) the word
"titletext".  Its word position 2 (used for phrase matching).  It's
"structure" is 87 hex.  "structure" is a bit flag that indicates where in a
HTML document that word is located, and in this case it's in " META HEAD
TITLE FILE"

META = because we said <html> was a metaname (so everything is inside of a
metaname)

HEAD = inside the <head> section
TITLE = inside the <title> section
FILE = is something I don't understand.  I think it says it's in a file,
which everything is in a file, so I'm not sure about that one.


 Adding:[1:html(10)]   'meta1text'   Pos:6  Stuct:0x85 ( META HEAD FILE )

Only difference is that this is not in a <title> section any more.  The
position "6" is bumped from "2" because of the </title><meta> tags --
that's to keep phrases from matching across tag boundaries.  But, that's
configurable per metaname if you want to match phrases across metanames.  A
classic example might be:
    <person>
      <first>Bill</first>
      <last>Moseley</last>
    </person>
where you would still want to be able to search -w person="bill moseley".

Those -T options were added to make debugging easier in development, but I
find that they are really useful to make sure my configuration is working
like I think it should work.

Anyway, does that help?  



-- 
Bill Moseley
mailto:moseley@hank.org
Received on Tue Feb 26 20:42:43 2002