lynx-dev Search engines for searching attributes in search engine URLs (was: [ no useful subject ] )

[David Woolley]

> <li><a
> href='http://research.amnh.org/cgi-bin/herpetology/amphibia?by=Distributi
> on&method=Contains&string=Afghanistan&Search=Search&.cgifields=method&.cgifields=
> by'>

This is syntactically invalid and relies on the error recovery of browsers
when presented with undefined entities.  It should be:

<li><a
href='http://research.amnh.org/cgi-bin/herpetology/amphibia?by=Distribution&amp;method=Contains&amp;string=Afghanistan&amp;Search=Search&amp;.cgifields=method&amp;.cgifields=by'>

> In other words, we would like to search the web, not for "link:" or
> "anchor:" but for "attribute:". :Attribute:" searches do not exist, but
> why not?

I suspect the answer is because it is a rather technical requirement for
which there is no significant demand.  Also, embedding simulated form
entries into href's is a relatively recent trend and, because of the
&amp issue; sites expecting to be used in this way are advised (HTML4.0
specification) to provide some alternative way of delimiting parameter.
The use of ; is suggested, but because this is for machine to machine use,
almost anything could be used, including /.

> We would like to find all sites that allow searches via their cgi-bins so
> that we can get world wide sites that can and do search for  "?by=" stuff
> such as country, date, type, etc.

I'm confused here.  Are you talking about using a search engine to find
search engines and what will you do when you find them.  Most search
engines are paid for through advertising and many will object to bypassing
their entry page with the sort of URLs quoted above.
 
> Thanks for lynx and your efforts. 

Still waiting for a Lynx question...