Hi,

There has recently been some discussion on this list about 
'backwards search' .. links to your site.

I have a slightly different problem .. I'd like to set up a search area
on my website so that people can do "targetted searches" ie search 
either a specific URL that I 'know' is good for the sort of information 
that they are after, or a set of such urls.

In theory, you can limit searches to sites and perhaps to specific 
pages, but my results are as follows :



A.  search engine  == lycospro.lycos.com

Searching for the word "forecasting" in the specific page 
www.bizmove.com/finance/m3b4.htm

   query = forecasting url:www.bizmove.com/finance/m3b4.htm
   query = url:www.bizmove.com/finance/m3b4.htm forecasting
    query = url:http://www.bizmove.com/finance/m3b4.htm forecasting

all work fine

but searching for "collocation" with the syntax
query = collocations url:http://www-nlp.stanford.edu/fsnlp/index.html
query = collocations url:http://www-nlp.stanford.edu/fsnlp
yields nothing . (The page appears to have been indexed, in that I 
can certainly find it by a search of the form
collocations "statistical natural language processing" "companion 
website")

Searching for "disease" with the syntax
disease url:http://lib.stat.cmu.edu/ gives me nothing 
but
disease u:lib.stat.cmu.edu
gives me a list of sites that have lib.stat.cmu.edu as PART of their 
URL, but not the main http://lib.stat.cmu.edu/ page

Searching for "mumps" with
http://lib.stat.cmu.edu/disease/
mumps url:lib.stat.cmu.edu/disease
mumps url:lib.stat.cmu.edu/disease/index.html
yields nothing



B. search engine = hotbot.lycos.com
     no joy. The supposedly correct keyword is domain instead of 
URL. 
It DOES work if you want to restrict the search to the domain
eg query = forecasting domain:www.bizmove.com
but it will search all the pages below that URL

C. search engine = alta vista advanced
I could not get through today, but previous results indicated that
searches of the form
forecasting url:www.bizmove.com/finance/m3b4.htm
work, but not for eg the lib.stat.cmu.edu type search

D. Google
forecasting url:www.bizmove.com/finance/m3b4.htm
forecasting site:www.bizmove.com/finance/m3b4.htm
did not work, but
forecasting www.bizmove.com/finance/m3b4.htm
ie without keywords, at least found the correct site as 
the top one on the list.

A search on
mumps lib.stat.cmu.edu/disease
did not get the right page exactly (it  found a couple of files on that 
page; I wanted the index.html page)



E. www.search123.com

no luck at all



The only thing I can make of all this is that 

a) lycospro works best for highly restricted searches on www pages
b) nothing works very well for pages that are not www .. dk why

Perhaps Google, without specific URL restriction keywords, will be 
overall the most appropriate, as it does appear to allow you to have 
the URL or part thereof as part of the query phrase - but with no 
guarantees that the search will be restricted to that URL.

If anyone has any ideas/comments ..

Thanks
  

• Web Designer Courses
• Off the Shelf Training Courses

HWG hwg-techniques mailing list archives, maintained by Webmasters @ IWA