CS>[List Owner] On subject morphing...
From: M. G. Devour
Date: Fri, 22 Oct 2004 16:04:32
http://escribe.com/health/thesilverlist/m74279.html

  [... trim to save everyone's time]

  > The search engine at the archives is extremely weak. It  will only
  > accept search terms of 4 characters or more. It does  only logical
  > AND searches  on  multiple  terms. No way to  do  logic  OR  or to
  > exclude results by a NOT function. It's very crude.

  > It does,  however,  search the entire  message,  Subject  line and
  > body. Therefore  searches  will  still  find  a  subject  based on
  > keywords in  the  body even when the subject  line  morphs  or the
  > topic has slid out from under the old Subject.

  > So searching will typically be a multi-stage process, using search
  > terms to identify an initial set of messages, then clicking on the
  > View This  Thread link within particular  messages,  reading that,
  > then going back to grab the next thread, and so on.

  > Seriously, I  can barely see an advantage to  *either*  method, as
  > far as  searching is concerned. Meanwhile morphing the  Subject to
  > keep up with the actual topic has the advantage of being easier to
  > skim through a subject or date ordered listing...

  [...]

  > We should discuss this openly here for now, so everybody  can know
  > what's being said. I'm listening.

  > Mike D.
  > da list owner guy

  Mike,

  It would  probably  be  difficult to set a  policy  on  changing the
  subject line. Sometimes it makes sense and sometimes it doesn't, and
  any attempt to set a policy would bring endless posts promoting both
  sides. As you point out, the archives are invaluable,  but searching
  is very difficult.

  Apparently it  only accepts single words of 4 characters or  more. I
  just tried  a  phrase  that I know exists in  the  archives,  and it
  couldn't find it.

  I would love to have a chance to write proper search code  to search
  the list  archives,  and  would be happy to  volunteer  my  time and
  effort. Perhaps  the  best  model is  Google,  where  they  look for
  different spellings  for the search terms you entered,  and  offer a
  choice. There  are many ways to do this - starting with  the Soundex
  method and going to more advanced techniques. For someone  who loves
  high-performance code, this would be an interesting challenge.

  Another option is to use one of the the free Google  search options,
  such as the free Public Service Search described here:

    "Google is   pleased   to   offer   educational   institutions and
    non-profit organizations worldwide free SiteSearch,  which enables
    users to  search your website, and free  WebSearch,  which enables
    users to search the Internet."

    "The features of the service are listed below:
    Free SiteSearch with optional WebSearch
    No advertising
    Unlimited queries
    Traffic reports
    High level of customization"

    http://services.google.com/publicservice/login

  This would give people who are familiar with using  Google immediate
  access to  the archives, and allow the system to  improve  as Google
  adds new features. All it would take is having the site  owner allow
  Google to crawl the archives, which is a bandwidth issue  that costs
  money.

  But after the site is archived, all searches are done  from Google's
  cache, which cost the site nothing. So it would actually  reduce the
  bandwidth cost for searches.

  This is definitely worth discussing with the site owner. If it would
  make any  difference, perhaps we could ask for a vote for  all those
  in favor of improving the search, and all those against:)

Best Wishes,

Mike Monett


--
The Silver List is a moderated forum for discussing Colloidal Silver.

Instructions for unsubscribing are posted at: http://silverlist.org

To post, address your message to: [email protected]
Silver List archive: http://escribe.com/health/thesilverlist/index.html

Address Off-Topic messages to: [email protected]
OT Archive: http://escribe.com/health/silverofftopiclist/index.html

List maintainer: Mike Devour <[email protected]>