About Vancouver Webpages'
jump! is a search tool with a rather single-minded purpose -
it finds Website homepages. For example, if you search for "kodak"
you will get www.kodak.com
but not a lot of pages with people talking about Kodak film.
How jump! Works
jump! first looks in its database for a match. If it finds
only one match, you are jumped right to the page (using an http redirect).
If jump! finds more than one match, they are listed. Sometimes
an apparantly identical listing appears twice. Entries are preceded
by a hidden comment which is part of the search key, so that
www.alstrucks.com "Al's Trucks Inc" will be keyed as "alstrucks"
and as "al's".
If jump! can't find a match in it's database, it guesses.
It will typically try "www.name.com" and "www.name.org", and also
try some case folding and punctuation folding, eg. mapping
"London Drugs" to "london-drugs". If it finds a match like this, it
updates its database automatically.
If jump! fails to guess a site, it presents the user with
a more detailed form including a netfind and InterNIC
search. These may give the user clues as to a possible Website.
If the user then uses jump! to try the new URL, then it
will be added to the database if successful.
What keys does jump! use?
jump! uses the following keys:
So "Welcome to Amazing Inc, manufacturers of Wizzo putty" would
be keyed under "amazing" and "wizzo".
- The domain title, eg. "kodak" from "www.kodak.com"
- The first word in the homepage title that is not "Welcome', "Index",
or a preposition.
- Any word in the title that isn't in a dictionary.
Where does jump! get its data from?
jump! gets URLs from people using it. It is also fed by a robot that
scans InterNIC and netfind information looking for Web servers.