Re: Search engine

[Norman Bauman nbauman@xxxxxxxx]

If you're creating a text search archive, you might want to peruse my
articles on text searching from Law Office Computing, which I maintain are
the best articles ever written on the subject (my challenge has always
been, if you know a better article, let me know).

http://www.nasw.org/users/nbauman/index.htm#txtsrch

Norman

At 03:43 AM 12/10/02 -0800, Robert Holmgren wrote:
>
>
>XySearch is Up and Running!
>-------- -- -- --- -------
>Thanks sooo much to everyone for the offers of files, and
>the interest in this quasi-archaeological endeavor to
>recapture XyWrite's past and learning experience. I've had
>a LOT of fun putting this together, no kidding. I would
>love to present you with the complete package, neatly
>wrapped, just in time for Christmas; but we're not quite
>there, and I think the more pointful approach is to open
>XySearch to your scrutiny, your SUGGESTION, COMMENT, and blunt
>CRITICISM, and then to take the site down for a day or two,
>implement improvements, and add the substantial
>contributions you have donated -- about three to five
>thousand "new" old messages -- to the database.
>
>The database is current through 9 Dec 2002. Presently it
>contains just shy of 16,000 messages, dated between 1994
>and 2002. Because I have to travel for a week or so, and I
>have yet to implement an automatic daily update of messages
>(how does 5:30am EST 24x365 sound?), there will be no
>changes until I return.
>
>Please explore the facilities. The PRIMARY PURPOSE is to
>reduce the amount of repetitive question-and-answering here,
>of revisiting the same issues, with the same results, over
>and over. XySearch is intended to be USED, to edify and
>occasionally to amuse. To put a sharper point on it, to be
>used *before* asking, if you suspect that the matter might
>have been broached before. I assume that everyone is
>familiar with DejaNews (now Google Groups), incontestably
>the most important problem-solving tool on the Net. The
>XySearch experience will be similar, although the query
>syntax diverges somewhat -- look at the examples (How to
>Search) -- we intend to enlarge this section of the site in
>the near future.
>
>A few tips:
>I've found that this engine generally produces too many
> results on the first essay, and that I need to add search
> terms (keywords) to prune to manageable numbers.
>Use the asterisk wildcard liberally, it's more necessary
> with this engine than say Google.
>Find a productive thread, then follow the thread, not the
> search engine; when a thread peters out, toggle to the
> "Threads in Chronological Order" index, because you'll often
> find other discussion threads with similar toplevel Subject-
> names, clustered in the same time period.
>The Index pages are large -- I cannot ABIDE looking at 10
> pages of 1/10th results, especially when they're
> alphabetical or chronological -- trying to remember what I
> just read two pages ago, the Back button, ugh! -- I want to
> see the whole thing at once. But that means longer download
> waits. So once you get a big page, keep it: open a New
> Window to look at individual messages, then just kill this
> secondary window when done (Alt-F4), and you're back on the
> big index page, no waiting around to reload. (Bear in mind
> that this applies only to you and not to me; I'm downloading
> at 100Mbps, whereas you won't top 120Kbps. Still, on a 53K
> *dial-up*, big indices appear within 2 seconds generally,
> and results seldom take more than 1-2 seconds to appear.
> Messages <1 second. Not bad for a 133DX P1.)
>
>So: You'll find the XySearch URL at EOF. But first ...
>a bit of memory lane?
>
>****************************************
>****  Message Section Commands  ****
>****************************************
>
>[R]ead Messages
>[Q]uick-Scan Messages
>[E]nter New Message
>
>[-]Previous Menu
>[0]Top Level Menu
>
>[U]pload
>[R]equest
>[F]eedback
>[L]ogoff
>
>Message Section: XyWrite Message Forum
>
>Command: L
>
>Logged on at 20:51:27
>Logged off at 21:12:34
>
>Thank you for calling the XyQuest Information Line
>
>>> 92/11/12 20:03:30 Carrier dropped, session ended.
>>>+ 92/11/12 20:03:30 Elapsed session time: 00:21:13
>
>NO CARRIER
>
>----------------
>
>http://users.datarealm.com/xywwweb/xysearch.cgi
>
>-----------------------------
>Robert Holmgren
>holmgren@xxxxxxxx
>-----------------------------
>
>
-------------------------------------------------------
Norman Bauman
411 W. 54 St. Apt. 2D
New York, NY 10019
(212) 977-3223
http://www.nasw.org/users/nbauman
Alternate address: nbauman@xxxxxxxx
-------------------------------------------------------