<P>Note:  This file has moved to <A HREF="/notablog">notablog</A>.</P>

     Remember our discussion about search engines?  Well, I went to
wwww.altavista.digital.com to test some searches out before sending
them to you.  However, it turns out that AltaVista has some new
technology so it can sort-of understand questions in english.  I asked
it "Where can I find out about chinese politics" and it gave me a bunch
of links, starting with the "Studying Chinese Politics FAQ" at:

     http://www2.mozcom.com/~w_hansen/faq.html

     Not a whole lot of info in this faq, but there're some references
and recommendations for books & music, plus a rather funny joke.  After
you're done there, you might want to go off to 

     http://www.altavista.digital.com

     and try the "Ask AltaVista a question" form.   Try the version of
the question I used above, and then try other questions in english.
If that doesn't do it for  you, then try the "advanced" link on the
top right corner of the page.  That'll take you to a new, more complicated
form.  Enter the following in the "Enter boolean expression" page.

     (china or chinese or asia or asian)
     near
     (politics or "political science" or ideology)

     And click on the "Search" key.  Then you can look at the results of
the search and add or subtract key words or key phrases to narrow the
search down some more, if you need to.

     I'm going to describe how and why I put these key words together.
Actually I skipped most of the process and put it together in my head,
but that's just because I have lots of practice.

     I'm looking for stuff on china, so I enter "china".  This will
find any web page with the word "china" in it.   But I'm specifically
looking for stuff about chinese politics, so I'll add the word "politics".
So I tell Alta Vista to look for the words:

     china and politics

     AltaVista recognizes "and" as a logical command to look for web
pages that have both key words in it.  But the word china could be at
the top of the page and the word politics could be at the bottom, completely
unconnected, so I'll use the "near" command instead:

     china near politics

     The near command works like the and command, but it also requires
the two words to be somewhat close to each other - in the same
paragraph, at least.

     People use several different words and phrases for politics.
So I'll add some more key words.

     china near politics or ideology
    
     There's a problem here, though.  AltaVista doesn't necessarily know
how to organize these key words.  It might show me all of the pages with
the word "ideology" in them, plus all of the pages with the word "china"
and "politics" in them.  So I have to group "politics or ideology" into
a single condition.  I can do this with parethenses, ( and ).

     china near (politics or ideology)

     This tells AltaVista that once it's found a bunch of web pages with
the word china in them, it should look at all of them for either politics
or ideology.

     Let's add another keyword, this time a phrase, "political science".

     china near (politics or ideology or political science)

     There's another problem here.  AltaVista doesn't necessarily know that
"political science" is a phrase.  It'll treat it like two separate words.
To group these two words together in a phrase, I need to put double-quotes
("") around it, like this:

     china near (politics or ideology or "political science")

     Actually, putting the "" around the words means "find exactly these
two words with a space between them".  So for example, if people liked
to put a dash between political and science ("political-science"), I'd
have to add another key phrase to look for those.

     This is getting a bit long and awkward, so I'll put it on several
different lines:

     china 
     near 
     (politics or ideology or "political science")

     Now I'll add some alternate key words for china.  How about "asia"?

     china or asia
     near 
     (politics or ideology or "political science")

     Whoops, now I need to group the first set of keywords with parentheses:

     (china or asia)
     near 
     (politics or ideology or "political science")

     Okay, that should work.  But it'll miss web pages where people use
the plural of china or asia.  So let's add those:

     (china or chinese or asia or asian)
     near 
     (politics or ideology or "political science")

     Actually, since asian contains all the letters in asia, AltaVista would
probably recognize web pages with "asian" in them, but why take chances?

     Okay, I run the search and AltaVista finds 114523 pages.  That's a lot.
AltaVista displays the links in sets of 10.  I'll browse through the first
thirty or forty and see if I find what I'm looking for.  

     Hm... actually these links all look pretty good, but for the
purpose of our example, let's say I find a lot of links to web pages
with chinese herbal medicine. That isn't what I'm interested in, and
they're making it hard to browse through the rest of it.  So I'll add
a condition to the search that says to ignore pages with the words
"herbal" or "medicine" in them:

     (china or chinese or asia or asian)
     near 
     (politics or ideology or "political science")
     and
     not (herbal or medicine)

     As I notice more trends of stuff I don't want, I can add more
key words to check for.

     And that's about it for now.  Try it out and let me know how it
goes for you.

     Wait, one last thing... let's do a search for mailing lists.
We'll add keywords to search for the phrase "mailing list" and also
for "listserv" and "majordomo", which are the names of two very
popular programs for running mailing lists.

     (china or chinese or asia or asian)
     near 
     (politics or ideology or "political science")
     near
     ("mailing list" or listserv or majordomo)

     Hm... still a lot of links, but you can look through them and
see if any of them lead you to a useful mailing list.

Steve

Steven J. Owens
puff@guild.net