It's very simple. First, make sure you've got a working, running installation
of Apache with PHP in it *and* available as a command-line binary, a working
MySQL database, and a working keyboard. That helps. THEN:

0) You've probably already unpacked the distribution file. Just put it all
   into a directory somewhere. If you don't put it under your web server's
   DocumentRoot, make a symlink *from* the DocumentRoot *to* wherever you
   installed it. This will be the URL you use to get to UniSearcher from a
   browser; if DocumentRoot is /var/www and UniSearcher is in /var/www/uni,
   the URL to get to it is http://your.host.com/uni/  (but you knew that.)

1) Confirm a few things. I distribute it as a tar because zip doesn't
   preserve file modes/permissions. So the following files *should* already
   be set "executable" (mode 755) but make SURE they are for me, please:
   "allchars1", "allchars2", and "parser". A helpful command for that:
       chmod 755 allchars? parser data
   ("data" is a directory, but it should be 755 too.) Every other file
   should be mode 644. By the way; the web server part of UniSearch never
   *ever* writes anything to either a file OR the database. Every bit of
   data is read-only as far as Apache is concerned. The only script that'll
   ever write anything anywhere is "parser", so it does need to be able to
   create files in both the installation directory and its "data" subdir.
   If you always run the parser as "root", no problem. Apache never even
   reads the data files themselves; it only reads from the database. This
   is why UniSearch doesn't even need an administration section (something
   that's rather unprecedented for one of MY applications). You could make
   "data" mode 000 and it wouldn't break anything as long as you're running
   "./parser" as root. But don't run it yet. Patience...

   Also confirm the "magic string" at the top of those three executable scripts.
   They all say "#!/usr/local/bin/php" right now because that's where my php is,
   so if yours is somewhere else, change that line in those three files (once
   again, "allchars1", "allchars2", and "parser").

   And now for a note about the "register_globals" setting in php.ini. If
   it's off, UniSearcher won't work right. I plan to go through and change
   *everything* to use the $G array ExtractGlobals() creates, but that's not
   done yet. Version 1.1, here we come... in the meantime, the .htaccess
   file I've included will turn that setting on FOR you just before calling
   the php interpreter (.htaccess files are just about the only place you
   can override that particular setting; once the script starts running,
   it's set in stone). So you don't have to worry about register_globals,
   in this or any future version. And thus endeth the confirmations.

2) Edit the "db.php" file and set up the parameters for your database. You
   can call the database itself anything you want; it defaults to "unicode".
   I recommend moving it somewhere else, where your web server can't serve
   it (somewhere not under the DocumentRoot), like "/etc/unisearch/db.php"
   or something. Just remember where you put it, edit the "functions.phps"
   file, and set the $DbConnector variable to the full path/filename.

3) Create an empty database named whatever you set $dbname to (in db.php).
   You do this through PhpMyAdmin or the "mysqladmin create unicode"
   command from a shell somewhere. Run "mysqladmin" by itself for help.

4) Import the DB structure by running this command from wherever you put the
   files from the distribution: "mysql unicode < unicode.sql". Use the "-h"
   and "-p" and "-u" parameters to "mysql" if you need them; "mysql --help"
   spits out a Library of Congress if you need help. And use whatever you
   entered as the $dbname variable in "db.php", of course, if not "unicode".

5) Run this command from the install directory: "./parser". Wait for it to
   do its thing and tell you it succeeded; then you can delete the data
   files it downloaded from the "data" directory if you want, or keep them
   around, whatever. If you want to update your database from Unicode Inc,
   you'll have to delete the data/* files first or it won't get new ones.
   They do actually update them from time to time. Incidentally, you'll
   need the "wget" application on your machine; the parser relies on it.
   But *everyone* has wget nowadays, right? If you don't, you'll have to
   download all five files manually and put them in "data/" yourself... and
   only THEN run the parser. If the files already exist, it won't even
   notice that it doesn't have wget available. All five's URLs are in the
   'parser' script itself, near the top. Have fun.

6) Open the URL to your install dir in a browser and enjoy.

If something doesn't work right, tell phee@isthisthingon.org and I'll see
if I can help.
