This fileset contains the VMSIndex search engine, originally developed
for use with the VMS Gopher Server, and now ported for use with the CERN HTTP
Server and OSU DECthreads HTTP Server.  We support it, and related aspects of
all three infosystem servers, via the VMSGopher-L@trln.lib.unc.edu email list.
(NOTE that none of us has any tolerance for WWW versus Gopher or Unix versus
VMS "religious wars", so be careful what you say should you seek our help!).

	Indexes are created with Build_Index.exe, which is common to all
three servers.  Its use is explained in Build_Index.Doc.  WWWquery.exe does
the searches for the CERN HTTP server, and query.exe is for the Gopher server.
Build them by executing Make.com, which will determine whether you are on a
VAX or AXP, and whether or not you have MMS, and then do the right things.
Use Make_DTquery.com, after reading its header, to build DTquery.exe for
the OSU DECthreads HTTP server.

	Start by reading the headers of VMSIndex.com and HTTP_IndexTree.com,
then the headers of VMSGopherServer.com and HTTP_IndexDoc.com, then study
Build_Index.Doc, then read the header of WWWquery.c or DTquery.c.  Scripts
for use with DTquery.exe have -dt appended to the .com.  Rename them to
foo.com before using them.

	A homolog of WWWquery.c and DTquery.c, Index_Search.c, is distributed
with the OSU DECthreads HTTP server.  The latter requires that the index (.IDX
and .SEL) files be in the same directory as the foo.html files.  WWWquery.c
and DTquery.c, like query.c for the VMS Gopher Server, allows them all to be
collected into a subdirectory of the data tree (e.g., WWW_Root:[INDEX]).
Also, our updates of Index_Search.c for keeping it functional with updates of
Build_Index.c for some reason weren't being included in the OSU DECthreads
server distributions, so we stopped bothering to update Index_Search.c, and
no longer support it (see info about DTsearch.com, below).

	In addition to indexing a set of whole foo.html files, WWWquery.c,
DTquery.c, and query.c can index single or sets of sectional database files,
analogously to WAIS.  The indexing with HTTP_IndexTree.com and searches with
VMSIndex.com are a model for sets of whole foo.html files.  The indexing with
HTTP_IndexDoc.com and searches with VMSGopherServer.com are a model for
sectional databases.

	Copy VMSIndex.com (and VMSGopherServer.com if you have reason to use
it) to your htbin directory, and set up a foreign command for WWWquery.exe
or DTquery.exe.  Copy SearchWords.html (and EGREP.html if you have reason to
use it) to an appropriate subdirectory of your data tree for access in
conjunction with searches.

	The command file interfaces for building indexes can be executed from
here.

	The VMSIndex distribution now includes the ISINDEX search scripts,
CERNsearch.com and DTsearch.com, which can be used, respectively, as
interfaces to WWWquery.exe and DTquery.exe for the CERN and OSU DECthreads
servers.  Read their headers for more information.  The ISINDEX document,
indexdemo.html, is included for testing CERNsearch.com or DTsearch.com with
the index built by HTTP_IndexTree.com.

	NOTE that any links in ISINDEX cover pages returned via direct calls
to an htbin procedure without a query should have references with absolute
paths (i.e, they should have href="/path/foo.html", not href="foo.html").
Otherwise, clients will fill them in as "/htbin/foo.html" instead of using 
"/path/foo.html".  Partial URL's can be used with ISINDEX documents if the
files are to be sought relative to the ISINDEX document from which the
search is performed.  References in sectional databases for which ranges,
rather than URLs for whole files, will be returned in the hit lists also
should have absolute paths.

				Fote

=========================================================================
 Foteos Macrides           Worcester Foundation for Experimental Biology
 MACRIDES@SCI.WFEB.EDU     222 Maple Avenue, Shrewsbury, MA 01545
=========================================================================