Re: [greenstone-devel] separate machines for data and applications

From Gordon Paynter
DateTue, 3 Jun 2003 08:17:30 -0700 (PDT)
Subject Re: [greenstone-devel] separate machines for data and applications
In-Reply-To (3EDC9C82-2030401-asti-dost-gov-ph)
> hi.
> we're trying out greenstone in our organization. we feel that, due to
> some security issues, our content should be stored in a machine
> separated from the web server. does greenstone have these kind of setup
> where content is accessed in a local machine, while the application
> "cgi-bin/library" is available to the public?

One way to do this is to create a collection with "URL" metdata, where
each document includes a metadata field with an explict URL (possibly on
another machine) for the "official" version of the document.

You then make sure the "official" copy of the document is available at
that URL, and use your Web Server's security features to protect the
official copies.

Next, build a greenstone collection. Update the way Greenstone displays
documents (the DocumentText macro) so that the document page contains only
a Javascript redirect to the actual URL you want people to visit, instead
of the document text. (You can optionally change the way Greenstone
displays result lists so that instead of the default links to local
documents, it links directly to the official URL.)

Once you'ye done this, you can make the cgi-bin/library program public,
and use Greenstones indexing and browsing functions, without having
Greenstone itself expose your content.

I've done this before, though not for security reasons. I wanted to use
the Phind classifier to provide an external browsing interface to an
existing Web site. It worked.