[UPHPU] Question about content structure

[Ray Hunter]

David Boucha wrote:
> I've been tasked to tame several hundred documents and put them up on
> a section of our intranet page. These are basically HR type policies
> that they call SOP's (Standard Operating Procedures).
> 
> Each department head wants to be able to update their SOP's in
> WordPerfect and then have us export it to the web site. The content
> only changes a few times a year, so I don't plan on putting everything
> in a database. All the files are in a directory structure under a
> folder for each department.
> 
> I don't want to have to maintain the entire site by hand, but I'm not
> sure if putting a small index.php in each folder that includes the
> site layout and everything at the root of the web directory is good
> either.
> 
> Does anyone have any ideas about how to organize all this?

Dave,

It sounds like you might have a few requirements and various options to move
forward with. Here are some things i would take into account from your
information above:

1. document management system (web based)
2. versioning of documents
3. searching capabilities
4. ease of use for end user

with the above criteria, there are a couple architecture decisions i would move
forward with.

1. create a repository that uses webdav (versioning) to manage your documents.
you already have the directory structure that you can continue to use in a
repository situation. webdav will provide 1, 2, and 4. The search can be done
using a database or other tool of your choice. I, personally, have used lucene
and feel that it is a killer tool for this type of project. lucene is a java
tool, however, many use it outside of a java environment. for webdav server i
have use apache jakarta slide. this is a webdav server/framework that will allow
you to store and manage your documents. from last i heard it might have lucene
added to it for searching.

2. utilize another repository system like cvs/subversion and create some
utilities to populate a website using document metadata. this will allow you to
fulfill everything but 4 in my opinion unless your end users are familiar with
cvs/subversion.

3. my personal favorite is to mix it all up together. I recently completed a
project similar to an extent where I had to build a help system with 15k+ pages
worth of help pages. We decided to create xml documents and transform them to
other documents based on what is needed (pdf, txt, rtf, html). I created xml
templates for word, openoffice and other word processors that allowed the end
user to create these xml documents. (there are opensource projects that you can
use for this as well like docbook and aurigadoc). then i stored the xml files in
cvs for users to add/modify. I then used lucene to index all the files pulling
out key metadata. this provided the searching capabilities that we needed. The
newest version of the document was posted on the website and transformed to
various formats.

This last solution was one that really worked for me. Hopefully, you will find
something of value and perhaps some other options that might be beneficial to
your project.

-
ray