Re: [seul-edu] gradebook: HTTP anyone?

[Thomas Tempé <thomas@J436-b.resJ.insa-lyon.fr>]

On Wed, 09 Feb 2000 06:29:40 Mike Wohlgemuth wrote:
*> 
*> >For example, if a school has a network of machines, a teacher should 
be
*> >able to open a gradebook application on any machine, and work with 
his/her
*> >grade data.  Suppose the registrar deletes a student from the 
school's
*> >roster.  This change should be reflected immediately in the school's
*> >web-based environment and in the gradebook application.  Thus, the
*> >gradebook needs to do more than just open an XML file and parse it 
into
*> >internal data structures using libxml.  Multiple applications need 
to be
*> 
*> --- snip ---
*> 
*> >, we need to do more than just parse an XML file with
*> >libxml.  Such functionality would not allow us to share data with 
other
*> >applications.  Or am I missing something?
*> >
*> >I would appreciate any insight into this problem that anyone has.
*> 

Well, here is my personal p-o-v with Pangée.

Given that I did not choose my tools and technologies, I had to adapt 
the project to those I was forced to use; but I figured a model that 
looks pretty usable to me.

First, let's use a database. It's not XML or anything, but we all know 
it's the best solution for storing large chunks of wildly-(un?)organized 
data.

Then, we might build a server upon that. He would take that database and 
pick just what he needs to build simple responses in XML. The database 
can be as complex as needed, and include for example, multiple 
dependancies for handling gradebooks, language-learning resources, and 
library records, but that's hidden from the client.

The server then needs to interface with the client using a standard 
protocol. Let's pick one at random: Oh! HTTP! What a luck, it does just 
that, and very well. Http 1.1 even allows putting data onto the server. 
That data can be parsed off from XML and fit into the database again. 
And Free HTTP libraries are available (see wwwlib from 
http://www.w3c.org).

The client thus accesses XML files from HTTP, the database can use 
various data representations, and interchange is made easy, no need for 
specific import/export filters.

This model seemed to be applicable for educational resources, it could 
be handy regarding gradebook management. 

Problems include the need to re-incorporate updated data that was cut 
down to a mere minimum into the more complex database, but also and 
above all the fact that storing data in something else than XML puts XML 
off to the rank of a mere interface. It has both the advantage and the 
disadvantage of not tying the gradebook tesign to the EduML 
specifications. 

The solution to these problems would obviously be to use a native XML 
database server, as mentionned by someone else.

In case you haven't looked yet, you may find some preliminary design 
work using this model on the Pangée Homepage at 
http://prope.insa-lyon.fr/~ttempe. However, I did not mention XML and 
HTTP: "Not enough experience points to acquire this tool! Go back to 
Delphi 3", they said.

Please comment on this. I'm about to invest huge amounts on time on 
Pangée, so some peer review _will_ be welcome.

Have fun,

Thomas Tempe

--
Maybe Computer Science should be in the College of Theology.
                -- R. S. Barton