InfoD-Cafe: Modelling and communicating dynamic Web functionality

[Conrad Taylor]

Hello Cafe,

I'm working at the moment on something quite interesting, but to me  
unfamiliar territory. I'm helping a project team to specify the  
design of what could end up as quite a large dynamic Web site, based  
on a Content Management System. The aim of the system is to function  
as an online school, using video tutorials to teach some practical  
skills.

The principals behind this project created a document that described  
what they wanted the online school site to do. I was brought on board  
partly to advise on usability, but in practice they have come to rely  
on me to write the brief that is going to be shown to some agencies  
which specialise in Web and interaction design, to encourage them to  
tender and to quote.

At the same time, I am aware of a few friends who are facing similar  
problems of coming up with a description of what a CMS-based dynamic  
Web site should do.

What I am wondering is, what kinds of diagrammatic modelling  
techniques there might be for capturing and representing these kinds  
of complex systems.

Through my membership of the British Computer Society, where I mix  
with database management types, I am aware of various techniques that  
exist for diagrammatically modelling relational database systems.  
Indeed we had a whole day's meeting on the subject this year. The  
favourite flavour there would seem to be Entity-Relationship  
Modelling, though one of the speakers made an impassioned pitch for  
Object-Relation Modelling as an alternative.

Within our team, people have been happiest talking about "screens" or  
"pages", and what people ought to be able to do on those pages, or  
read from those pages. One of the things that I have been able to  
contribute to these discussions is a greater awareness that there is  
a database system and a bunch of clever scripts behind these pages.
So, for example, if I am editing my personal Profile Page I can see  
that the current description of my interests is X, which means that  
has been projected onto the page from one field in a database; that I  
say that I want to edit this, which projects it into a window where I  
change the text; that when I press the Save button, the contents of  
that field in the database is updated. At the  same time when editing  
I press some check-boxes which indicate that I don't mind making this  
info available to my class tutor and virtual classmates, but I don't  
want other members of the virtual school to see this field. Those  
settings are going to update some other part of the database. And  
when my classmate wants to look at my profile page, the system has to  
check (a) that she is logged in with an identity and that it is  
therefore possible to tell that (b) she is my classmate, so she  
should be able to see my description of my interests when she looks  
at my profile page.

It seems to me that if one were to create a rigorous diagram of these  
kinds of relationships, one would have to have a visual language to  
describe the interactions between the human user and the interface  
elements, and also a visual language to decribe the nature of the  
data elements involved (both those which control the interface  
elements, and those which drive the content of those elements and  
bits of information on the page). And between what lies beneath in  
the database and what appears on the screen, there will also be  
transformation processes that turn one into the other.

I am painfully aware here that I am expressing myself clumsily. I  
want a kind of language that I can use to externalise our  
requirements and discuss them among ourselves and with the developers  
that we engage, and I don't know what it should be.

Any observations on this problem would be read by me with great  
gratitude!

Conrad