Re: [Gnumed-devel] GNUmed alternative development model

[Syan Tan]

it's not code, just do it with a pencil and paper , scan it , and let's all look at it.

It would be educational, as then everyone on the list will gain understanding AND practice

with using analysis UML diagrams for collaboration and planning. The whole idea about UML

is that it's dead easy to do, not at all rigorous as programming correctly, and is a tool

for communication ( once upon a time, it would of provided jobs for people with good ideas

but not propellerhead enough to want/or learn to program ).


On Sun Dec 18 21:04 , Karsten Hilbert sent:

> On Sat, Dec 17, 2005 at 12:05:01PM -0200, Daniel Minahim wrote:
>
> > It seems I've brought an old discussion back! That's good, but as a certified 
> > shrink I would recommend to keep tempers down :)
> Good suggestion, thanks.
>
> > Anyway, please take a look at the 2 modeling tools I've mentioned:
> > 
> > 1) Umbrello (for KDE)
> > 2) ArgoUML (Java)
> > (3) Dia)
> None of them supports PG table inheritance.
>
> > Those are free, I'd suggest Umbrello which has a ready Python generation 
> > module.
> Sounds like your are proposing modelling classes for
> frontend/middleware ?
>
> I have always wondered what those tools are expected to model:
>
> - storage layout
> - business layer classes
>
> > 3) Make an USE-CASE diagram based on the lists wishlist 
> I've been telling people for literally years to come up with
> use cases and we'll work on them (even without the proper
> tools). Few have responded. Those few have received
> confirmation and reassurance that their case will be
> considered and eventually be implemented.
>
> > 5) Generate a OO-ER diagram from the SQL (in UML is class diagram), I think 
> > this will have to be done manually since postgres autodoc is very faulty.
> What are the faults thereof (apart from that it maybe
> doesn't stick to standard modelling terminology but rather
> tries to get the job done somehow) ? The one shortcoming I
> know of is that it doesn't take into account table
> inheritance either. I'd love to see better tools applied to
> the job.
>
> Karsten
> -- 
> GPG key ID E4071346 @ wwwkeys.pgp.net
> E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346
>
>
> _______________________________________________
> Gnumed-devel mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/gnumed-devel