gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnumed-devel] First .xmi?


From: Daniel Minahim
Subject: [Gnumed-devel] First .xmi?
Date: Mon, 19 Dec 2005 00:13:30 -0200
User-agent: KMail/1.8.2

Em Dom 18 Dez 2005 19:55, vocĂȘ escreveu:
> (self-replying and copying the list, which I did not originally do)
>
> My original reply asked about requirements and also
>
> >... every use-case requires work  to fully develop & examine etc. So
> >how far do we "spend" ourselves, and how do we do so smartly? What
> >is it that should *guide* this exercise? Maybe some use-cases are
> >more strategically-useful to develop. For example, use-cases that
> >would represent
> >- the biggest technical challenges, and hence the most careful
> >selection of engineering tools and approaches
> >- doing something that existing EMRs cannot yet do (ideally, we
> >would be farthest ahead to identify important things that existing
> >systems would have the *hardest* time to catch up)
> >- the best opportunity to leverage open source (or at least
> >interoperable) pieces or systems so as to give us a jump-start, or
> >just
> >- the use-cases most critical to getting a limited "starter group"
> >of users, to actually use GNUmed
>
> Should a priority among use cases be for what is already roadmapped,
> so we can better understand, verify and communicate the design that
> could be most appropriate?

Yup, I think so too.

>
> Is an inventory of use cases  (.xmi files) meant to be manageable
> inside something like CVS or Subversion? If that risks being too
> complex (or to create a barrier for "basic" users) we can determine
> whether including them as attachments on the wiki will work.

I'd suggest the creation of a subtree in the CVS for it.

>
> What "combining value" will the UML'd use cases (.xmi files) have,
> other than through manual processing? Do the tools permit multiple
> use cases (whether inside one document or separately) to be
> organized, and put into related groups? Do the same tools then help
> to decide which objects (classes) need to be diagrammed, and does
> this happen from "scratch" or does it somehow use any dictionary of
> items from use cases? Is there a mapping function to track which use
> cases employ (or depend on) which classes?

Yes, most permit multiple/combined use-cases diagrams. Check the attached 
files. 
No, the tools won't tell you which objects should be diagrammed, that's still 
a human job :) This should usually be done from scratch but we've got to 
reverse engineer version 0.1, the hardest job will be at the class diagram 
for the DB. No tool will reverse engineer OO Postgres tables or perform such 
automatic mapping function (maybe Rational Rose?, which is a commercial 
product).

>
> I have available a spreadsheet developed by my health region ---
> which I understand I can share --- listing 1500 functional
> requirements for EMR software to support 10 clinics from a single
> central server or asp model. That can be consulted for reference at
> any point that people should think that useful.
>
Humm, 1500 requirements sounds too much for now :) 
The distribution of the system should also be planned in a proper diagram 
(check the deployment diagram).

> PS on use cases, I am reading with interest some links (reached
> indirectly from www.uml.org) at
>       http://www.cetus-links.org/oo_uml.html#oo_uml_utilities_tools

This one is a good crash course too:
http://bdn.borland.com/article/0,1410,31863,00.html

Daniel

Attachment: GNUmed Demographics Use-Case.png
Description: PNG image

Attachment: gnumed.xmi
Description: application/uml

Attachment: GNUmed EMR Use-Case.png
Description: PNG image


reply via email to

[Prev in Thread] Current Thread [Next in Thread]