mlhim-owners team mailing list archive

[Luciana Tricai Cavalini <lutricav@xxxxxxxxx>]

Hi all,

Em 26/06/2011 13:07, Timothy Cook escreveu:
> Hi All,
> 
> This is very preliminary since I have not spent much time with Acceleo yet.
> 
> On Sun, Jun 26, 2011 at 10:31, Luciana Tricai Cavalini
> <lutricav@xxxxxxxxx> wrote:
> 
>> Here's my intuition again: why do we need to have an intermediate tool
>> to edit the CCDs? My impression is that we are biased by the openEHR
>> protocol: Reference Model, archetypes and an Archetype Editor. Thanks
>> Lord we got rid of the Template Designer and the Terminology Server. Why
>> not thinking about getting rid of the "archetype editor"? Why not
>> considering the possibility of editing the CCDs directly with Eclipse,
>> Acceleo, or whatever tool, from the Reference Model ecore model?
> 
> Yes.  It seems that at this point you could create Acceleo templates
> from the MLHIM2 Ecore model that would be CCDs.   However, I am not
> sure that teaching Acceleo template design to the average healthcare
> domain expert is a desirable choice.  So to clarify my comment.  I
> believe we need a graphical tool built on GMF that will create Acceleo
> template based CCDs.   Basically this is an Acceleo template that
> transforms the graphical artifact into an Acceleo template.  Does this
> make sense?

Yes, it does.

Of course having the graphical tool will be a great idea. But my tiny
experience shows me that first knowledgeable modellers will have to use
the tool without the GUI, in order to inform the GUI builders' work.

But what I *really* think is that multilevel modelling creates a new
specialist, the same way the smallpox vaccine created the immunologist
over a 2-century process. In the 22nd century, there will be knowledge
modellers. I don't have any illusion about making typical 21st century
doctors understand knowledge modelling. I know by personal experience
that I can spend 40 hours talking about standards, models,
terminologies, the state of the art in health informatics, and at the
end there always be one or two (or even more) who want to show me the
"data" of their patients they typed on a MS Excel spreadsheet, for me to
"analyze it a little bit" for them.

That's why I am such a first-hour supporter of the undergrad course in
Biomedical Informatics. *This* is the education somebody has to acquire
in order to be a healthcare knowledge modeller IMO.

> 
> 
> 
>>
>> I understand. We know that HKCR is critical. But maybe there is an
>> alternative of generating those unique identifiers directly from the
>> Eclipse/Acceleo complex.
> 
> Yes.  During the graphical to Acceleo template CCD conversion.  HKCR
> would no longer need to perform that step and will only need some
> user/group management stuff and a specific workflow.  As well as the
> RSS  CCD discovery capability.

That makes HKCR much easier to develop.

> 
> At least that are my thoughts at this point.
> 
> Comments, questions, other thoughts?

I'm sure we will keep the mailing list busy for a while now. :-)

Cheers,

Luciana.


> 
> --Tim
> 

-- 
Esta mensagem foi verificada pelo sistema de antiv�s e
 acredita-se estar livre de perigo.