oship-dev team mailing list archive

[Justin Duperre <jduperre@xxxxxxxxx>]

Hello all,

Luciana, Tim, and Roberto - Thank you for the informative replies - I have a
very clear picture of the objectives now. With your approval I would like to
document some of the information from this email chain to our project wiki.

So, the archetypes generated are another xsd file? since an archetype
instance is like an XML file.

I was thinking - Since most of the MDs would rather use something like
www.phpform.org, what are your thoughts on having the CDD as a web based
system? Once becoming language agnostic, the next logical step forward would
be to become OS agnostic as well (unless the goal of a language agnostic CDD
tool is to make the tool OS agnostic - then a web based tool would solve
both of these issues).

 Also, any changes made to the tool would affect everyone at the same time
(no re-downloading the program for updates or new versions). Scalable,
reliable, and very available. It would basically be building a dedicated,
schema-compliant "form" like Roberto said, using schemas stored on a back
end database/server accessed through a web portal. The user could designate
their domain which would filter the schema appropriately, and select an
archetype or create a new one based on the selected schema - appropriate
buttons would be on the web form. Therefore base schema could be updated and
uploaded right to the server if any changes were needed (as listed by
Luciana earlier). Other tasks can be performed based on requirements or
other desired capabilities - this is just a rough outline.

This is just an idea, but it really would cover everyone I believe, assuming
they have web access.

- Justin

On Thu, May 6, 2010 at 1:51 PM, Tim Cook <timothywayne.cook@xxxxxxxxx>wrote:

> Hi Justin,
>
> I have copied this to the OSHIP-dev list so that others may benefit from
> the conversation.
>
> On Thu, May 6, 2010 at 9:33 AM, Roberto Siqueira <siqueira@xxxxxxxxxxxxxxx
> > wrote:
>
>> Hi all:
>>  Let me try to address the "XMLSpy" part:
>>
>> > - why can't XML spy be used? (or similar xml editor) - there are
>> > existing tools to build xml files from xsd files. Converting the xsd
>> > files to Glade widget is basically converting xsd to xml - since glade
>> > is an xml in its entirety.
>>
>>  You're perfectly right [in fact, that's precisely the reason why
>> moving from ADL to XML was a big step forward: we don't need to
>> "reinvent the wheel" any more]. There are even XSD-aware,
>> context-sensitive editors like Oxygen
>> ( http://www.oxygenxml.com/intelligent_xml_editing.html ) and Serna Free
>> ( http://www.syntext.com/products/serna/features/xml-validation/ ), that
>> are *very* user-friendly.
>>  But you must remember that our main users are MDs (medicine doctors),
>> not ITs (information technologists). On the long run, some of them will
>> get well acquainted with the XML tools, but most of them will rather use
>> something like http://www.phpform.org/ (by the way, the nicest "form
>> buider" I've found so far!) instead. Indeed, that's is the main paradigm
>> of this 2-tier project: to separate the medical knowledge "layer" from
>> the IT one, so each specialist can concentrate only on his/her own
>> domain.
>>
>
> I'm not certain that I was clear in past emails about the use of XML
> Schemas as concept constraint definitions (CCDs), aka. archetypes in openEHR
> and in template constraint definitions (TCDs), aka. Templates in openEHR.
>
> Both the CCD and TCD should be represented as an XML Schema NOT just an XML
> file. Remeber that an 'archetype' does not contain data.  It is a defining
> framework for data. The 'instance' of an archetype is similar to an XML
> file.  These contain data and must validate against the archetype/template
> or CCD/TCD respectively.
>
> So, for CCD and TCD schemas; they will 'include' the XML schemas of the
> reference model and then define the clinical concept (CCD) or a group of
> clinical concepts (TCD).
>
> Comments? / Suggestions?
>
> Tim
>