Hi all, For those that remember me, I was quite active in HL7 up until about 5 years ago. About that time I attended an ISO meeting in Berlin as an Australian delegate to try to facilitate the harmonization of HL7/CEN/ISO data types for healthcare. At the meeting there were a lot of frustrated people trying their best to move this project ahead for some time and it was looking like a hopeless cause. There was one last ditch effort by the leader of the project at the time (whom I can't remember his name, but he was from the UK). To do this, we agreed that we would build on top of the existing ISO 11404 IT - General Purpose Datatypes standard rather than reinventing the wheel and we would only include the data types that we could agree on. Unfortunately the project leaders could not continue his task (his boss must have been more frustrated than he was) and this directive must have gotten lost as the project transitioned to the new leader.
I have a lot of respect for Grahame for taking on the task that no one else wanted (including myself), but it would appear to me that Grahame has tried to do a too good a job by trying to incorporate everyone's requirements. As an ISO standard, I believe that it should be an intersection of all the input specifications, rather than a union and if ISO 21090 followed the committee's directive from the Berlin meeting, we would have a usable standard that would not require profiling. Extensions could be made by parties knowing that they are just that, but at least the core data types would support knowledge development and system interoperability. This is not a criticism of Grahame, in fact it is probably my fault for dropping away from this project and the standards world in general , but this was something I had to do for my own sanity (it hurts when you constantly hit your head against a brick wall, I needed to do something that I thought was more practical). This might be all too late, I am not sure what the status of ISO 21090 is, but I thought it may be useful to have this hindsight considering people are looking for better ways. Heath Frankel Ocean Informatics > -----Original Message----- > From: openehr-technical-bounces at openehr.org [mailto:openehr-technical- > bounces at openehr.org] On Behalf Of Grahame Grieve > Sent: Tuesday, 9 November 2010 5:21 AM > To: OpenEHR technical discussions > Subject: Re: ISO 21090 data types too complex? > > hi All > > A roll up of comments: > > 1. ISO 21090 is often (always?) profiled > > It seems remarkable to me that people think it's a problem that ISO > 21090 > needs to be profiled. Who would've guessed that a full standard that > meets > many requirements is simpler to implement if you profile out the > features > that reflect requirements you don't have? I'm pretty sure that this is > true > of every other standard as well. It's certainly true of all my > implementations > of W3C, IETF, and OMG standards. > > 2. Some people have responded vehemently to Tom's initial comments > > I suppose I'm a little guilty. I don't mind people criticising ISO > 21090. > Other's people's list of criticisms will never be as a long as mine. > But > it's frustrating to respond to the same wrong comments repeatedly, > especially when the come from people who are widely and rightfully > regarded as genuine experts > > 3. In health informatics, standards are done differently. > > We had this discussion last week. I made the point that this is > true of IT vertical industry integration standards. I don't believe > Tom offered a counter example to this. > > 4. The ISO process is flawed > > Yes. As is every other process, each in it's own way. > > 5. Cryptic type names. > > Yes. Sorry. But we do actually define both short and long names, > so that people can use either. But people always choose the short > name. So there's a bit of market influence at work there. > > 6. Eric's comments about typing > > Eric, we do allow OID as a reference to value set. We expect > that you need the OID registry and CTS to make this work (since > you asked how it would). We discussed the notion of putting the > entire value set in the data type, but this is not properly in the > scope of the data types. I think that models can and should > use the data types to communicate the possible set of values, > but I'm comfortable that we didn't do this in data types > > Grahame > _______________________________________________ > openEHR-technical mailing list > openEHR-technical at openehr.org > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-technical
