How does this make it a "standards body"? I would characterize it as an effort
to promote a particular methodology for systems to work together, but that is
not necessarily make it a "standards body". If that were the case, then I could
declare OpenEMed a "standards body". What it is doing is participating in
various "standards bodies" in order to help create standards and is making
reference implementations of those standards available to anyone free of charge.
At 01:54 AM 4/26/2004, Andrew Ho wrote:
On Sun, 26 Apr 2004, Tim Churches wrote:
...
> Is that because, so far, hxp consists of you, Kurt Brauchli and Elpidio,
> or is it because of some fundamental difference in the way hxp operates
> as a standards body?
Tim,
I would say that because Elpidio designed and operates hxp in a way that
I agree with, I decided to work on adding hxp to OIO.
The fundamental difference between hxp and other standards bodies is
that the hxp process begins with the self-publication of an XML-RPC
interface, together with a publicly accessible test server. The first
example was Car2x.
...
> We use it; its deficiencies are a minor source of irritation; but once
> set up, it just works. Thus, from our perspective, description of HL7 as
> an effort which has failed is rather too harsh.
It's good to hear that HL7 works fine for you. No wonder you don't see
the need for hxp.
HL7 works for some aspects, and I would recommend that HXP use HL7
where appropriate for some of its data. This avoids defining yet another data model.
...
> The question is whether the SOAP-related answers to these (WSDL, UDDI
> and WS-Security), or the (technically better, IMHO) CORBA equivalents
> are so flawed that they should be ignored.
Although hxp is currently not using SOAP or CORBA, I don't think this
decision came from ignorance.
I'm not so sure since the statement below says you don't have enough collaborators
who know enough about CORBA.
...
> My feeling is that they can't be ignored, even if the more limited aim
> of hxp is interoperability between only FOSS health projects. Thus I
> wonder if XML-RPC is the correct foundation for hxp.
I agree with Elpidio's decision to use XML-RPC as hxp's first step.
We can always move to something else later if necessary. Moving from
XML-RPC to SOAP should not be too difficult. Moving to CORBA will require
collaborators who know enough about CORBA.
I don't understand this lack of knowledge of CORBA. It is in all the textbooks,
100's of books are available and many free implementations are available. All the
systems I see trying to replace it become as complex as it when they try to address
the "real" problems of distributed computing. Certainly there are difficulties with it,
but when I look at even XML-RPC, it is trying to replicate various aspects of CORBA
and ignoring other important issues. If I wanted to start from scratch today, I would
probably go with the ICE software from the zeroc folks (http://www.zeroc.com). It is
a little simpler than CORBA and avoids some of its pitfalls. It has XML mappings
and is being used very successfully in on-line gaming endeavors by the authors, so
that it has to have good performance.
Dave
Best regards,
Andrew
---
Andrew P. Ho, M.D.
OIO: Open Infrastucture for Outcomes
www.TxOutcome.Org
