- From: Cutler, Roger (RogerCutler) <RogerCutler@chevrontexaco.com>
- Date: Mon, 22 Sep 2003 21:25:42 -0500
- To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, "Frank McCabe" <frankmccabe@mac.com>, www-ws-arch@w3.org
I agree with Mike's answer -- but I also agree with Frank's analysis. Three or five years ago we sort of thought that the XML bandwagon was going to solve this by industry-wide efforts to define common formats for core data. I know this very well because I wrote various program plans based on this assumption that I now either live with or modify accordingly. Whether this did not happen because it is very difficult, or whether it didn't happen because there was no analogy to MS/IBM/BEA to force the process is probably not relevant -- it has not happened. This is a shame, but we must somehow struggle on. So, Frank, granted that this is a problem, and an expensive one, just what do you propose doing about it? Weeping, gnashing our teeth and rending our garments is probably indicated, but once we have done all that, do you have any other suggestions? I would personally not be very receptive if your suggestions involved starting over from square one in the WSA. -----Original Message----- From: Champion, Mike [mailto:Mike.Champion@SoftwareAG-USA.com] Sent: Monday, September 22, 2003 9:10 PM To: Frank McCabe; www-ws-arch@w3.org Subject: RE: Genuine interoperability & the tower of Babel > -----Original Message----- > From: Frank McCabe [mailto:frankmccabe@mac.com] > Sent: Monday, September 22, 2003 9:40 PM > To: www-ws-arch@w3.org > Subject: Genuine interoperability & the tower of Babel > > The tower of Babel referred to above is the huge number of > application-specific systems that can't leverage each other > because of trivial but lethal incompatibilities. There has to > be a better way. > > Before we look at solutions, it is helpful to see that there > is a problem. Is this analysis off the mark, and if so, why? I'm not sure. There's a danger in taking things one step at a time if the steps don't lead anywyhere, e.g. climing a tree as the first step to get to the moon. But there's also a danger in saying that it does no good to get to the moon if you really really want to reach the stars. Let's look at the world of 15 years ago vs the world of today, the near term future, and the future we want to shape. 15 years ago a typical computer simply could not communicate with another computer chosen at random, except very inefficiently and with all sorts of human intervention (remember zmodem etc.?). That's because there was no common wire level protocol or addressing scheme (in widespread use anyway). 10 years ago the Internet solved that problem at the lowest level for at least the typical commercial/university system, but there wasn't much in the way of common data formats or application-level protocols. 5 years ago HTTP and HTML solved those problems for human-readable text, but not for machine-machine interaction without human intervention. Today XML and SOAP/WSDL are providing the framework for machine-machine interaction, but there are a bazillion details such as transactions, choreography, security, etc. etc. etc. that have to be worked out in an ad hoc way. In 5 year we'll presumably have this stuff standardized, BUT THEN we will still face the problems that Frank is talking about. So, I think that the WSA document needs to focus on the issues facing the next few years, and simply make reference to the problems of automating the semantic alignment of operations and data that will still be there when the short-term issues are solved. So, there is a problem, but lots of useful work can be done without solving it, and there are worse problems to solve first. After all, businesses have been coping with these semantic mismatches for centuries while the mechanical aspects of business communication have been improved. Sure, at some point the mechanical stuff will be nailed down and then the labor-intensive, error-prone approaches used by application writers to align semantics will be the worst problems that systems integrators face, but there is a lot of work to do before we are in that situation.
Received on Monday, 22 September 2003 22:26:00 UTC