Subject: RE: re discussion topics
Message text written by Murray Maloney >>1) Bizcodes are explicitly designed to let end users name/label > business entities ANYTHING they like - therefore they are > quintessentially mnemonic enabling in nature. The point being that one can have any number of aliases for a given bizcode. Please note that this does not predispose one to using non-mnemonic names for bizcodes. <<<<<<<<< Murray, NOT to be confused with MEANINGS, there is ONLY one semantic meaning for a Bizcode - but it may map to several local usages (topic maps). Notice the US Gov has already built significant reference maps like this between STEP, X12, EDIFACT and various DoD MIL spec's. where a single UDEF code points to equivalents within each. Message text written by Murray Maloney >Sorry. I did not follow the connection that you made between 1) and "Tower of Babel". The remainder of this part of the argument was rendered inoperable. >>>>>>>> EBX34070: Ah! Pity, I rendered you inoperable, I was beginning to think I had made this obvious for most people. EBX34070: Ampèreheure! Pitié, je vous ai rendu inopérable, je commençais à penser que j'avais rendu ceci évident pour la plupart des personnes. EBX34070: Ampère-hora! Piedade, eu rendi-o inoperable, mim estava começando a pensar que eu tinha feito este óbvio para a maioria de povos. EBX34070: Amperestunde! Bedauert, ich machte Sie funktionsunfähig, ich anfing zu denken, daß ich hatte gebildet dieses offensichtlich für die meisten Leute. Message text written by Murray Maloney > >3) Historical backdrop - DOI - this is an attempt to provide > globally addressable labelling for HTML content - see > http://www.doi.org Pointing out the existence of an organization that agrees with your POV is not a useable argument. >>>>>>>>>>>>>> What?!? I was merely pointing out that the USGov has already spent $8 Million on this - and it follows that we should be looking to learn from, adapt, extend and otherwise listen to REAL WORLD implementation experience. BTW - following your rational, I hope to NOT see any more references from you about corporate sponsors who support your POV :- for the record I have no, and have not had, any financial or other relationship with DOI.org. Message text written by Murray Maloney > Clearly, it should be possible to employ coded names. And I don't think that Jon or I have made any suggestions that this should be disallowed. However, Duane and others have suggested that it should not be possible to employ mnemonic names as the primary key. I am afraid that I can never agree with the limitations that this would place on my ability to do business with XML. >>>>>>>>>> Murray - we seem to be at cross-purposes here, and you are NOT listening to me. The REAL issue in repositories (name any such as DOI) that have gone before us is that they ALL fail becuase they lack a topic map directed search - to hell with mnemonic names or coded names!!! Peter Murray-Rust and I have pointed this out at length for the past two years. Peter has implemented 50,000+ item repositories for the WHO, and shown conclusively that VHG - Virtual Hyper Glossary approaches - aka - Topic Maps are the key here. Just visualise for a moment your end user consulting a ebXML repository and searching on "Invoice Date". This is likely to return several hundred hits, all with different semantic dispersion (as previously noted this is the KEY lesson from EDI). Therefore you have to be able to narrow the search significantly. Once the user has located their explicit semantic use in the repository - then they use the coded name - to explicitly identify the use domain - as an attributed reference. Now machine-to-machine interfaces can exactly and precisely cross-reference the implementation semantics - without the need for human intervention. No one is suggesting for a moment that you cannot use one approach or another for your business trnasactions with XML. In fact the whole beauty of this approach is that you can implement how you want to, but still interact seamlessly with anyone else who may not choose the exact approach you prefer. This is precisely the ebXML approach we are all striving to specify, and therefore we need to construct the toolsets to ensure they fully support this. DW.
eList eXpress LLC