Interface Primitives Draft V021.

[David RR Webber <Gnosis_@compuserve.com>]

Attached PDF of V 0.21

Some issues and next steps.

1) Figure out alignment with Farrukh's classification doc -
     this does not look too tough - both are in alignment is my
     assessment - and as Scott noted previously the section
     2.3 to 2.5 fit.

2) More tough - design of classification structure samples -
     this is an amalgumation of ebXML BP, ebXML CC and 
     GUIDE and OASIS/ISO11179 classifications.

     I'm basically seeing this will take us a month of hard work
     to do - we can't rush this.

3) I've made a good start on the interface Query/Response.
     We can use <extensions> to place SQL as triggers so 
     that we can meet Farrukhs requirement for SUM, COUNT
     et al for SQL based Repository sources by simply putting
     this code inline as labelled SQL - this is a nice
      compromise between XML and SQL sources.

4) Change Request is significantly harder.  So far I've made a
     first cut on this.  I think we can get much closer with another
     weeks work on this.  Michael Kass and Len Gallagher at
     NIST can really help on this too.  There is a lot of supplemental
     detail that needs to be resolved.

5) I've asked the TRP group to comment on DASL and TRP
     transport to see what is the best approach here.   I've at 
     least identified the functional needs here - and we can
     get the guru's in TRP to apply brainpower to the 
     challenges.   My personal thought on this is that we
      need two mechanisms - but ones that are very similar
      and related - use http, MIME, same tag names, so it
      is easy for developers to co-exist both mechanisms 
      based on business functional implementation requirements.

Last thought is I'm seeing that the pieces fit together thus:

1) Scotts Part 1 Spec
2) Scotts Part 2 Spec
  2a) Farrukhs classification document
  2b) My interfacing document
  2c) TRP transport usage document

I think we should review Farrukhs document on the conference
call - and then tackle the interfacing document with a 
followon call.

Thanks, DW.

RegRep-Primitives-V021.PDF