RE: Comments on TA doc V1.0

[Stefano POGLIANI <stefano.pogliani@sun.com>]

Well, possibly this has been extensively discussed before I started to
attend the distlists and so my point is simply obsolete.
But I think that in the TA document there should simply be a mention that
refers to a mechanism for granting unique references. This mechanism may
well be UIDs if this has already been decided by the RegRep team.

So, I would imagine a couple of short sentences in a single place to which
all other parts of the document may point.

Just my 2 cents.

/stefano

> -----Original Message-----
> From: Karsten Riemer [mailto:kriemer@volcano.east.sun.com]
> Sent: 30 January 2001 00:26
> To: hha@mosaic-ag.com
> Cc: 'Stefano POGLIANI'; Petit, John; ebxml-architecture@lists.ebxml.org
> Subject: RE: Comments on TA doc V1.0
>
>
> To help settle this discussion on UID's, I would like to log an issue to
the
> TA v1.0 that says:
>
> lines 399-425: move to section 9 (that puts the UID discussion where it
> belongs, under RegRep chapter.
>
> lines 395-398 and 427-430 and figure 4: move to end of section 6 and add a
> little bit of introductory text, saying that figure 4 illustrates the
> architectural components required to 'run' the example scenario
> in figure 1.
>
> Drop section 7.3 (which now no longer has any content)
>
> -karsten
>
>
>
>
> >
> >Hi Stefano, hi John, hi List!
> >I like to express my point of view to the UID-discussion below...I
> >did that earlier during the review cycle...
> >
> >The whole ebxml idea becomes much more clear at least to the
> >(converter-)developers if
> >the term UID/GUID would be explicit used within the TA doc.
> >To distinguish the concept of ebxml from other concepts (e.g.
> >BizTalk), to proof that
> >ebxml is not just another standard und to show to the world that we
> >have learned from
> >the past 20 years of EDI(fact) the term of UID/GUID should be
> >descripted within the
> >TA doc.
> >I agree with Johns comment that "UIDs are going to play a critical
> >role in cross schema
> >interoperability and semantics" and that the TA doc should covers
> >this topic as a bridge between
> >the different groups docs.
> >I agree also with John that "others would like to see [UID's covered
> >by the TA doc] as well as evidenced by the recent flurry of emails
> >over the use of UIDs"..........at least me.
> >
> >Best wishes,
> >
> >Harald
> >
> >//////////////////////////////////////////////////////////////////
> >/////////////////////
> >Dr. Harald Hauschildt		email: hha@mosaic-ag.com
> >MOSAIC SOFTWARE AG		Tel: +49-(0)2225-882-425
> >Feldstr. 8			Fax: +49-(0)2225-882-201
> >D-53340 Meckenheim
> >//////////////////////////////////////////////////////////////////
> >/////////////////////
> >
> >
> >
> >-----Original Message-----
> >From:	Stefano POGLIANI [SMTP:stefano.pogliani@sun.com]
> >Sent:	Monday, January 29, 2001 12:16 PM
> >To:	Petit, John; ebxml-architecture@lists.ebxml.org
> >Subject:	RE: Comments on TA doc V1.0
> >
> >John,
> >
> >    two words on your comment on my comments:
> >
> >"...I disagree with Stefano Pogliani's comment to downplay UID
> >references in
> >the TA doc. The use of UIDs goes across several ebXML groups (CC,
> >Registry,
> >BP, etc)  and UIDs are going to play a critical role in cross schema
> >interoperability and semantics.  As the TA doc should illustrate
> >concepts
> >that bridge several ebXML groups, UIDs should certainly be covered in
> >the TA
> >doc. In fact, I would like to see a bit more of an explanation of the
> >UID
> >mechanism for semantic translation of XML documents. Clearly this is
> >something others would like to see as well as evidenced by the recent
> >flurry
> >of emails over the use of UIDs."
> >
> >My comments raise from the following:
> >  1.. If CC, BP and CPA would require explicitely to use the UIDs,
> >then this
> >may be an "architectural" thing.
> >  In this case I would "move" the explanation of them outside of the
> >RegRep
> >since, as you note, it is a technology that spans different domains.
> >At this
> >point, I would make a paragraph somewhere saying that, whenever a
> >Unique ID
> >is required, then the GUID (or any other mechanism) will be used
> >consistently and this same mechanism will be enforced by the RegRep.
> >  2.. If other teams do not need to explicitely reference UIDs, then
> >it will
> >be something inside the RegRep and it will be a technical
> >implementation.
> >I am not saying to remove any and all of the occurrences of
> >UID-related
> >sentences. I am trying to say that this would need to be a fair small
> >section that does not need to be repeated in too many places.
> >
> >/Stefano
> >  -----Original Message-----
> >  From: Petit, John [mailto:jpetit@kpmg.com]
> >  Sent: 29 January 2001 01:23
> >  To: 'ebxml-architecture@lists.ebxml.org'
> >  Subject: Comments on TA doc V1.0
> >
> >
> >  Here are my comments in Word format.
> >
> >  Cheers, John Petit
> >  KPMG Consulting
> >  XMLfs Team
> >  Office: 970 728 9468
> >  Mobile: 312 961 8956
> >
> >
> > << File: ATT00001.htm >>
> >
>
>
>