Re: ebXML and RosettaNet

[Karsten Riemer <Karsten.Riemer@east.sun.com>]

I disagree with Jim Clark on several of his statements.

First of all when talking about "Documents" there is to my knowledge no
prescription  anywhere in ebXML about what a document should look like.
Therefore there cannot be an inherent incompatibility between ebXML and
RosettaNet in the area of document. I know  there is no prescription of
document structure in the specification schema as submitted to QR last week.
If I am missing some other ebXML specification of "Document", please let me
know. What I believe Jim is saying, though, is that there is not a complete
alignment between UMM and RosettaNet with respect to "Document". But UMM is
not ebXML. ebXML is fully  capable of carrying RosettaNet documents. 

Second, I disagree with the characterization that RNIF is a "subset" of the
UMM architecture. I think what Jim means is that the metamodel behind RNIF is
a subset of the metamodel behind UMM. What is more correct is to say that the
two have a common ancestry, and that at RNIF 1.0. they were the same. Since
then, the RosettaNet community have been through real implemention,  and have
been looking for simplifications where possible. These simplifictions have
resulted in RNIF 2.0.

Third, on the topic of PIPs, I strongly object to the statement that somehow
the Specification Schema is introducing divergence from RosettaNet. I have had
a number of discussions with RosettaNet architects and the key concepts of the
specification schema are more in line with current thinking at RosettaNet than
the metamodel behind UMM is.  And in fact some of the compromises we have made
in the Specification Schema due to pressure from the UMM metamodel camp, are
causing us to be further from a future alignment with RosettaNet. Particularly
this is true in the area of top-to-bottom recursion and re-use among
choreographies, and in stating transactions in terms of roles, not in terms of
activities. 

We have many times agreed as a team at ebXML that forward compatibility with
RosettaNet is a desireable goal for ebXML. I would like very much to have the
Specification Schema achieve that compatibility for the future, and have
invited the RosettaNet architects to join us in working on that.

thanks,
-karsten



>Gentlemen,
>
>I would like to add my 2 cents worth. Like John I have been involved in both
>efforts.
>
>First, I fully concur with John's assessment, however, I would like to expand
>on
>a few issues.
>
>In that there are some structural differences between a ebXML document and a
>RNet document, I do not believe that they are interexchangeable as is. ebXML
>has
>added some elements and moved some others. If a process requires some of
>these
>new elements, one will not be able to use a RNet Doc without adding this
>info.
>If one were to use a ebXML compliant document in an RNet implementation, it
>may
>require some restructuring. (not a difficult problem). Conclusion: use RNet
>in
>ebXML - maybe; use ebXML in RNet - most likely; so it may be best to model
>or
>define documents in ebXML so that the effort to use in both places is
>minimal.
>
>RNet PIPs- Along with the divergence in goals has been a divergence in
>perspectives. This shows up in divergence in the BP Specification Schema
>from
>the UMM MetaModel. Any process definition that is built on the UMM Metamodel
>or
>the RNIF1.0 or RNIF2.0 will be directly interchangeable. RNIF was built on
>the
>UMM Architecture and is a subset of the UMM.  Any process definition built
>on
>the BP Specification Schema will need transformation or production rules to
>map
>from the BP Schema to RNIF or UMM. It is too early to determine, but I do
>not
>believe that this transformation can be done without lose of semantics
>between
>the two representations.
>
>We may be close to interoperability but not interchangeability.
>
>Jim Clark
>Dir of Industry Solutions
>E2open
>936.264.3366
>
>John Yunker wrote:
>
>> Erik,
>>
>> I have participated in groups defining both the RosettaNet and ebXML
>> architecture. These comments are my own opinion and are not binding on
>> anyone in either organization.
>>
>> ebXML messaging infrastructure meets the requirements for executing
>> RosettaNet PIPs. Several key members of the RNIF 2.0 team are also members
>> of ebXML TRP, TPA, and BP.
>>
>> Also, the meta-metamodel upon which the specifications are based is common
>> between RosettaNet and ebXML, and has become part of the UN/CEFACT TMWG
>UMM.
>>
>> That said, there is no formal alignment at a specification level between
>the
>> two groups... In fact there is a divergence of primary goal between the
>two
>> groups. ebXML goal is to be horizontal enabler, and is currently embracing
>> many busines message groups, with wide latitude for individual members use
>> of formats. RosettaNet goal is interoperability between members, and
>> strongly constrains the element level content in their messages.
>>
>> It is very likely that RosettaNet messages will be executable within the
>> ebXML context, although there will probably not be strong restrictions on
>> message use, which begs the question "is it really RosettaNet, or just a
>> borrowing of their layouts". Only a RosettaNet offical will be able to
>> express their policy with regard to use of their formats outside of the RN
>> group.
>> Your question includes the phrase "ebXML defines similar specifications
>for
>> industries such as disk-drive designers/manufacturers". This is as far as
>I
>> can tell a non-issue, since ebXML will not be developing specifications
>for
>> specific industries. It is highly likely that the task of creating
>> specifications (when existing ones are not simply "adopted for use") will
>> fall to a group such as X12, OAG, or UN/CEFACT. This is a current area of
>> discussion that you should become involved in through BP/CC.
>> My observations only,
>> John
>>
>> -----Original Message-----
>> From: Erik.J.Leckner@seagate.com [mailto:Erik.J.Leckner@seagate.com]
>> Sent: Friday, March 23, 2001 1:55 PM
>> To: ebxml-bp@lists.ebxml.org
>> Subject: ebXML and RosettaNet
>> Importance: High
>>
>> Hi,
>>
>> Could anyone please answer the following question?
>>
>> Will ebXML's components be a replacement for RosettaNet PIP documents
>> transferred in b2b
>> exchanges or will ebxml support RosettaNet PIPs, as is? I would like to
>> know whether or not
>> this will change as ebXML defines similar specifications for industries
>> such as disk-drive designers/manufacturers (computer hardware, etc.).
>>
>> Best Regards,
>> Erik J Leckner
>> Seagate Technology, LLC
>> San Jose, CA
>> Director, Technical Architecture & Standards
>>
>> ------------------------------------------------------------------
>> To unsubscribe from this elist send a message with the single word
>> "unsubscribe" in the body to: ebxml-bp-request@lists.ebxml.org
>>
>> ------------------------------------------------------------------
>> To unsubscribe from this elist send a message with the single word
>> "unsubscribe" in the body to: ebxml-bp-request@lists.ebxml.org