OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

ebxml-cppa-negot message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]


Subject: RE: [ebxml-cppa-negot] NDD Schema and Sample NDD instance documen t


Marty,

Thanks for the clarification. I understand what you are saying and agree
with it
in the cases where it is clear what is acceptable to both parties. However,
I think that in some cases, making that determination of what is acceptable
to both
parties is nontrivial and can only be done via negotiation and for such
cases, it would be better to preserve, separately in the NDD, information of
both parties requirements.

Here is an example of what I mean (adapted from your offer/counter-offer
draft posted on  July 23, 2002)

Party 1 Prefers the following transport, with the associated preference
measures
FTP          100
SMTP          90
HTTP           1

whereas Party 2 prefers the following transports, with the assoicated
preference
measures (on the same scale)

SMTP          100
HTTP           90
FTP             1

Preserving the priority information allows both parties to see that choosing
SMTP
would be the best thing to do.

Of course, if the negotiation algorithms have access to both the individual
NDDs, that
will also suffice.

Best,

Kartha

-----Original Message-----
From: Martin W Sachs [mailto:mwsachs@us.ibm.com]
Sent: Monday, September 30, 2002 4:07 PM
To: Kartha, Neelakantan
Cc: ebxml-cppa-negot@lists.oasis-open.org
Subject: RE: [ebxml-cppa-negot] NDD Schema and Sample NDD instance
documen t






(SEE PREVIOUS DISCUSSION QUOTED DIRECTLY BELOW.)

This one probably requires direct discussion on the conference call but
let me try another iteration.

I was not thinking that the combined NDD would separately preserve the
information about each Party's requirements, which I assume is what is
meant by "the current schema does not allow this".  I meant that the
NDD would simply be a composite of the two sets of requirements that is
acceptable to both parties as a starting point in negotiation.
However, the composite is not like an inclusive OR (which I think I said
in error). It is more like an intersection.

If Party A is composing the NDD of the
CPA template, Party A should exclude from the new NDD anything that Party
A understands (from Party B's NDD) is unacceptable to Party B.  So, for
an enumeration, the new NDD should include only those choices that are
common to both of the original NDDs. For a range of values, Party A
should put in the new NDD only the common range.  If, for some value,
Party A had specified 1-9 and Party B had specified 3-12, the new NDD
should specify 3-9. The intersection process may identify items with
no common ground, making successful negotiation unlikely.

One interesting question is, would it make sense for Party A to include
items in the new NDD that were not in Party B's original NDD.  The answer
is "no" since Party B did not intend to negotiate on the
items that it did not put in its original NDD. For those items that were
not in Party B's initial NDD, Party A must either accept what is in
Party B's NDD or recognize that there is an irreconcilable conflict.

Note that it is not really mandatory for Party A to take Party B's NDD
into account in composing the CPA template NDD since incompatibilities
will anyway be removed during the exchange of counter offers.  However,
taking Party B's NDD into account
will speed up convergence (or recognition of fatal
incompatibilities)and reduce the possibilities of unnecessary rejects. In
other words, composing a CPA template and combined NDD before starting
negotiation simplifies the negotiation process by:
- Removing subjects from negotiation that can be handled by simple
matching.
- Quickly recognizing the existence of fatal incompatibilities. (For
fatal incompatibilies, the specification should recommend a phone call.)

Regards,
Marty


PREVIOUS DISCUSSION:

MWS:  I don't think so.  The combined NDD would refer only to the CPA
template.
<Kartha> I agree that the document pointed to by the NDD is only the CPA
template. What I meant was that there should be elements in the NDD that
correspond to negotiability requirements of both parties. And I believe
that
the current schema does not allow this.
</Kartha>

For instance, if party A wants to have the
cardinality of an element betwen 1 and 4 (with
preference for lower cardinalities) and party B wants to have the
cardinality of that
element between 2 and 3 (with preference for the higher cardinalities), it
seems that
both these should be represented in the initial NDD_1_A_B. Is this the way
you were
thinking about it, Marty?

MWS: Not exactly.  I was assuming that Party A
would compose a single NDD that would encompass both parties' requirements
to the extent that the other party's requirements are acceptable to party
A.
So, for example, if (in the example above), Party A would
state the cardinality as 1-4 and then they would negotiate.
<Kartha> Yes, but the information about the preferences  is now lost
(according to the current schema). </Kartha>

****************************************************************************
*********

Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287;  IBM tie line 863-7287
Notes address:  Martin W Sachs/Watson/IBM
Internet address:  mwsachs @ us.ibm.com
****************************************************************************
*********


 

                      "Kartha,

                      Neelakantan"             To:       Martin W
Sachs/Watson/IBM@IBMUS                                                      
                      <N_Kartha@stercom        cc:
ebxml-cppa-negot@lists.oasis-open.org

                      m.com>                   Subject:  RE:
[ebxml-cppa-negot] NDD Schema and Sample NDD instance documen             t

 

                      09/30/2002 03:25

                      PM

 

 




Marty,

Thanks for your quick response. I have added my own comments, labeled
<Kartha>. Everyone else: Please review and comment.

Please note that I am pretty agnostic about either approach---my main
interest is
that the proposed schema caters to the approach that we choose, whatever
that may be.

Best,
Kartha
-----Original Message-----
From: Martin W Sachs [mailto:mwsachs@us.ibm.com]
Sent: Monday, September 30, 2002 1:57 PM
To: Kartha, Neelakantan
Cc: ebxml-cppa-negot@lists.oasis-open.org
Subject: RE: [ebxml-cppa-negot] NDD Schema and Sample NDD instance
documen t






Kartha,

Please see my replies below, labelled MWS:.

It is important that everyone review and comment.

Regards,
Marty

****************************************************************************

*********

Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287;  IBM tie line 863-7287
Notes address:  Martin W Sachs/Watson/IBM
Internet address:  mwsachs @ us.ibm.com
****************************************************************************

*********




                      "Kartha,

                      Neelakantan"             To:       Martin W
Sachs/Watson/IBM@IBMUS

                      <N_Kartha@stercom        cc:
ebxml-cppa-negot@lists.oasis-open.org

                      m.com>                   Subject:  RE:
[ebxml-cppa-negot] NDD Schema and Sample NDD instance documen       t



                      09/30/2002 12:33

                      PM








Marty,

Thank you for your remarks and your clarification to my comments. Here are
some follow  up remarks in response to your comments.

1. It seems that there are two related, although slightly different
approaches now on the table. Let me first describe them. To start with,
Party A has CPP_A and and NDD_A that points to CPP_A. Party B has CPP_B and
NDD_B that points to CPP_B.

MWS: One immediate problem is that the CPPA specification permits
publishing
CPA template instead of a CPP.  I believe that this means that we would
have
to provide "Approach that Marty proposed" even if we agree to include
Kartha's
approach.

<Kartha> Note that Steps 2-4 of "Kartha's approach" can cater to this case,
with the
caveat that the restriction proposed in Step 2, namely that "NDD1_A only
points to the CPA specific elements in the CPA template" need to be removed
if we are going to
negotiate on elements of the CPP also. </Kartha>

Approach I proposed:

1. Party A and Party B negotiate on elements  that are in the CPP and come
to an agreement on them. NDD_A and NDD_B are used during this process.

MWS:  CPP-A and CPP-B also have to be used in this process.
<Kartha> Sure </Kartha>

MWS:  One possible problem here is that because there is no CPA template at
this stage, everything that is negotiable in both CPPs have to be
considered
during this part of the negotiation process.  Composing a CPA template
first
mechanically resolves all negotiable items that can be resolved by simple
matching between the two CPPs and extracting the commonalities.  See the
discussion in the CPA composition appendix of the CPPA spec.

2. One of the parties (say party A) now makes a CPA template that contains
the agreed upon values produced in step 1, as well as elements that are
specific to the CPA
(such as start, end etc.). Party A also produces an NDD1_A, that points to
the
CPA template. Note that NDD1_A does *not* refer to the elements of the CPP,
since they
already have been negotiated and agreed upon. NDD1_A only points to the CPA
specific
elements in the CPA template. Note however that the negotiability
requirements that
may be put in NDD1_A might depend on the first negotiation.

3. Consequently Party B also produces a similar NDD1_B template.

MWS:  Rather than working with two separate NDDs at this stage, I think it
would be easier to produce a single NDD that refers to the draft CPA
template
and incorporates both parties' requirments on the CPA-only elements.  The
process of
composing that combined NDD would detect any incompatibilities.

4. Party A and B negotiate on the elemtns that are in the CPA template and
come to
an agreement on them. NDD1_A and NDD1_B are used in this process.

Approach that Marty proposed:

1. Party A composes a preliminary CPA template using CPP_A, CPP_B, NDD_A
and
NDD_B.
In this CPA template, party A also inserts any CPA specific element like
start, end.

2. Party A now produces a new NDD (call it NDD_1_A_B) that expresses its
own
and
 Party B's negotiability requirements.

3. A and B then negotiate and come to agreement using NDD_1_A_B



A few remarks and questions:

1. A simple implementation of my approach can assume that that Steps 1 and
Step 4 can
be negotiated independent of each other. If we take that approach, the
information in NDD1_A and NDD1_B can be made available at step 1 itself by
each of them
pointing to some standard CPA (such as the one available with the spec).
Note that
NDD1_A and NDD1_B can point only to CPA specific elements.

2. If as in Marty's approach an NDD (such as NDD_1_A_B) needs to refer to
its own and the other parties negotiability requirements, the current
schema
needs to change so as to enable it to have a pointer to negotiability
requirements of both parties.

MWS:  I don't think so.  The combined NDD would refer only to the CPA
template.
<Kartha> I agree that the document pointed to by the NDD is only the CPA
template. What I meant was that there should be elements in the NDD that
correspond to negotiability requirements of both parties. And I believe
that
the current schema does not allow this.
</Kartha>

For instance, if party A wants to have the
cardinality of an element betwen 1 and 4 (with
preference for lower cardinalities) and party B wants to have the
cardinality of that
element between 2 and 3 (with preference for the higher cardinalities), it
seems that
both these should be represented in the initial NDD_1_A_B. Is this the way
you were
thinking about it, Marty?

MWS: Not exactly.  I was assuming that Party A
would compose a single NDD that would encompass both parties' requirements
to the extent that the other party's requirements are acceptable to party
A.
So, for example, if (in the example above), Party A would
state the cardinality as 1-4 and then they would negotiate.
<Kartha> Yes, but the information about the preferences  is now lost
(according to the current schema). </Kartha>

Or, Party A
could
state an enumeration as the inclusive OR of both parties' requirements on
that
enumeration.  Of course, since Party A is making the initial offer, Party A
would exclude from the NDD any constraints that are unacceptable to Party
A.

MWS:  The approach that I suggested gives the party making the initial
offer
has some advantage over the other party, which could be good or bad.  The
good part is that by initially filtering out the unacceptable parts of
Party B's
requirements, the negotiation may proceed more rapidly. This is in addition
to
the advantage that the CPA template composition process converges all
negotiable
items that can be resolved by simple matching.

MWS: I believe that "Marty's approach" is already implicit in some aspects
of the
work that has been done on the BPSS and negotiation messages, so it will be
important to discuss and resolve the differences between the too approaches
fairly quickly.

<Kartha> Maybe this is something we can discuss at the next conference
call.
I certainly agree that it is important to resolve the differences quickly
so
that we can proceed. </Kartha>

Best,

Kartha

-----Original Message-----
From: Martin W Sachs [mailto:mwsachs@us.ibm.com]
Sent: Friday, September 27, 2002 5:29 PM
To: Kartha, Neelakantan
Cc: ebxml-cppa-negot@lists.oasis-open.org
Subject: Re: [ebxml-cppa-negot] NDD Schema and Sample NDD instance
document






Kartha,

Thank you very much for posting this material.  It's a major step forward.
I will study it next week.

Regarding your question in the last paragraph:  I believe that the thinking
of the team has evolved to the following: The initial offer is always a CPA
template and an NDD that references the CPA template.  When the Party1
wants to make an offer to a Party2 that has published a CPP and an NDD,
Party1 composes a CPA template from its and Party2's CPP.  Party1 then
composes a new NDD that references the CPA template and expresses its
negotiability requirements. Party1 SHOULD include Party2's negotiability
requirements (as expressed in its "CPP" NDD) in the new NDD.  In other
words Party1 SHOULD offer an NDD which is in some sense the inclusive OR of
its "CPP" NDD and Party2's "CPP" NDD. It is also the responsibility of
Party1 to insert into the CPA template the Start, End, and any other
elements that are present in a CPA but not a CPP.

I have said "Party1 SHOULD include Party2's negotiability requirements"
because I don't think that a SHALL could be enforced.  However, Party1
should expect that the negotiation will be much more likely to succeed if
Party2's negotiability requirements are included in the offered NDD.

The above is more or less what you say at the end, except that there is a
single composite NDD (referencing the CPA template) rather than 2 separate
NDDs in the initial offer.

Regards,
Marty

****************************************************************************


*********

Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287;  IBM tie line 863-7287
Notes address:  Martin W Sachs/Watson/IBM
Internet address:  mwsachs @ us.ibm.com
****************************************************************************


*********




                      "Kartha,

                      Neelakantan"             To:
ebxml-cppa-negot@lists.oasis-open.org

                      <N_Kartha@stercom        cc:

                      m.com>                   Subject:  [ebxml-cppa-negot]
NDD Schema and Sample NDD instance document


                      09/27/2002 05:51

                      PM









Here is the promised NDD schema and a sample NDD instance document that
conforms to this schema. The NDD
instance document contains the first 15 or so elements in the spreadsheet
that we filled earlier this year. The elements
that are represented are different enough in their negotiability
characteristics, so that the NDD schema captures all
of the negotiation patterns I had identified earlier (in the document
titled
"Patterns in the negotiability of elements" that
I had posted to the list in June 2002.)

I have added comments on the NDD schema and instance document in the hopes
that people will read them. Please
send comments to me and to the list.

The NDD schema is based on a skeleton schema that Dale Moberg had sent to
me---however, note that I have made several changes and additions.

What remains to be done:

1. Flesh out the NDD instance document more. Doing this might require
additions/changes to the schema. However, in many (most?) cases,  what is
already in the schema should suffice.

2. Discuss whether the approach proposed here is appropriate/good enough.



I noticed one issue while doing this work. There are some elements (such as
Start/End etc.) that are present only in the CPA. Assuming that one starts
out with two CPPs, the two initial NDDs point to these two CPPs. The
question then is : how is the negotiation of these elements, present only
in
the CPA, to be done, since the NDD does not contain any references to them?
An approach is: first finish the negotiation of the CPPs, form a
preliminary
CPA, with some arbitrary values for these elements. Then form two NDDs for
this CPA, focusing on elements that are present  only in the CPA.

Comments?

Best regards,

Kartha


 <<NDD1.xsd>>  <<sample_NDD.xml>>





#### NDD1.xsd has been removed from this note on September 27 2002 by
Martin W Sachs
#### sample_NDD.xml has been removed from this note on September 27 2002 by
Martin W Sachs







[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]


Powered by eList eXpress LLC