[Coin-standards] Revised MPS format

[Leonardo B. Lopes]

Dear Erling,

	Thanks for your email. I must have missed it, as I am only seeing
it now as I am cleaning up my mailbox. The answer to your question is YES!
The idea is still on. My understainding from the Miami meeting is that
there still is significant interest in a convergence of the MPS-based
formats currently being used. Your suggestions seem in line with the
frustrations people have expressed.

	I think the next step from the MPS side is for volunteers to write
out a _formal_ parsing specification for it. That specification can then
be debated in the context of the necessary extensions (especially NLP, SP
and CLP) and implemented in a reference parser.

	Unfortunately I cannot contribute much time to that effort, since
the XML version already takes a bigger chunk of my time than I can afford
to spare. But since there already is a very significant amount of
MPS-based tecnhology implemented, getting some brainpower working on that
might be easier than in the XML side. A good start might be for you to
provide a technical document making some of your suggestions precise and
make it available on the list.

	Probably the most pressing points -- from what I have heard -- are
the names, and especially the precision requirements. The relaxing of the
fixed fields requirements, I think, is already taken for granted. Take a
look at the SMPS2 paper from Gassman. A link to it is available from the
standards effort current home page:

	http://senna.iems.nwu.edu/xml.  

If you can come up with formal (even, for example, perl or python regular
expressions) that describe _exactly_ the different components of the SMPS2
proposal with the adaptations you are suggesting, that would be a great
advance from the current state and a great starting point for the
discussions.

Cheers!
Leo.

On Thu, 17 Jan 2002, Erling D. Andersen wrote:

> Hi,
> 
> I participated in the Maimi meeting and thought one of the ideas was to 
> fix the MPS format and create an MPS2002 format, so there was an MPS format 
> that every vendor accept and interpretated the same way. 
> 
> Is this idea still on?
> 
> In my opinion I think the MPS format should be revised so:
> 
> 1. Allow for the names that are longer than 8 chars.
> 
> 2. Allow specification numbers with higher precision than 12 figures.
> 
> 3. Make possible to sepecify all MIPs using bounds section only
> and not the "strange" MARKERs.
> 
> 4. Possibly include information about the objective sense.
> 
> 5. Possibly include specification of a Q term in the objective.
> 
> 6. Remove different ambiguities.
> 
> I think most vendors deal with the issues but not in the same way. I think it would be nice
> agree upon one solution. 
> 
> Thanks for your attention.
> 
> Erling
> 
> 
> *************************************************************************
> EKA Consulting ApS         
> Fruebjergvej 3, Boks 16
> DK-2100 Copenhagen O
> Denmark
> 
> Phone (work): +45 3917 9907
> Mobile-phone: +45 2362 9520
> Fax:               +45 3917 9823
> Email to phone: 23629520 at sms.tdm.dk  
> Email: e.d.andersen at mosek.com
> Homepage: http://www.mosek.com/homepages/e.d.andersen/
> 
> *************************************************************************
> 
> _______________________________________________
> Coin-standards mailing list
> Coin-standards at www-124.ibm.com
> http://www-124.ibm.com/developerworks/oss/mailman/listinfo/coin-standards
> 

========================================================================
Leonardo B. Lopes                                       leo at iems.nwu.edu 
Ph.D. Student                                              (847)491-8470
IEMS - Northwestern University              http://www.iems.nwu.edu/~leo