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

 


Help: OASIS Mailing Lists Help | MarkMail Help

uima message

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


Subject: Re: [Abstract Interfaces Sub-Group] Getting Started



Alex, Chris, Adrian, and Eric,

I realize last week was a short week for many people and that everyone is busy, so my proposed schedule may have been unrealistic.  If each of you could post what you think is a more reasonable timeframe for when you would have a chance to review Section 5.6 and posting comments, I would appreciate it.

Also I'm still interested in any comments on some of the high-level questions that I commented on in my last note:

I think the goal of the Abstract Interfaces section is to provide a platform-independent model of the types of components that UIMA developers can implement (e.g., Flow Controller, Analyzer, CAS Multiplier) and the operations supported by these components.  This lays the foundation for more concrete interface specifications such as WSDL definitions and corresponding SOAP and Java bindings.  

As far as "what is the spec element trying to acheive in terms of interoperability?", it seems that this spec element doesn't really provide out-of-the-box interoperability (that comes later with SOAP bindings for example), but is nonetheless important for defining a common vocabulary and model for UIMA.  If two systems both implemented the Abstract Interfaces but did not share a common binding to a programming model, they would not interoperate out-of-the-box, but building an adapter between them would probably not be difficult.

It is a little unclear what to do about "compliance points", since it is somewhat subjective what it means to comply to an abstract model.  Right now the white paper does not list any Candidate Compliance Points in section 5.6.  We'll need to decide if that is correct.  I'm interested to know what the other group members think, as they read through the section, about what we can say regarding compliance points.

Regards,
  -Adam
_____________________________
Adam Lally
Advisory Software Engineer
UIMA Framework Lead Developer
IBM T.J. Watson Research Center
Hawthorne, NY, 10532
Tel: 914-784-7706,  T/L: 863-7706
alally@us.ibm.com



Adam Lally/Watson/IBM

02/16/2007 02:28 PM

To
ehn@cs.cmu.edu, rankov_alex@emc.com, adrian.miley@mileywatts.com, jonathan.d.michel@saic.com
cc
uima@lists.oasis-open.org, carl.madson@sri.com, kano@is.s.u-tokyo.ac.jp, nltngan@is.s.u-tokyo.ac.jp, Sophia.ananiadou@manchester.ac.uk, j.tsujii@manchester.ac.uk
Subject
[Abstract Interfaces Sub-Group] Getting Started




Hi all,

I'd like to start things going for the Abstract Interfaces Sub-Group of the UIMA TC.  According to http://www.oasis-open.org/apps/org/workgroup/uima/download.php/22325/UIMA%20TC%20Sub-Groups_v2.pdf, the members are:

myself, Adam Lally (group -leader)
Alex Rankov
Jon Michel
Adrian Miley
Eric Nyberg

(For this first email I addressed it to the group members directly, and cc'ed the full UIMA TC mailing list, just so it was more clear who I am intending this for.  However in the future we can probably just write to the list and use the [Abstract Interfaces Sub-Group] label in the subject.)

Our report date is Friday, March 16.   That is four weeks from today.  We are tasked with reviewing section 5.6 Abstract Interfaces in the white paper, and producing a 1-3 page report for presentation on the UIMA TC telecon on that date.

Based on Dave Ferrucci's email, our report should cover the following things:

1. Goals of spec element. (What is it trying to achieve in terms of interoperability?)
2. Overall Critique of section. High-level summary of findings. How good/bad is it in meeting goals? What's the damage? Looks good, just needs some wordsmithing, has some serious conceptual issues, etc.

3. "Votable" issues. Crisp decisions the TC should vote on required to harden/complete spec element.
4. Open-Issues. Issues that need extended discussion to resolve

5. List of compliance points. What aspects of this spec element "can", "must" be adhered to in order to  be "compliant"
6. Action Plan. Very important -- List of tasks required to bring spec element to completion.



How does the following schedule sound to the group?
        One week for everyone to read Section 5.6 of the whitepaper and submit comments. (Submit comments by 2/23, in email, or hopefully, we'll have a Wiki set up by then.)
        Two weeks for general discussion of everyone's comments and to identify items to be included in report.  
        One week to do a  final write up of the report.



I can kick-off the discussion with some comments about the "Goals of the spec element":

I think the goal of the Abstract Interfaces section is to provide a platform-independent model of the types of components that UIMA developers can implement (e.g., Flow Controller, Analyzer, CAS Multiplier) and the operations supported by these components.  This lays the foundation for more concrete interface specifications such as WSDL definitions and corresponding SOAP and Java bindings.  

As far as "what is the spec element trying to acheive in terms of interoperability?", it seems that this spec element doesn't really provide out-of-the-box interoperability (that comes later with SOAP bindings for example), but is nonetheless important for defining a common vocabulary and model for UIMA.  If two systems both implemented the Abstract Interfaces but did not share a common binding to a programming model, they would not interoperate out-of-the-box, but building an adapter between them would probably not be difficult.

It is a little unclear what to do about "compliance points", since it is somewhat subjective what it means to comply to an abstract model.  Right now the white paper does not list any Candidate Compliance Points in section 5.6.  We'll need to decide if that is correct.  I'm interested to know what the other group members think, as they read through the section, about what we can say regarding compliance points.


That's all for now.  As you read throught the whitepaper, if something is not clear please send questions to the list, and I can try to clarify what it was we meant when we wrote that.

Regards,
  -Adam


_____________________________
Adam Lally
Advisory Software Engineer
UIMA Framework Lead Developer
IBM T.J. Watson Research Center
Hawthorne, NY, 10532
Tel: 914-784-7706,  T/L: 863-7706
alally@us.ibm.com



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