RE: [nmrg] draft-irtf-nmrg-im-dm-00.txt status

["Andrea Westerinen" <[e-mail]>]

Discussion points up here:
- Your revised wording of #3 is ok with me.
- About reference [3]... If you don't add the DMTF document URL that is ok
with me, but you do have a URL in reference [11].  IMHO, it would help the
readers of this document to (at least) get the high level URLs for DMTF
(www.dmtf.org), OMG, ITU, etc since they might not know where to look for
the references.

Andrea

-----Original Message-----
From: Juergen Schoenwaelder [mailto:[e-mail]]
Sent: Tuesday, September 03, 2002 1:18 PM
To: [e-mail]
Cc: [e-mail]
Subject: Re: [nmrg] draft-irtf-nmrg-im-dm-00.txt status



>>>>> Andrea Westerinen writes:

Andrea> A couple of comments on the draft:

Andrea> 1. The sentence, "In particularly the terms "Information
Andrea> Model" (IM) and "Data Model" (DM) (Section 1) turned out to be
Andrea> controversial."  should read "In PARTICULAR ...."

fixed

Andrea> 2. The sentence, "... its outcome, which is a better
Andrea> understanding of the terms "Information Model" and "Data
Andrea> Model", as presented in this document." (Section 1) should
Andrea> read "... its outcome, ..., IS presented in this document."

fixed

Andrea> 3. The sentence, "The purpose of an IM is to model managed
Andrea> objects at a high conceptual level, which is easy to
Andrea> understand for the human designer or human manager.  ..."
Andrea> (Section 2) needs some clarification.  "High conceptual level"
Andrea> implies broad concepts, but IMs can define detailed data. I
Andrea> would say "The purpose of an IM is to model managed objects as
Andrea> abstractions, independent of specific implementations or
Andrea> protocols used to transport the data.  The degree of
Andrea> specificity (or detail) of the IM is dependent on the modeling
Andrea> needs of its designers."

I would like to keep "conceptual level" in that sentence because we
use this word later on in several places. What about this:

     The purpose of an IM is to model managed objects at a conceptual
     level as abstractions, independent of specific implementations or
     protocols used to transport the data.  The degree of specificity
     (or detail) of the IM is dependent on the modeling needs of its
     designers.

Andrea> 4. I do not understand the importance of "managers" in the
Andrea> following sentence, "An IM is primarily useful for designers
Andrea> and managers" (Section 3). I would expand the sentence and use
Andrea> the "guide" word from John Strassner's mail - "An IM is
Andrea> primarily useful for designers to describe the managed
Andrea> environment, for managers to understand the modeled objects,
Andrea> and for implementors as a guide to the functionality that must
Andrea> be described and coded in their DMs."

fine with me

Andrea> 5. The following sentence is not grammatically correct and
Andrea> should be reworded.  "SMIng also supports some object-oriented
Andrea> principles and provides an extension mechanism which allows to
Andrea> add more features such as support for methods when the
Andrea> protocols support them without breaking SMIng
Andrea> implementations." (Section 4) My suggestion would be: "SMIng
Andrea> also supports some object-oriented principles and provides
Andrea> extension mechanisms allowing the addition of new features
Andrea> (such as support for methods).  New features can then be used
Andrea> when supported by underlying protocols, without breaking SMIng
Andrea> implementations."

fixed

Andrea> 6. The sentence, "Instead, the SMIng working group currently
Andrea> focusses resources on developing a third version of the SMI
Andrea> (SMIv3) which ..."  (Section 4) should be simplified as
Andrea> "Instead the SMIng working group is developing a third version
Andrea> of the SMI (SMIv3) which ..."

fixed

Andrea> 7. Could you have reference [3] include the URL
Andrea> http://www.dmtf.org/standards/published_documents.php?  This
Andrea> is where DSP 0004 is permanently located.

Not sure (a) how I do that with xml2rfc and (b) if I like to go down
the path of adding URLs to references at all (I am old fashioned).

/js

--
Juergen Schoenwaelder    <http://www.informatik.uni-osnabrueck.de/schoenw/>