[UCR] Issue 7 discussion summary - completes Action 97

This email completes Action 97 on Deborah Nichols ("summarise the
discussion [from 29 Aug] to issue 7").  Discussion notes below are to
capture what was already said in re issue 7, which will be dealt with
at the next (26 Sept) telecom.  Notes follow the issue text below.

 

Issue 7: UCR Document needs multiple restructuring, revisions, and
clarifications

 

Issue 7 includes a summary of comments and suggestions made by Sven
Groppe; the issue is posted
http://www.w3.org/2005/rules/wg/track/issues/7
<http://www.w3.org/2005/rules/wg/track/issues/7> .

See also Sven's original comments, linked from Issue 7 and posted at:

http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/att-000
2/ReviewRIFUseCasesAndRequirements_23_3_2006.txt
<http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/att-00
02/ReviewRIFUseCasesAndRequirements_23_3_2006.txt>   

(copied from an attachment to his email in the archives: 

http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/0002.ht
ml)
<http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/0002.h
tml)> 

 

Published UCR Working Draft is at http://www.w3.org/TR/rif-ucr/
<http://www.w3.org/TR/rif-ucr/> . 

 

Text of Issue 7:

 

Raised by: Sven Groppe.  Posted to Issues by: Deborah Nichols [DLN].
 
Description:  In his comments (full version linked below), Sven Groppe
identifies the following problems with the Use Case Document, which
need to be addressed item-by-item:
 
1. Document should not begin with use cases given as long stories.
2. Document should begin with a discussion of RIF and enumerate its
required 
features.
3. Use cases should be presented as illustrations for each required
feature. 
4. Use cases need to be more worked out, with examples in which rules
are 
translated from a specific language via the intermediate form to a
target 
language.
5. The Abstract needs to clarify which kinds of use cases pertain to
Phase 1 
vs. Phase 2.
6.  Need to clarify what RIF is about, e.g., intermediate language or 
framework for mixed-language support.
7. (Section 1.1)  Is there only exchange of the facts or also of the
rules 
themselves?
8. (Section 1.2)  Reorganize the document to eliminate confusing
overlap of 
features/requirements between different sections.
9. (Section 1.7)  First emphasize what the requirements are, then give 
examples in RDF/OWL-DL.
10. Add a summary with conclusions to the document.
11. Need to give references to resources for the real-world examples.
 
Issue:  Specific suggested improvements to the Use Case document.  For
each 
numbered item above: 
* Has this already been addressed?
* Do we agree with the comment?
* Who will make the changes to the UC Doc?
 
Sven Groppe's comments on the UC Document are attached here:
http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/att-000
2/ReviewRIFUseCasesAndRequirements_23_3_2006.txt  
(was an attachment to his email in the archives: 
http://lists.w3.org/Archives/Public/public-rif-comments/2006Apr/0002.ht
ml)
 
[This was item 5 in: 
http://www.w3.org/2005/rules/wg/wiki/UCR/Document_issues.]
<http://www.w3.org/2005/rules/wg/wiki/UCR/Document_issues.%5d>   

 

 

Discussion Notes on Issue 7 from 8/29/06 telecom (notes by Deborah N.
[bracketed comments DLN])

 

1. Chris W disagreed with point 1 that UCR should not begin with use
cases.  

Sven's point may have been that the use cases are long and their focus
isn't obvious.  Hassan suggested that summary sentences for each use
case might help.  

[Either UCR editors or original abstract UC authors would need to
provide summary sentences.]

 

2. Sven's comments were made on a UCR version with no requirements
included.  Requirements will be included in the finished version, at
the end.   

[Sven's comment also recommended starting the UCR with a RIF
description and the requirements, rather than with the UCs.  No one in
the discussion supported the suggestion to change the order of
presentation in the UCR document.]

 

3. There will be links from each use case to some requirement(s), which
will answer this item.

 

4. Several sub-issues were discussed in connection with #4.

(a)  Is it clear in every UC that interchange of rules is involved?  

Some in discussion thought that this is not clear for every UC.  

Allen noted that since it seems pretty clear that the RIF is itself a
rule language, then translating into RIF would itself be an
interchange.  

ChrisW:  Specific criticism is needed if we think that UCs don't
adequately cover interchange.

csma:  Initially tried to put each UC into a processing model to show
where and what kind of transformation occurred in particular
[original?] UCs.  But that processing model (with UC #s) was lost when
the UC abstraction was done.

Request for csma to post examples [of the process models?].

 

(b)  If Sven is suggesting that specific languages (and translations)
should be presented in the UCs, then that is something the WG agreed
not to do, at the 1st F2F (ChrisW).  

(c)  Being able to translate seems to presume a RIF already, and the
knowledge of how to translate between it and specific languages.

            

5. There was apparent general agreement that the UCs need not be
divided into Phase I and Phase II, even if the Requirements are phased.

 

6. Wrt the question of clarifying whether RIF is an intermediate
language or a framework for mixed-language support:  

Someone stated that the new abstract in the 2nd working draft clarifies
that RIF is about interchange.  

[The Abstract says the UCR doc is "for a format that allows rules to be
translated between rule languages and thus transferred between rule
systems."

The Introduction of the UCR doc indicates potential advances obtainable
from "enabling exisiting [sic] rule-based technologies to interoperate
according to standards-based methodologies and processes," and states:
"The basic goal of the Rule Interchange Format (RIF) Working Group is
to devise such standards and make sure that they are not only useful in
the current environment, but are easily extensible...."

In the discussion of CSCs, the Introduction states: "One of the
critical factors for a successful RIF is that it be useful for
interchange of rules among the set of rule languages it is intended to
cover."]

Also see Sven's original email for elaboration of the alternatives.

 

7. Wrt whether there is exchange of rules or only facts in Section 1.1:
Allen will check.  [Allen, what was the upshot?  I think the section is
now 2.1.]

 

8. Wrt overlaps of features/requirements in different use cases:  This
is natural.  But it will be clearer after the links mentioned above (in
(3) are inserted).  

ACTION on Allen: (Re-)post the previous work on linking UCs to
Requirements.

 

9. This point should also be answered when the UC-to-Requirements links
are in.  [Actually, links won't provide examples in RDF/OWL if we don't
mention specific languages.]

 

10. A summary with conclusions will be added to the UCR document.  That
will answer this point.

 

11. Wrt needing reference to resources for real-world examples:

ChrisW: Not if it means mentioning specific rule systems.  Yes, if it
means clarifying the test cases.

Hassan:  From the technical discussion, RIF is already converging as a
dialect of some XML-based language.  [See also point 6, above.]  If RIF
is incorporating features of other (languages), then the resources
should be posted [linked] for those.

Harold:  Test cases will contain specific languages.  We will add test
cases which will have specific languages, but the test cases will be in
a separate document from the UCR.