[UCR] use case response

I find the way this strawpoll was set up is rather ridiculous and 
makes for a lot of extra work and email, but what the heck, I'll play 
-- here are my comments on the sections and this is the email I will 
point to in the web-based form.    I apologize to Sandro for not 
breaking it into 8 separate section, but since there is a single 
theme running through my comments, I choose to show how that theme 
informs all of these together


Let me make clear my motivation about the following - I think too 
many people in this group are still working on use cases for 
rule-based reasoning, not for Web rules or rule exchange.  Those of 
you who were in WOWG will remember that I went ballistic on that 
group over similar issues in the early days, and insisted we throw 
out all the use cases that said "ontologies are good" and insisted on 
ones that emphasized their use on the Web for real Web problems.  I 
think several of these use cases are far from something I can 
understand seeing RIF in


1 - Info Int  (No)
  I don't understand what this section has to do with rules exchange 
or even really rules on the web.  The examples seem contrived.   It 
seems to me that applying rules to facts in a standard KB meets this 
use case, but standard rule application doesn't seem to me to talk to 
the work of this group.

2 - Decision Support (No)
  Let me get this straight - a use of RIF is e-learning in a single 
system that is running on a laptop?  There is a section in the middle 
wherre somethign called "MEDIC" is discussed, and it might actually 
be a rule deployment problem (as discussed in 2.3)  but the use of it 
here isn't really made clear.
  It might be possible that if this was reworded it would have 
soemthing to do with decision support instead of training bad doctors
  It may be the case that the intent is that the info from MEDIC was 
transmited to the e-learning program and that these interns and 
residents were thus able to use the specific case they were involved 
in, rather than generic knowledge, to prepare answers in advance 
(let's not make them lazy).  If that was the case, this might be a 
RIF scenario, although I still am not sure where the decision support 
part comes in.

3 - Cross Platform Rule Devel and Deploy (Yes)
  This is what I thought RIF was all about - I like this one

4 - Policy based transaction authorization (Yes with editing)
  I think this is a good example, but as written it concentrates on 
the use of rules for authorization, not on the standardization of 
formats and the exchange.


5 - Interchange of human-oriented business rules (No - but see section 6)
  It took me a long time to figure out what this one was about, and 
then what I got to the publishing one it took me a long time to 
figure that out as well.  I think these two can be combined, see below

6 - Publication (yes, with editing)
I think whether the target is humans or machines (which seems like it 
may be the difference between 5 and 6) the idea in both these use 
cases is that sometimes it is useful to have a declarative format for 
rules that could easily be mapped to various formats for human 
readability (the thrust of 5) and for eventual machine processing 
(which I think is the intent of 6, although I may be reading that 
into it)
  I think these two use cases could be combined into a very useful one 
that says sometimes it is important for humans to be able to read the 
rules in a browser or etc, rules are generally very hard to read, and 
thus having a standard format would allow for the development of 
tools which had good ways to display rules based on their syntax, not 
domain, and thus developers could create third-party plug-ins for 
displaying rules to users (and, for that matter, we could mention for 
creating machine-readable rules).
  These would be the essential HCI drivers for RIF, and I believe are 
critical to its use (and to why a standard, web-exchangeable syntax 
for rules is needed.

7 - third party interachange  (yes with editing)
  I think this is really important, but it is very wordy and one can 
get lost before getting to the end, which is where the use of RIF is 
explained.  I would suggest rewriting with the scenario first and the 
broad areas afterwards.

8 - Rich KR (potential formal objection)
a separate message will be sent on this one


-- 
Professor James Hendler			  Director
Joint Institute for Knowledge Discovery	  	  301-405-2696
UMIACS, Univ of Maryland			  301-314-9734 (Fax)
College Park, MD 20742	 		  http://www.cs.umd.edu/~hendler
Web Log: http://www.mindswap.org/blog/author/hendler

Received on Monday, 20 February 2006 17:48:54 UTC