W3C home > Mailing lists > Public > public-swbp-wg@w3.org > June 2004

Re: [OEP] Draft of a note on n-ary relations

From: Natasha Noy <noy@SMI.Stanford.EDU>
Date: Thu, 10 Jun 2004 14:44:26 -0700
Message-Id: <58157000-BB27-11D8-A5C9-000A958B5C28@smi.stanford.edu>
Cc: public-swbp-wg@w3.org
To: Fabien Gandon <Fabien.Gandon@sophia.inria.fr>

Fabien,

This is a very good point and something that we should probably discuss 
as a group. I see two mutually-exclusive arguments for whether or not 
to include all approaches, including the ones we would generally rule 
out, in the OEP note.

Argument 1 (pro): Including patterns that we do not recommend but that 
a reasonable person may come up with is good because it preempts 
questions from readers such as "Why didn't you do in way X?" It often 
helps understand the patterns that we do recommend.  It also shows that 
we did our "due diligence" in considering other approaches.

Argument 2 (cons): The WG is the "best practices" WG, so the patterns 
that we produce should be the ones we think are good alternatives in 
some common set of circumstances. Therefore, including considerations 
for other patterns and then saying that we don't generally recommend 
them would dilute the message, in a sense; besides, someone who just 
wants a simple answer to his question "how do I do X?" may not want to 
read through all the reasons why you should not do it in a particular 
way.

I personally am leaning towards argument 2, argument 1 is often more of 
interest to academics than practitioners (Incidentally, 2 is also 
usually  less work :) But I can see the argument going either way.
	
Natasha

On Jun 10, 2004, at 9:33 AM, Fabien Gandon wrote:

>
> Alan Rector a écrit :
>> I think this approach has a serious problem in that
>> it assumes that there an only be one temperature (...)
>
> Alan,
>
> I completely agree with your remark and my message was confusing.
>
> In fact, the purpose of this message [2] was not to defend a 
> particular approach (most of the time I use the option of reifying the 
> instances of the n-ary relation).
> My point really was that, IMO, a maximum of different possible options 
> should be listed in the design pattern with their pros and cons like I 
> tried to propose in a previous message [1].
>
> What I was trying to avoid is to have a reader thinking we did not see 
> an alternative approach while it was in fact implicitly discarded 
> because of reasons that seemed obvious to us.
> I think all the options we can think of should be listed and 
> discussed, even briefly, to provide the design rationale that really 
> makes the design pattern reusable.
>
> But then again, may be I completely missed the point of this note,
>
> Fabien
>
> [1] 
> http://lists.w3.org/Archives/Public/public-swbp-wg/2004May/0128.html
> [2] 
> http://lists.w3.org/Archives/Public/public-swbp-wg/2004Jun/0054.html
>
> -- 
> "men build too many walls and not enough bridge."
>                                 -- Isaac Newton.
>  ____________
> |__ _ |_  http://www-sop.inria.fr/acacia/personnel/Fabien.Gandon/
> |  (_||_) INRIA Sophia Antipolis - ph# (33)(0)4 92 38 77 88
>
Received on Thursday, 10 June 2004 17:44:31 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:09:39 UTC