W3C home > Mailing lists > Public > public-owl-wg@w3.org > July 2008

Re: ISSUE-114 [RE: Teleconference.2008.07.09/Agenda]

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Wed, 9 Jul 2008 09:17:22 +0100
Message-Id: <F68F6D6E-139A-44A8-A8B1-0F91D2BB9058@cs.man.ac.uk>
Cc: "OWL 1.1" <public-owl-wg@w3.org>, "Alan Ruttenberg" <alanruttenberg@gmail.com>
To: "Michael Schneider" <schneid@fzi.de>

On 9 Jul 2008, at 08:41, Michael Schneider wrote:
[snip]
> And I am its raiser.
>
> Let me say that I originally thought it would be peculiar to have
>
>   (class|datatype) / (object|data|annotation)property punning
>
> in OWL DL. But now, I have heard several people arguing pro this  
> kind of
> punning. So you won't see me opposing this any further.

So withdraw the issue.

> Btw, I don't think that it was "improper" to raise this issue.

As extensively argued at the time, it wasn't a substantive issue.  
I.e., you raised no specific technical or user point. It was a  
fishing expedition. It's perfectly reasonable to fish on list (we  
need review) but to make it an *issue* just distorts the meaning of  
"issue".

> For example,
> it eventually brought to everyone's attention that class/datatype  
> punning is
> now disallowed, too.

That there were possibly useful side effects doesn't make it proper.  
You also asked today for there to be a *very* heavyweight issue  
resolution process. That only works if issue *raising* is similarly  
weighty. If it's easy to raise issues and hard to get rid of them,  
then we make progress very difficult.

(Also, there are several review points ahead of us where the  
documents are subject to a lot of scrutiny. Why is *now* a good time  
to bring this to everyone's attention?)

> However, if I was about to raise this issue today, I
> would rather start such a discussion off-issue-list, in order to  
> see what
> people think (the preferred method, before raising an issue, anyway).

Yes. Hence, it being improper :) Please note that I don't have  
*animous* against you for that. We just, at the time, had different  
views of how issues per se should work.

Here's an analogy. My understanding is the if you do a full body scan  
(with an MRI) you will *always* find "suspicious" stuff. (Nodes,  
lumpy things, etc.) Everyone has that. The problem is that once  
something suspicious shows up, the doctors are *obliged* (by ethics  
and by sensible fears of malpractice) to take them all seriously.  
This can involve invasive and inherently dangerous methods such as  
biopsies. Thus, random scans can put the person at *greater* risk  
(and cost a lot of resources that could have been used eleswhere).

Raising issues like this seem to be like doing random scans, esp. if  
we add a heavy resolution process.

Cheers,
Bijan.
Received on Wednesday, 9 July 2008 08:25:05 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:05 UTC