Re: Datatype disjointness implemented

On 16 Mar 2009, at 08:58, Alan Ruttenberg wrote:

> On Mon, Mar 16, 2009 at 4:50 AM, Bijan Parsia
> <bparsia@cs.manchester.ac.uk> wrote:
>> On 16 Mar 2009, at 06:52, Alan Ruttenberg wrote:
[snip]
> Your arguments are not persuasive.

Actually, they are conclusive. That you refuse to be persuaded is a  
feature of you and not the arguments.

I notice that instead of *refuting* the arguments, you merely reject  
them. That is a sign of the quality of your refusal.

By the by, picking the weakest argument against your position (or  
nominally against your position) and ignoring the rest is arguing  
against straw men, and a fallacy. You have been doing this  
*continually* throughout this debate (for example, ignoring the  
reasons other than RIF harmonization; trying to make a *PROCEDURAL*  
appeal against our *VOTING* on the issue!! (that was, even with chair  
hat off, chair malpractice)).

Give that you and Rees are not experts (nor, even, educated laypeople,  
afaict[1]) in 1) XML Schema, 2) logic, or  3) mathematics and the  
philosophy of mathematics there's no prima facie reason to find your  
*brute* rejection of my *standard* (at every level!) arguments  
remotely in court. This is just a form of denailism[2].

Just to recap:
	1) XML Schema primitive type disjointness (including numerics) is the  
standard theory and enjoys wide consensus as a coherent (indeed as the  
intended) reading of the spec. For example:
	http://www.w3.org/TR/swbp-xsch-datatypes/
explicitly faces the cases of floats not entailing lexically similar  
doubles.

	2) Primitive disjointness has been widely implement (e.g., Jena and  
Pellet) and has not received any user complaints. At all.

	3) Standard, consensus understanding of mathematics and its ontology  
supports a coherence reading of the XML schema spec.

To all this, you've given no substantive rebuttal. None. Lots of brute  
denials, cherry picking, goal resetting, etc.

The only sensible argument that you can make, roughly, that you prefer  
non-disjointness and believe that most users would too to the point  
that the implementation and specification coordination pain is worth it.

Leaving aside the pain judgement, there's little reason to believe  
that your preferences are remotely representative esp. with your  
intensity. We have several users in the group who work heavily with  
datatypes who are fine with disjointness. We have people who work  
heavily with a broad range of users who believe their users are fine  
with disjointness. We have the years of Pellet and Jena users (who  
have used floats and doubles and integers, btw) without any complaint.

Afaict, only you and Rees (and, historically, Rob Shearer and his  
reports on Cerebra users) have a problem with it. As neither of you  
have any expertise in user analysis and investigation, we really have  
to treat your preferences as fairly ungrounded preferences. You've not  
presented a cognitive walkthrough (even) where disjointness would be a  
serious problem.

Thus, I ask the chairs to stop such discussion. I also ask that you,  
Alan Ruttenberg, make clear in any discussion and negotiation with the  
RIF working group that the decision of the working group is in for  
disjointness and that any discussion on your part is purely in  
representation of your organization. I believe that any such  
discussion should *not* occur in back channels, so I ask Chris to  
conduct such discussions in public.

Cheers,
Bijan.

[1] By which I mean you make elementary mistakes, do not spot them,  
have a very hard time understanding them, are not aware, even  
remotely, of relevant concepts or literature, etc. IOW, you don't seem  
to have the background of someone having taken an (advanced)  
undergraduate class on these topics.

[2] http://scienceblogs.com/denialism/about.php

Received on Monday, 16 March 2009 10:26:41 UTC