Re: no names allowed, we serve types only

paul c wrote:

> Bob Badour wrote:
>

>> Jan Hidders wrote:
>> ...

>>>>>> C.Date presents this argument very well in section 20.9 of an
>>>>>> Introduction to Database Systems where he claims that a coloured
>>>>>> circle is not a subtype of circle (or vice versa).
>>>>
>>>>> The tuple that represents the circle is not the same thing as the
>>>>> circle itself. I find Date's argument rather unconvincing, to put it
>>>>> very mildly. He is by no means an authority in this area, and those
>>>>> that are mostly disagree with this position.
>>>>
>>>> Since when do you find argumentum ad verecundiam convincing? Hmmm?
>>>> [peers over rim of eyeglasses]
>>>
>>> I don't, nor do I think it is without any meaning whatsoever.
>>
>> What happens when one disagrees on what makes authority. For example, 
>> I consider Date an authority in this area--your ad hominem 
>> notwithstanding--because he put considerable tuition into the subject 
>> over a period of a decade or more. He did so with full knowledge of 
>> what others before him had to say including Cardelli. Further, he did 
>> so with an eye to obviating inconsistencies and flaws in those earlier 
>> works.
>>
>> More to the point, I find his arguments convincing.
>>
>> If you cannot offer a convincing reply to them, can you at least 
>> direct me to someone who has replied convincingly?

He hasn't. His discussion of circle and ellipse relate to data types not compositions.

> Personally, I have a great deal of trust in what he writes. In large
> part my reason is that he is pretty thorough about examining the
> statements of other authors (not to say 'author' means authority, nor
> 'academic' for that matter), whereas he's pointed out that a number of
> them have declined to correspond with him. His body of writing is
> pretty large and I would say it deserves respect and public examination
> by all serious academics. I don't always agree or understand what he
> says but it is usually thought-provoking. Actually his dogmatic
> positions seem rather few, for example I don't think he advocates that
> every relational technique must be founded on mathematical principles
> that weren't necessarily originated for the purpose of information
> theory or a practical dbms.
>
> Many of the people who have publicly argued with him struck me as
> hand-wavers.

Indeed.

> Can't comment much on type theory (in fact I don't even understand how
> this thread crossed into that territory). This is a young field. It
> might be useful to remember that many medical 'authorities' of the
> 18th-century are not remembered that way today.

It crossed into the territory when David BL (you inexplicably cut out the attribution above) introduced the discussion of circles and colored circles. Received on Thu Feb 25 2010 - 01:47:50 CET