Re: C# 3.0 with pseudorelational extensions

From: BobTheDataBaseBoy <"xxx>
Date: Sat, 17 Sep 2005 17:01:00 -0400
Message-ID: <h6mdnWY6-sZtHrHeRVn-ow@rcn.net>

Marshall Spight wrote:
> BobTheDataBaseBoy wrote:
>

>>>I have no idea whether DLinq is well-executed or not. But I'm
>>>convinced that unifying the capabilities of the dbms and the
>>>general purpose programming language will lead to some great
>>>things.
>>>
>>
>>prolog appeared within months of RM, and was based on just that premise.
>>  it hasn't taken the world by storm in the decades since.

>
>
> I dunno. Prolog is still alive and kicking 30 years after its
> debut, and that's more than most programming languages can say.
> But Prolog's not really what I have in mind. It's decidedly
> single-paradigm, for one thing.
>
>
>

>>>I also agree that O/R mapping, as it is commonly practiced, is
>>>mostly about throwing out relational functionality and putting
>>>an OO face on what's left behind. And I'm convinced that this
>>>is *not* a useful approach.
>>
>>it's an idea made be people who don't know how computers actually work.

>
>
> Well, that's a bit stronger than I'm comfortable with. It's generally
> the work of people who understand OO just fine, but don't understand
> relational all that well.

the notion of the mismatch derives from the notion that an object must:

i) be realized in the computer's memory whole and therefore
ii) be persisted whole

i) isn't true, thus ii) isn't true.

Bob
>
>
>
> Marshall
>
Received on Sat Sep 17 2005 - 16:01:00 CDT