Re: Shared game-data

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Mon, 01 May 2006 13:45:00 GMT
Message-ID: <who5g.1345$A26.37775@ursa-nb00s0.nbnet.nb.ca>

Alvin Ryder wrote:

> Alfredo Novoa wrote:
>

>>>Alfredo, exactly which part is wrong?
>>
>>This:

>
> Alfredo, thanks for your reply.
>

>>>>>OTOH game objects contain some atomic types but they also contain
>>>>>pointers, other objects, trees, collections ... which is all very
>>>>>non-atomic and very non-normalized.
>>
>>To work with pointers is always painful and they might be eliminated
>>with the RM. Trees are very well handled by the RM, and relation is the
>>most convenient collection to manage data. Normalization consists in
>>the elimination of some database design errors, and it is always
>>possible.

>
> You could eliminate pointers but at what cost. In some applications
> performance and not theory conformance is what matters most.

One can eliminate them from logical discourse with no cost at all. At the physical level and internal to the dbms, one will always have pointers of some sort.

>>>Are you saying objects used in games, or any OOP, don't need to contain
>>>references to other objects, trees or collections?
>>
>>They don't need to contain pointers. They might refer to other objects
>>using values.

>
> That extra level of indirection is ok in some applications but not
> others.

Nothing says there is any additional physical indirection. I suggest you pay more attention to the requirement for physical independence as a means to separate the concern for correctness from the concern for performance. Received on Mon May 01 2006 - 08:45:00 CDT