Re: Announcing New Blog

"dawn" <dawnwolthuis@xxxxxxxxx> wrote in message
news:1137039711.753911.222510@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
> mountain man wrote:
>> "mAsterdam" <mAsterdam@xxxxxxxxxxx> wrote in message
>> news:43c59fa2$0$11068$e4fe514c@xxxxxxxxxxxxxxxxx
>> > mountain man wrote:
>> >> dawn wrote:
>> >>>See http://www.tincat-group.com/mewsings
>> >>>Leave a comment :
>> >>
>> >> My impression of Codd was that he preferred to avoid the
>> >> complexification
>> >> via normalisation, as outlined in an article by Gittens earlier
>> >> referenced
>> >> ("A Codd inspired ..."
>> >>
>> >> OTOH my impression of Date is that he prefers to promote
>> >> complexification via normalisation.
>> >
>> > I read some books and attended some lectures by Chris Date.
>> > He doesn't shy away from difficult subjects.
>> > Could that be the trigger your impression?
>>
>> No, there are more difficult subjects that he avoids,
>
> To be fair, we each have subjects that are not our bailiwick, even
> within broader areas of expertise.


Date's publications are regarded as university level educational
material for database systems. As such, I see that he some form
of obligation to stay in touch with reality.



>> such
>> as schema evolution theory,
>
> While I don't fault him for ignoring or underestimating some subjects,
> I would say that a person's opinions about how to model data can be
> flawed because they do not take some important things, such as schema
> evolution, into account.


Database systems is an evolving environment.
Date's pedagogy on the RM is not.
It has been static for nigh on 30 years.



>> and omits others entirely,
>> such as any formalised theoretical interface to the software
>> side of the (data/software) coin. (eg: stored procs).
>
> I agree that marginalizing either of these topics has implications and
> Date's writings sometimes favor "data in a vacuum" thinking.
>
>> > Could you provide some examples?
>>
>> Normalisation of tables in order to pedantically avoid
>> the use of nulls as though they were some form of an
>> alien life form.
>
> They are a shame, but if using SQL it is just a game and not highly
> practical to eliminate nulls. I understand going out of your way to be
> able to use a 2VL, but reworking the database schema repeatedly just to
> avoid the introduction of nulls seems silly at best and could be both
> risky and costly.


And counter-productive.



>> Date approaches complexity ad nauseum
>> but offers no simple solutions, and cannot, because of
>> his data centric view of a data _processing_ environment.
>
> I agree but still like to read what he writes. Cheers! --dawn


Why? After a certain stage, it is mainly repetition.






--
Pete Brown
www.mountainman.com.au



.

Relevant Pages

  • Managing Nulls in the DB
    ... Do you think the best way to avoid the problems of nulls in the database is ... just to provide default values via the db schema? ...
    (microsoft.public.dotnet.languages.csharp)
  • DataTable with Column Default Values not working as execpted
    ... columns that have default values and do not allow nulls. ... So I added a manual schema fill for default values and everything works fine ... schema information as ), which I then try to set as the default ...
    (microsoft.public.dotnet.framework.adonet)
  • Re: more closed-world chatter
    ... I just don't see how you can avoid it. ... relation schema, and he felt that order should be dispensed with. ... one domain can interact with those from another provide the means whereby ... comparisons beyond equality for values from a single domain and comparisons ...
    (comp.databases.theory)
  • Re: When is a NULL Not a NULL ?
    ... I would always *try* to avoid them, Nulls that is, with careful database ... State what version of SQL Server you are using and specify the content ...
    (microsoft.public.access.adp.sqlserver)
  • Re: On specialization constraints time of application
    ... that drawbacks of using nulls far outweights any advantages. ... For example, a schema for ... SQL vs OO ... is not in the layer of RM but for instance the layer of computing ...
    (comp.databases.theory)