Re: Mapping arbitrary number of attributes to DB

Marshall wrote:

> On Oct 26, 7:31 am, Bob Badour <bbad..._at_pei.sympatico.ca> wrote:
>

>>>I believe the problem is the cost of up front modeling.
>>
>>As I stated clearly, you were begging the question. You invented an up
>>front modelling cost that doesn't necessarily even exist and then
>>alerted us to this 'problem'.
>>[...]
>>You assumed a non-existent large up-front 'modelling cost' to prove a
>>large up-front 'modelling cost'.

He posited a situation where the data need stored without any further knowledge of the requirements (with the requirements presumably future discovery.) Creating a schema that is a 1:1 mapping of a file may not be free, but it is no more costly than any other option.

> I want to make sure I understand what you are saying. It
> *sounds* as if you are saying that, after gathering requirements,
> turning those requirements into a relational data model (example:
> SQL CREATE TABLE statements) costs less than epsilon.
>
> My experience is that such process usually involves lots
> of wailing and gnashing of teeth. But perhaps it would
> make sense to attribute that effort to extended
> requirements gathering?

I suggest you go back and read what Roy wrote and how Sempo responded. Sempo posited a large up-front 'modelling cost' at the same time he stated the requirements were largely unknown except for simple storage for future analysis. How large a cost does one incur to say: "We don't know what we are going to do with this so we better preserve its original structure as closely as possible" ?

If my requirement is "I need a place to dump this for future use", dumping it someplace that offers extremely effective management and manipulation capabilities strikes me as a fair choice. That "even the crappiest so-called SQL DBMSs on the market already solve the problem [of dumping data without further analysis] extremely well for practical purposes" strikes me as almost self-evident. What Sempo wrote in no way contradicts Roy's assertion.

> Not sure I follow. The idea that there is a cost
> to figuring out the right schema seems reasonable
> to me. It is my experience that the cost is vastly
> outweighed by the benefits almost all the time,
> but I'd still say there was an associated cost.

What requirements? Sempo posited a situation with no known requirements: "we only know that we have some data, some proper subset of it will be useful later on, and so we want to store it, to be sorted out later." Received on Thu Oct 26 2006 - 19:50:37 CEST