Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> comp.databases.theory -> Re: Embedded SQL, again

Re: Embedded SQL, again

From: paul c <toledobythesea_at_oohay.ac>
Date: Wed, 15 Nov 2006 23:13:17 GMT
Message-ID: <haN6h.318363$1T2.59062@pd7urf2no>


David Cressey wrote:
> ...
> Third, I want better transaction management than SQL provides, although I'm
> not quite sure what I mean by "better".
> ...

Me too. My idea of 'better' would be a system that allows designers to invent different transaction models for different part of app's but allow them to prevent the average (ie. incompetent) developer from issuing their own concurrency directives.

At heart, any transaction model or facility is just a way of binding assertions about various states that already exist or will exist during the transaction. I think the transaction theorists have fallen down on the job by failing to abstract this.

Also, as far as the topic is concerned, if other people are interested in it and want to tread the same uninspired and tired ground, so be it, but I would rather see a language that doesn't have the conventional 'general purpose' verbs such as ones that iterate - they are just mirrors of various typical machine instructions. Would rather have one that is predicate-based from the ground up. For example, if one wants to do something a million times, why not just JOIN a built-in relation named 1000000 to whatever relations one wants to do the something to?

p Received on Wed Nov 15 2006 - 17:13:17 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US