Re: So what's null then if it's not nothing?

"DonR" <donr_work_at_yahoo.com> wrote in message news:1132350557.673092.92800_at_g49g2000cwa.googlegroups.com...
> Hello Hugo,
>
> Since I'm a "Pickie", I'll throw my 2 cents in here. Caution, a long
> post.
>

[big snip]
>
> I would be glad to continue this discussion with anyone who so desires.
>
> I told you this was a long post ;-)
>

Don,

Your post was far and away the most valuable thing I've read in this forum as an overview of Pick for those of us not familiar with it. It was worth the length, and a fine read.

I do want to continue the conversation. Hopefully, given the tone you've started with, it will be a discussion with more "sense and sensibility" and less "pride and prejudice" than the other Pick vs SQL discussions that have preceded it.

I'm going to be quoting it in bits and pieces over time, and trying to add my insight (such as it is) into what the comparison might mean, going forward.

For now, I want to limit my comments to an overview of the situation.

Let's imagine some yardstick of total cost (net present value), and let's further imagine that the average cost works out to be:

3 for a Pick application and 1 for an Oracle application, after an initial cost of 10 for the development of the database to be shared by all the apps. By "an application" I mean a whole suite of application programs that cover a major function, like Order Processing of Purchasing. These numbers are purely fictitious.

Now let's say someone observes the costs, in an environment where most databases ashared between 2 or 3 applications. That observer is going to note that the cost of the Pick applications is lower, every time. Unless the observations are extremely careful, the obeserver will have no inlking that, once the number of applications that share a database grows to 10, the cost edge is now on the Oracle side.

And that imaginary scenario illustrates much of my thinking about databases. My work was mostly done in an era where the construction of a new database was more of an enterprise wide issue than the mere support of a single application or two.
As time has gone by, that scenario has tended to change. Many, many of today's databases are simply not integrating large numbers of applications anymore. So the perception I have does not resonate with today's programmers.

I still think that "large shared data banks" will enjoy a comeback some day.

More details to follow. Received on Tue Nov 22 2005 - 01:27:47 CET