This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them.Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.
It's always a pleasure to read anything Joe Celko writes. His column (Oct .30, 2004) on Microsoft SQL Server stored procedures hit the nail on the head.
Oracle has had the ability for developers to write stored procedures in Java since v8, and Microsoft is now playing catch-up by letting you do the same with .Net and SQL Server '05. (The integrated debugging in Visual Studio, which the Oracle environment lacks, is actually quite nice.) The problem is that Microsoft doesn't provide enough guidance in its documentation as to when to write stored procedures/functions in a .Net language and when not to, and let good old SQL do the job. Some of the "hello world" examples provided on its Web site are laughable, and in reality constitute the kind of code you would never want to write in a production scenario.
This situation is being rectified slowly. For example, one of the Microsoft New England evangelists (who doesn't pull his punches and gently criticizes his employer in public lectures when appropriate) warns that, if you create code similar to some of those examples in the hope that your system will run faster, you may be in for a shock. It will often run considerably slower than the plain SQL equivalent would.
There are a few good uses for the new feature. For example, if you wanted a set of routines that operated on BLOBs that happened to be fingerprints — in effect, extending the core system to handle special data types the way Illustra/Informix did through "data blades" — you'd go crazy trying to do this in any language that didn't support 2D arrays, and this is when you use the '05 .Net stored procedures. My experience was the Illustra data blades were a nightmare to write, and you really need an integrated development/debugging environment that hooks right into your database.
You're absolutely right, however, that the current Microsoft printed blurb has the effect of placing loaded guns in the hands of babies.
Prakash Nadkarni, MD Associate Professor Yale Center for Medical Informatics New Haven, CT
On the Truth
I enjoyed Joshua Greenbaum's "The Truth About the Truth" (Sept. 18, 2004), but to tell you the truth, I felt a little let down at the end that he didn't have something prescriptive to offer.
Neil Raden Hired Brains Research/IE Contributor
Greenbaum responds: The lack of prescriptive solution is partly the fault of the limits of the column, partly due to the fact that, other than the caution I suggest, I'm not sure what the solution is. We do need answers to unanswerable questions, public and business decision-making require it. But it's not really clear how that can happen in a "processed" information world without necessarily erring one way or the other.
We welcome your comments on this topic on our social media channels, or [contact us directly] with questions about the site.