Re: Serious article on comparison between MS SQL Server 2005 Yukon and Oracle 10g

On Thu, 18 Nov 2004 03:38:57 -0800, Niall Litchfield wrote:

> Those
> guys who wrote the article they did some serious piece of work.

Yes, there were several revelations in the article: 1) On page 12: "ROLLBACK SEGMENTS don't support PCTINCREASE and hence don't "autogrow" so you must get the size right when they are created"

I could have sworn that I've seen parameters like initial, next, minextents and maxextents which do have some relation to the size of the object. My memory must be playing tricks on me.

2) On the same page, just below:
"MSSQL 2005 Oracle
READ COMMITTED No Equivalent
(locking)
REPEATABLE READ No Equivalent
SERIALIZABLE No Equivalent"

I have a dream. I have a dream that the statements below will implement precisely the features that are called "unsupported". "SET TRANSACTION LEVEL [SERIALIZABLE|READ COMMITTED]" and SET TRANSACTION READ ONLY (repeatable reads).

On page 20:
"In our opinion, MSSQL databases structure is simpler and designed better than Oracle tablespace. Each MSSQL database contains its own dictionary. On other hand, Oracle has more recovery options for corrupted database, redo log or datafile than MSSQL."

I believe that the ancient debate of centralized dictionary vs. partitioned or distributed dictionary was settled with Ingres and RDB. Centralized dictionary proved better then "complete entities".

Page 24:
"
.NET in MSSQL vs. Java in Oracle
It is looks like MSSQL 2005 has a serious advantage over Oracle. Java inside Oracle database can be called only indirectly meaning that it should be wrapped by PL/SQL or by using either CORBA or EJB interfaces while .NET in MSSQL can be called absolutely transparently as Transact-SQL. In addition, uploading and maintaining of Java objects in Oracle is the same as maintaining a complete java application server. On the other hand, MSSQL provides a simple interface to develop, debug and use .NET triggers, stored procedures and packages in the MSSQL 2005."

First, this is an outright lie. One can call Java procedures directly, using the CALL command. Second, it remains to be seen whether this .NOT stuff will become the greatest source of viruses ever invented. Visual basic in documents and spreadsheets looked fabulous until some teenager found the way of cajoling people into formatting their drives by opening an email. God bless .NOT and Winduhs!

Finally, the debate that got my eyebrows raised was the discussion of the identity field vs. sequences, claiming sequences to be a "major overhead".

Well, sequences are certain (but by no means a major one) overhead because they allow things like clustering (see "ORDERED" and "NOCACHE") as well as good caching for performance. The article doesn't say anything about he performance of the identity fields in the clustered environment.

By all standards this is a substandard article in which some important facts are plain wrong and conclusions are shallow at best. Verdict: waste of time.

-- 
Artificial Intelligence is no match for natural stupidity.
 .