Examining the SQL Server Release Cycle

the theory is correct but problem is 2005 was full of new features that needs more than 3 years from administrators and developers to digest, so the simple reaction is to ignore the 2008 release totally and concentrate on the next version, it was better from MS to wait till end of 2009.

We are in the process of making our customers to migrate to SQL2005, and we will stop support of pre-SQL2005 versions in june. What make us sorry about SQL2008, is that it will be hard to justify to our customers to go to for it soon because the just spend some money to go for SQL2005.

It would be great if Microsoft would reduce cost upgrade depending on early you get the previous version. For example buying SQL2008 right when it goes to manufacturing could be significantly less expensive than buying it one year later. Microsoft would get income from the new version earlier, and customer would find some incentive to go to the new version.

The alternative, i.e. support of both version SQL2005 and SQL2008 into the same application is unfortunately not a good incentive to use the great new features of the newest version.

Yes, five years was too long, but not because it was five years.
It was too long because Shiloh was Microsoft's first 'ground-up' version of the product after the split with Sybase and the lack of features became painfully evident after time. Not to mention EM wasn't exactly rock-solid.

Yukon was the version designed to get the attention of the enterprise crowd and it has succeeded rather well by that measure. But we need a pause to get caught up with 64-bit hardware, applications, drivers, etc.

As a DBA that is forced to support whatever version our vendors decree, we are frequentlly held captive in 32-bit land, not to mention Shiloh, or even (shudder) Sphinx.

It would be nice if Microsoft could somehow light a fire under these pokey vendors (ending mainstream support for Shiloh in April '08 doesn't seem to be getting the job done) so we could move forward with a modern version - be it Yukon or Katamai.

I would like to see Microsoft stick to a three-year release schedule. No sooner and no longer. In addition to letting us digest the current/previous version, it would dovetail nicely with SA.

Finally, someone in the media with both a memory and intellectual honesty. I distinctly recall every single SQL Server "journalist" with an opinion, and much of the user community, lambasting MS for taking 5 years to bring a new release to market with SQL Server 2005. They also complained that the release was too significant a change from SQL Server 2000. In other words, the upgrade was not incremental enough. Now the hue and cry is that SQL Server 2008 is coming too soon. I don't believe anyone is being forced to upgrade to this release, and I suspect many organizations will simply choose to skip a release - either SQL Server 2005 or SQL Server 2008, depending on their needs. No harm, no foul. Let's find a legitimate issue to fuss about.

Just disagree with the authors opion. Many people just recently upgraded to 2005 paid thousands of dollars for that upgrade. Why can't Microsoft provide better upgrade incentive for customers who stay current?