Cogent argument, Tony!

There's one way SQL Server 2008 is distinctly *not* a Service Pack, no matter how accurate that moniker is: switching to 2008 from 2005 will cost you real currency!  Fortunately, the additional cost of migration should be quite small, but in licensing Microsoft will have their due.

That said, it seems to me the only strong reasons to upgrade from 2005 are:
- new servers (adding licenses to your organization)
- need of a 2008 only feature (particular SSIS improvements, GIS data types, etc.)

If you have MSSQL 2000 or prior, though, no more excuses!  This sounds like the perfect opportunity for jumping straight to 2008.

Will

I disagree. Where Tony said: "there have been polite murmurs of interest around ... data and backup compression ... However, in terms of added bells and whistles, there seems not to be a standout reason to upgrade to SQL 2008"

I would have set data compression squarely into the category of "standout reason to upgrade to SQL 2008."

I would argue that data compression provides at least as much value as the 2005 improvements you mention: DDL triggers and TRY - CATCH functionality. Had you mentioned partitioned tables, I would have to eat my words.

I think data compression will become the "sleeper hit" of 2008.

I have been experimenting with data compression, and depending on the nature of your data, it has a lot of potential for databases that tend to have more reads than writes. It will really be interesting to put it into practice in large environments and see how it fares. The sad thing is that it is only available for the Enterprise Edition, which is the case for many of the new features being introduced in 2008.

I have been piddling around with analysis services now for the better part of 2 years.
I really like the idea and concepts surrounding these services, however I really think there are some overall lacking features in AS (analysis services).

The biggest one is that in order for AS to work properly, the dimension tables can't have data that is not in the fact table. I think this is a terrible design and AS should compensate for this instead of making the "developer" delete dim data in order to get the process to function.

And secondly, there has to be better ways of delivering up data to users in AS.
I have worked with proclarity, and that seems to be very nice tool, but even proclarity has to be "installed" on a users system to get the data to do tricks.

Also, I think AS could be a little more mathematically friendly. From performing analysis, it should have some built in statistical functions to find outliers and show apparently interesting things about the data.

That's all I got,
Foxjazz

As a developer/consultant, I am disappointed in the lack of changes for a major release to SQL Server 2008 itself.  I was really hoping for some improvements in the Reporting Services area.  Although there is no reason to not move directly to 2008, there isn't much reason for my clients to spend the money to upgrade now - some of which are just moving to 2005 now.

However, the biggest thing that may get me to move to 2008 particularly for my development platforms is the changes, not on the server end, but within SQL Server Management Studio.  Some of the additions in that tool (code completion, T-SQL debugger, object search, etc.) will make me more productive as a developer.

"lack of advancements in areas such as reporting services"

Tablix, Report Designer, and all of the various visualization enhancements aren't enough for you?  Or did you just not notice that they were there?  Perhaps it would be good to spend some quality time with Google before posting on this topic.

As for many of the other features you mentioned... First of all, it's quite odd to mention DDL Triggers as one of the top SQL Server 2005 features and then lambaste PBM as something no one is interested in.  PBM takes the basic idea of DDL Triggers and shoots it upwards by several orders of magnitude.  Next I should address TDE, which is a -huge- feature for those of us who are somewhat security-minded. I already have customers talking about upgrading just so they can use it to comply with regulations.  The compression features are equally important in their own right, and TVPs answer one of the biggest problems developers have historically had with SQL Server.

If the people you're talking to are only making "polite murmurs of interest" around these features, they probably have not yet bothered actually looking at the product -- at least, beyond posts like these that tell them not to bother.

"Service Pack" or not, 2K8 has been the release we've waited for to avoid the excessive expenditure of a jump from 2K to 2K5 then 2K8.  2K5, for us, had the compelling Snapshot Isolation that Oracle's had for longer than I care to remember.  Having to deal with lock contention by (despised) dirty reads has been too long a necessity.  Before anyone shouts too loud, non-profits have severe resource shortage issues that others may not, including no DBA or SQL Server-versant Sys Admin, and although separate reporting databases are recommended/necessitated "we don't have the resources!" (here endeth the whine).

RS2K freaks out on large Excel exports - 2K8 delivers faultlessly!

Just wrote a trigger begging for MERGE - bring it on...

2K5's CROSS APPLY is a boon, especially for those string-splitting problems - too bad we're on 2K, can't wait for 2K8 licenses, which I heard were priced by CPU, just like 2K5 (dual quad-core server, please).

Native encryption please!  Tired of saying, "sorry, can't decrypt account numbers from the DB - it's .Net code" (yes, there's the CLR but we're 2K...)

Policy-based management - similar thought process to RS Role based, great!  Tired of trying to work out who has what at too granular a level and then grant to the newbie.

Built-in auto-complete - no thanks, tried it, SQL Prompt is much better.

"Facets" in SSMS is VERY informative - had to go lots of "elsewhere's" before - glad it works with old DBs too.

Activity Monitor with "Recent Expensive Queries" with Show Execution Plan built-in - THANK YOU MS!  Probably saved us buying a 3rd party tool to ID them...

Sparse Columns - can't wait.

Data change auditing (or whatever its real moniker is) built-in - more "thanks MS"...

So many opportunities for TVPs it's not funny - bastardized .Net/T-SQL interfaces (the ubiquitous name-value pair or clunky XML in 2K) will finally be a thing of the past.

For RS2K8 I would have liked to have seen report-job chaining/dependency functionality included (maybe it is and I've not spotted it...).  It would ease juggling shared schedules to obviate lock contention and enable me to run related reports as close together as possible.  Would that MS would make the jobs non-GUID named, or at least give us the option to assign a friendly name - what a pain!

List could continue - suffice to say (IMO) for a 2K shop the features and proven stability of the underlying 2K5 core are compelling reasons to adopt 2K8 as the new baseline DBMS.

BTW, a HUGE "THANKS" to Red-Gate for ensuring that the Toolbelt is out with full 2K8 support so swiftly.  Tried to run Idera's Admin Toolset on the test 2K8 instance and "No!" came the stern response...

For those of us still in the 2000 world , it's clear that our planned move to 2005 just had to be reconsidered. Just fine we still didn't buy the product for our production servers , since we'll have to retest all our migration on the 2008 road as we did for the foreseen transition to 2005.

And maybe this is the only point of contention with MS at this point. While 2008 in effect seems to be a massively reworked 2005 , i'd agree its more a marketing thing than a real brand new product. Calling it 2005 SPx would be more appropiate and , to the users , a lot cheaper , since it would be covered by the 2005 fee.

While i'd usually have in comptent the "last version" nutcases , this is clearly more a case of applying a nice service pack than moving on , but as ever , even when applying minor hotfix , you have to test it before moving to production.

btw , Mr. Davis , do you read your inbox mail in this forum? I'm waiting from an answer for about two months ;)

MVV and others:

How many features does a product need to be called the "next version"?  Was SQL Server 2000 really just SQL Server 7.0 SP++?  How about if we use some other program as a benchmark; should Red Gate not call the new version of SQL Backup Pro a new version just because they added only one major new feature?  Is it a percentage thing?  How about SQL Compare?  How many new versions have there been?  In your mind are they all exactly the same?  Where should a company draw the line?

Like everyone else, I was hoping for various features that didn't make it into SQL Server 2008, and some of the features that did make it in have left me wondering what the SQL Server team is thinking.  But to say that it's nothing more than effectively a service pack for 2005 is quite harsh...

Mr. Machanic , I have not an honest answer to that , just my gut feeling.

What should be the milestone where users and MS would agree that it is , in fact , a new product ?

Adam,
As I read what Tony has written, I feel certain that he means it as a compliment to  SQL Server and the team to call it the biggest Service Pack in history. I think that 2008 is basically a healed and smoothed SQL Server 2005. We're all big fans of it here at Simple-Talk as you can see from the articles and books that we're beginning to publish. I'm sure Microsoft feel the same way, which is why we're getting the message that conservative and cautious companies can feel safe moving from SQL Server 2000 straight to SQL Server 2008. As Tony says, everything just seems to work better.

What about now being able to declare and assign values to variables in a single statement? Surely that alone is reason enough to upgrade! ;)

Ted Kummert opened by reiterating the message of Microsoft's increased level of support for the PASS...