SQL Server 2000 is losing ground -- Where the H**L is Yukon?

Ok, with the recent news Yukon is going to be delayed (again!) until 2005 I really have to ask what the SQL server team is thinking delaying this product this long?  I have seen demos and quite frankly it looks amazing, but, it is now due early 2005.  Some of the things in SQL 2000 which looked really good when it was released are starting to look quite frankly pretty clunky by todays standards.  These things include the Analysis services UI, Partitioning Support and DTS.  Quite frankly some of the other vedors are starting to get it right and the ease of use of some the new tools just blows the SQL 2000 design environment out of the water. 

Microsoft, get this thing out the door.  I don't even care about CLR functionality at this point, I just want a more usable product for delivering business intelligence functionality.  You can give me the CLR bits later.  Management is starting to ask if SQL was the right decsion? 

http://weblogs.asp.net/sjsmith/archive/2004/04/13/112845.aspx

I would rather wait for the database to be written the right way and be more sucure, then rush it out and have tons of problems after its released.

Scenario 1: The released a buggy application that needs a bunch of patches, its unsecure, etc...

Scenario 2: They are taking to long to release it. Its delayed again.

Either way, soemone is going to be upset by the outcome, just have to look at it from a MS point of view i guess.

Personally, i'll wait, but if i didnt want to wait, i would just go with another database.

I can see both points.

5 years is along time between versions of SQL Server. Oracle 9i & 10g are pulling ahead.

I would rather see a mature product brought to market that has gone through ALPHA, BETA & Dog-food, than early just to stay ahead. I think Microsoft have made that mistake before.

It is so important that Yukon is an Enterprise class database if Microsoft want to use it as the flag-ship that takes them into the heart of the Enterprise.

I think we are going to have to be patient. Rome wasn't built in a day. But there is a down-side, Microsoft better not decide to faze out SQL 2000 in a hurry after Yukon is launched.

Don't get me wrong we can get the job done with SQL 2000 in most areas. I think what I am looking for is the big bump in productivity other tools have come out with in the same time period. The one feature I really want is an expanded Analysis Services experience. It was a very nice version 2.0 release when it came out in SQL 2000. Now it is 2004 and we have not realized any productivity gain in this area for 4 years! This is becoming unacceptable, I could not imagine other vendors not shipping a new version of product in the BI space for 5 years when it is all said and done with Yukon. I expect more from Micrsoft, they have been the big productivity booster in the past, now we are looking at buying other tools in the near term to do more with less people. The reality of 2004.

sjsmith wrote:

Ok, with the recent news Yukon is going to be delayed (again!) until 2005 I really have to ask what the SQL server team is thinking delaying this product this long?  I have seen demos and quite frankly it looks amazing, but, it is now due early 2005.  Some of the things in SQL 2000 which looked really good when it was released are starting to look quite frankly pretty clunky by todays standards.  These things include the Analysis services UI, Partitioning Support and DTS.  Quite frankly some of the other vedors are starting to get it right and the ease of use of some the new tools just blows the SQL 2000 design environment out of the water. 

Microsoft, get this thing out the door.  I don't even care about CLR functionality at this point, I just want a more usable product for delivering business intelligence functionality.  You can give me the CLR bits later.  Management is starting to ask if SQL was the right decsion? 

http://weblogs.asp.net/sjsmith/archive/2004/04/13/112845.aspx

That is exactly what they did with Windows ME and look at how it turned out.

They are not going to do that again and pretty much everyone prefers that it be done before it is released.

Not to mention, a finshed version of Yukon is going to be more effective at competing than an unfinshed version that is later completed as a service pack.

The DTS experience could be much better also. I have only seen a few bits of what this will look like in Yukon and I must say I can't wait for it to get here.

Take this for what its worth... it's late and I've been reading "moon in a dewdrop" again.

It really amazes me how many people just don't get "it" about the CLR in Yukon. What if I were to tell you that us mere mortals having access to the CLR is a side effect of Microsoft using the CLR to realize many of the functional features like the UDM, XML as type and DTS. It's certainly the bedrock of Reporting Services. Yet, so many would have Microsoft burn away the CLR functionality just because Oracle 10g could potentially be more promising or because other tools have newer UIs.

They never had to give us access to the CLR in Yukon, you know. We would never have had to know it existed. The product might still be delayed anyway.

Yes, they are making the right decision, but it seems either you get it or don't. I suppose its the fault of folks like me who are trying to evanglize Yukon the way that others are Whidbey for not communicating the message clearly enough.

As to managers asking if SQL Server is the right thing or not, I'd want to know why they went to SQL Server in the first place. If it only was for the BI platform, well, that's an interesting choice. True, it hasn't evolved as quickly as even I would like, but also true, IMHO, that few other db-vendor offered packages have either. At least, not by direct development. After all, its easy to buy an analytics company with one hand and join with the DOJ in beating up MS with the other for being a "monopoly."

This isn't a sprint, this is a marathon. There's plenty of room in the market for scale up and scale out. One size doesn't seldom fits all.

UDM UI and DTS UI are written in Managed code but not the plumbing (although they can be used from managed code) and the XML Datatype is not written in Managed Code.

Almost all of the UI for the product is written in Managed Code in addition.

-Euan

Ok, I want a good stable product.  I suspose I just want a good stable product more offten then 5 years apart.  At the enterprise level we are being forced to squeeze every last ounce of productivity out of the tools we own today.  If tools are driving down development cost or delivering lots of end user satisfaction then there is a problem, period.  From my point of view Analysis Services looked like a sure thing bet for the future.  The reasonably quick iterations beteen 7.0 and 2000 provided lots of good enhancements.  At this point my disapointment comes from these facts:

1) We are still struggling to bring OLAP to the masses.  However the design experience in SQL 2000 was nice back in 2000 but 5 years?!
2) The integration between Analysis services and Excel in the XP timeframe got lots better.  Then 2003 shipped and we got nothing....
3) SQL reporting services was a great add-on.  Don't get me wrong I was in the launch video, but we still need an end user power query tool. (Don't even get me started about MS Query in Excel).
4) Partitioning...  yikes.

I think at this point I would have been happy with the 2000 product with some productivity enhancements.  A more incremental approach where functionalty improved ever 2 years would have bee welcomed.  (Yet, again don't get me started about the SA value we have received from SQL 2000.  I agree reporting services came out but how many people run the web gui on SQL server?? without extra licenses.)

Shawn

EuanG wrote:

...the XML Datatype is not written in Managed Code.

I know you know that, but I could swear that I read that such was the case elsewhere in the PDC bits.

sjsmith wrote:

Ok, I want a good stable product.  I suspose I just want a good stable product more offten then 5 years apart.  At the enterprise level we are being forced to squeeze every last ounce of productivity out of the tools we own today.  If tools are driving down development cost or delivering lots of end user satisfaction then there is a problem, period.  From my point of view Analysis Services looked like a sure thing bet for the future.  The reasonably quick iterations beteen 7.0 and 2000 provided lots of good enhancements.  At this point my disapointment comes from these facts:

1) We are still struggling to bring OLAP to the masses.  However the design experience in SQL 2000 was nice back in 2000 but 5 years?!
2) The integration between Analysis services and Excel in the XP timeframe got lots better.  Then 2003 shipped and we got nothing....
3) SQL reporting services was a great add-on.  Don't get me wrong I was in the launch video, but we still need an end user power query tool. (Don't even get me started about MS Query in Excel).
4) Partitioning...  yikes.

I think at this point I would have been happy with the 2000 product with some productivity enhancements.  A more incremental approach where functionalty improved ever 2 years would have bee welcomed.  (Yet, again don't get me started about the SA value we have received from SQL 2000.  I agree reporting services came out but how many people run the web gui on SQL server?? without extra licenses.)

Shawn

I, on the other hand, really struggle to keep with the testing and costs of licesing a new version of SQL Server, so I've benefited from the respite of not having a new version of SQL thrown at me for five years. But then, we're small by comparsion, five servers supporting maybe 500 applications for 3500 users. And we don't use Analytics at all. We are in the same boat about having to squeeze the services well past the bleeding point, though. The stability and predictability of SQL2000 has been a godsend in that sense.

I also suppose that as far as product releases go its very easy to go from nothing to something, but its very hard to go from good to really good or to great.
One thing that I'm hopeful about systems like Channel9 for is that folks on the Office Team will see your comments about the non-improvement of Excel as an Analytics client and take that to heart. I'd be hard pressed to know what more you could really need in Excel, maybe you'd care to share that list?

There's also a significant problem for many of us in giving the users a more powerful query tool. We can generally express the need for growth in services on a per-project basis, but making such systems really function well as both OLAP and OLTP systems is a much harder sell because management doesn't "get it" about the value of data mining. Rather than giving them a great query tool -- which would make my job easier in the long run -- I'm all to frequently forced into a "oh, we'll just write some reports for them." No evolution of SQL Server is going to solve that problem in and of itself.

Finally, we seldom partition: we just don't need it nor could we deal well with the complexity and latency it would introduce.

Lastly, nobody that I know of and helped with RS runs the RS Services on the meta data hosting database. We also seem to be able to putting them on a dedicated server. Somehow. 

I think we represent the traditional market for SQL Server, and you, Shawn, represent the Market where MS wants to be. Its gotta be tough pleasing both of us.

sjsmith wrote:

Ok, I want a good stable product.  I suspose I just want a good stable product more offten then 5 years apart.  At the enterprise level we are being forced to squeeze every last ounce of productivity out of the tools we own today.  If tools are driving down development cost or delivering lots of end user satisfaction then there is a problem, period.  From my point of view Analysis Services looked like a sure thing bet for the future.  The reasonably quick iterations beteen 7.0 and 2000 provided lots of good enhancements.  At this point my disapointment comes from these facts:

1) We are still struggling to bring OLAP to the masses.  However the design experience in SQL 2000 was nice back in 2000 but 5 years?!
2) The integration between Analysis services and Excel in the XP timeframe got lots better.  Then 2003 shipped and we got nothing....
3) SQL reporting services was a great add-on.  Don't get me wrong I was in the launch video, but we still need an end user power query tool. (Don't even get me started about MS Query in Excel).
4) Partitioning...  yikes.

I think at this point I would have been happy with the 2000 product with some productivity enhancements.  A more incremental approach where functionalty improved ever 2 years would have bee welcomed.  (Yet, again don't get me started about the SA value we have received from SQL 2000.  I agree reporting services came out but how many people run the web gui on SQL server?? without extra licenses.)

Shawn

Be careful what you ask for, you just might get it.

If Microsoft was to ship Yukon now, there might be a devastating bug in it (for anyone that has used MySQL 4.0.x, remember the table corruption bug that was in it until 4.0.15?) that would completely ruin it while if they shipped Yukon when it is ready, they would probably find and fix it.

That is just an example of what could happen if it shipped early.

Shining Arcanine wrote:

Be careful what you ask for, you just might get it.

If Microsoft was to ship Yukon now, there might be a devastating bug in it (for anyone that has used MySQL 4.0.x, remember the table corruption bug that was in it until 4.0.15?) that would completely ruin it while if they shipped Yukon when it is ready, they would probably find and fix it.

That is just an example of what could happen if it shipped early.

Too true. I think we all care too much about our data to put them in what would essentially be a beta product masquerading as a full release version. If Microsoft releases a poor version of IE the worst thing that could happen is we couldn't access the internet, we wouldn't lose anything but we wouldn't gain either. The worst thing that could happen if an SQL server screws up is all the data is corrupted or wiped out, could cost thousands could even end the company!

There is nothing wrong with SQL2k, it seems to do what it was designed to well. So until the new version is released they should stick with SQL2k and if that doesn't suit their needs they should look at alternate vendors (which there are lots).

Are you maybe getting confused with Date, Time and UTCDateTime? They are written in .Net Languages.

-Euan

EuanG wrote:

Are you maybe getting confused with Date, Time and UTCDateTime? They are written in .Net Languages.

-Euan

Nope, I knew those too. Dr... Err... Michael Rys chimmed in with this, too: http://sqljunkies.com/WebLog/mrys/archive/2004/04/15/2075.aspx

Manip wrote:

There is nothing wrong with SQL2k, it seems to do what it was designed to well. So until the new version is released they should stick with SQL2k and if that doesn't suit their needs they should look at alternate vendors (which there are lots).

Wow, two things I agree with Manip about!

So are you saying it was the Tim Anderson interview that led you to believe it was managed?

EuanG wrote:

So are you saying it was the Tim Anderson interview that led you to believe it was managed?

No, Euan, I've not read/seen the interview yet. That is was an .NET-Defined type is a (mis)perception I've had since last August or so. Michael's comments helped me understand more about it, as it did yours.