..do you really have to let us know Mark ? Admittedly I can think of duller things... oh - wait

I expect this will provide me with a series of articles I can use to back up telling my clients why they should wait.

I'm also assuming you aren't looking to get better support from MS, just because you are going public. BTW is publishing performance numbers still a violation of their licensing agreement?

It would be more interesting if Mark had decided to migrate his project to 100% open-source alternatives. Then we, and he, might actually learn something useful.

i would be more than interested to find out what you find. I am about to take over the role of team lead for the largest (4.3Tb) SQL database in Australia, and I know I will be faced with the decision to upgrade to 2008 very soon. its been less than a year since they went to 2005 and it was pain for all involved.

I am taking over this team in the next 4 weeks, and would like to be able to say, we shouldn't upgrade for the following reasons.

Please continue this article, and i will be more htan happy to read them.

I await your trials and tribulations with interest...just don't try throwing any symantec products in there for good measure as you will doom yourself to failure!

Windows 2003 is common now, SQL 2005 less so as most are still on SQL2000.

The only reason to have a production system on the 2008 bleeding edge is to puff out your CV.

It's really not in the business' interests. Go and lie down in a dark room for a couple of years.

I've got production systems running SQL 7 on Windows 2000. They run fine, still no need to upgrade. In fact, server hardware specs dictate that we stick with that, at least for the foreseeable future.

OK, why not color data types? They could store not just numbers, text, binary etc. but CSV type RGB type etc. Why not car model data types? Why not celestial star classification type?

Sorry, but to me they look like arbitrary compound data types. Whenever I see people do this in a database, I think they're trying to avoid writing a server side component by using the database as the complete server. Making everything client-database rather than client-server.

Nice and easy to do, right up until the database doesn't do what they want and then it becomes botch city. Or your project is waiting on a new feature to appear in the database... like spatial datatypes for instance.

I look forward to reading more in this series as we are currently wondering about deploying Server 2008 and Visual studio

I think you should use MapReduce instead.

Tim, I assume your comment was ment in jest right?

If not, keep on reading...

Whilst this may not be the most exciting 'article' on El Reg, it serves far more purpose than most of the shite they are publishing recentley.

I'm looking forward to this, as a developer who is tasked with researching these new systems (BTW, Server 2008 is starting to look very very good) I am always going to be interested in seeing what others think of the products that I am looking at in order to gain a better understanding of the issues I might face.

Keep it up Mark, looking forward to the next installment.

Be bold ditch micrsoft ?

Then he would be an old bold dba and there are non of them stability is king, more than a reason to stick with what you know (execpt the new features)

...to upgrade to Vista at the same time :)

I'm with Stephen - it's nice to see an article about a large, real world project on El Reg.

I'm certainly interested to see how he gets on with 2008 - it'll come to all us MS shills at some point. More please !

And btw, I'd have no problem reading about somthing similar done with an open source project. It's all good.

My comment was rude but funny. Pity they pulled it. But then the echo chamber nature of IT continues to reinforce itself in odd ways.

Wishing you best of luck with the project Mark.

"But, while the decision adds to the workload now, it should reduce it in the future."

No, still laughing.

As to "Spatial data types?"

With ORACLE the spatial cartridge allowed you to do spatial queries, its not about storing the data as such, as all it is is a compound data type, but being able to do some useful queries which arent really all that easy to do.

Use Oracle, its had spatial for 10 years know, and SQL has had it for oooooh nano seconds, assuming it dont get descoped of course.

Mark,

Shame that Light blue touch paper and retire ....By amanfromMars Posted Friday 25th January 2008 11:05 GMT...... didn't appear here for Peer Review, for it was almost something which Bill Gates might have dreamt up if this is anything to believe .... http://www.theinquirer.net/gb/inquirer/news/2008/01/25/gates-calls-revision-capitalism

It certainly painted AI in a novel Sunnier Vista with ITs Programmers in Control and at the Controls.

Wouldn't go there myself but good luck to you and keep us updated

Damn this should be some good reading. I would not touch this with a 20' pole!

Yes we do have some SQL2000 in production (where we are forced to) but we do all our real Database work in Oracle, on Unix. Nothing beats the performance that combination brings for the price tag.

even if it is anecdotal, or a platform you don't have to use somebody somewhere _has_ to do this stuff it doesn't matter if it's crap it's still being used.

"if we don't like problem solving, what are we doing working in computing?"

Pretty much it exactly, it's why you do this shtuff it's not good for your stomach , or your love life, and there are easier ways to make a living. I am glad to see this sort of thing every time someone is brave/foolish enough to do it and the very best of luck to you, Mark. Knowledge is always useful even if it's knowledge of evil (some would say especially then).

According to today's article by Kelly Fiveash (http://www.theregister.co.uk/2008/01/28/sql_server_2008_delayed/), Mark is going to have to find an alternative to SQL Server 2008, especially if he needs to roll out a version of his new IT project before the summer.

Are you sure you don't want to consider an open source alternative for this project, Mark?

Happily, as I say in the next, exciting, episode (which should appear soon), having lived through the release of SQL Server 2003/2004/2005 I'd already allowed for substantial slippage in the 2008 version. This doesn’t mean I am happy about the slippage, but I was prepared. The project will go live in the summer, but only with a subset of the users and we can do that on the beta (sorry, CTP).

"Are you sure you don't want to consider an open source alternative for this project, Mark?"

We have already done so. When the project started we drew up a spec. and then matched that against the available products; some of which were open source. A subset of open source products was potentially capable of meeting the database requirements but not the BI (specifically the OLAP and data mining) requirements.

However, for other projects I am always open to suggestion and the world is constantly changing. So if any Reg. Dev. readers are aware of good open source OLAP engines and data mining tools (in particular those capable of scaling well) please do let me know.

"Killer features continue to appear in software that confer such competitive advantage that procrastination is impossible.

The "it" in question here is the ability to handle spatial data. SQL Server 2008 has it. End of story. Yes, I know Oracle had spatial data types first"

That doesn't give you any competitive advantive as you'll be playing catch-up to the companies already using Oracle (as you pointed out)