Does Microsoft communicate more ... or do we learn more? | The Coffeehouse

Sabot
My name is Dave Oliver. I'm a Technical Architect.

Today I was on a course at Microsoft UK on SQL Server 2008 upgrade where it struck me that wooooow there is sooooo much new stuff in SQL Server 2008 it isn't funny ... and with a new version, SQL Server 2008 R2, just around the corner more good stuff is coming.

Now I've been using SQL Server 2008 for a year now and I learnt some new features I didn't know existed (like TABLIX) and didn't realise that *= & =* joins had been deprecated amongst other things. All of which I found that it would have been handy to know some of these things over the last year.

OK so there is some nice tools like the Upgrade Advisor that will tell you but how do you know about these tools? I mean recently a colleague spend ages writting a Access 2 SQL process when the SSMA tool does it very well, boy was he miffed when he found out!

So to the point, Microsoft doesn't have a hope in hell of telling us all about new stuff in it's products ... and all that info we just aren't going ot retain. So, how do we go about finding out if there is something that can help us?

They answer is one of theses,

Bugging a colleague and asking if they know,

Throwing ourselves at the mercy of a search engine ... and hope you used the right words

Posting on a forum and hoping someone knows the answer.

Plough though a book

Paying for it in the shape of a course or consultancy.

Wait for the C9 video!

In this day and age of knowledge management, is this all the answers we have? They are all very random and hap-hazard? My moan is .... there has to be an easier way! What are your suggestions?

As you can see this would be in eveyone interests if we can crack it ...

My idea as a starter for 10 would be a version of BING just with SQL Server info, where you could access it from a right click context menu option when hovering over a feature or highlighted command and that it gives you the top 5 pages. The page titles could be 'Warning this feature has been deprecated!" ... then you click on it, it loads to tell you more! Good idea?

Bass
www.spreadfirefox.com/5years/

I don't view it as a good thing that a piece of software who's fundemental purpose is to store and retreive data is so complex you have to create a post like this. In a time when databases are becoming fatter and more complex there are some people who yearn for simpler, more transparent databases.

I know I am going to piss off some DBAs, but I really hope that entire career field goes the way of the dodo.

Sabot
My name is Dave Oliver. I'm a Technical Architect.

Nothing wrong with little-weight databases, they are perfectly ok and good at what they do and what they are intended for.

However those things just don't come close to solving all the data needs of many organisations, mine included. Such as billing 10 million customers a day and moving 5 GB of data per hour, all of which has to be reported on instantly, daily, weekly, monthly and for 7 years in the past, spotting trends and patterns on the way all of which is guaranteed to work on kit from HP, Dell, Bull, IBM, EMC, Brocade, Toshiba, Cisco, Oracle etc. Good luck doing that on a little-weight database!

Agreed, that the DBA's role is under threat! Why because it's becoming so complicated that those guys are going to have to specialise.

Organisation such as Google rule because they derive knowledge and meaning from all the data they colate. Knowledge is power so you have to look the after data with respect, something you would be wise to learn.

Bass
www.spreadfirefox.com/5years/

In reply to Sabot

#Sep 23rd @ 2:52 PM

I like your enthusiasm but it might be a little misplaced, IMO.

I agree with your last paragraph. Google is a great example of an organization who has a lot of data, and does some fairly complex processing on that data. Most of how they do it is secret.

But not everything is. So we have can study Google's methods. Google is a huge, complex system as a whole. But at the same time they are deceptively simple about it. Do you need to be a DBA to search Google? No!

This simplicity is at the core of everything Google does, including the interfaces that are exposed to the majority of developers internal and external. Google's parallelization algorithm, MapReduce, has an interface of exactly two methods, Map() and Reduce(). Should your algorithm implement these two methods, they can be transparently scaled to any number of processor cores, with fault-tolerance and everything.

What database does Google search engine run? Oracle? SQL Server? MySQL? None of the above. They use their own database software called BigTable. BigTable is actually partly the inspiration for Drizzle. BigTable is designed to be ridiculously fast, and ridiculously scalable. It's also machine invariant. But it's so simple it isn't even closed under relational algebra. So standard SQL doesn't work on it. It's a quazi-relational database.

The beef of what Google does has nothing to do with their database. It's a giant artificial intelligence project mostly. I think your enthusiasm for databases is really an enthusiasm for artificial intelligence.