Read the rest of Happy 4th of July at SQLServerCentral.com

Someone at Intel blogged that there are some fundamental changes coming for developers in the near future. They are starting to talk about shipping many more cores in their chips, possibly hundreds, and this will fundamentally change that way applications need to be written to take advantage of this.

I have done a little parallel programming in college, and I'll say that I didn't really understand it that well. Moving to multi-threaded architectures and synchronizing threads was something that seemed esoteric and in its own niche twenty years ago, but I bet that by the time I retire the vast majority of computers will have more than 8 cores and I'd hope that most software will take advantage of this situation.

But if you view IT as a strategic group, an asset to the company, than it must make the business better by more than just reducing costs through automation. At that point IT must really find ways that not only provide more benefits than it costs, but must really drive the business forward.

But do we really have a lot of information?

There's a fundamental difference between data and information. I could give you data like 1, 2, 3, 5, 8, 13, 21, 34, but that doesn't necessarily mean anything unless I can use it. anyone recognize that sequence?

A month or so ago I got a shipment of LED bulbs from the C. Crane Company, from one of the IT people there that saw these updates. They wanted me to do some testing of the bulbs and see what I thought. Now one of the things that I've done out here on the ranch is replace some our bulbs to try and get more efficient in how we use power. We have compact fluorescent (CF) bulbs for a number of our outdoor lights that run at night, solar powered lights in a few places like on the shed, and I'm wiring an energy efficient sodium light for the barn. However most the internal house lights have dimmers installed and so CF bulbs don't work well with dimmers.

Wikipedia runs a large site that serves lots, and lots of requests. In fact, they get a decent number from me when I'm researching editorials and my kids think "everything" is on Wikipedia, their first site of choice for looking things up.

They have a nice article on their infrastructure, which includes 80,000 queries / sec. That's a pretty high load, higher than I've worked on for one service. However unlike Google, Microsoft, and others, they don't have a huge budget to spend on hardware. It's an interesting look at how they do a lot with a little and might be something those smaller companies out there should think about.

So you are reviewing service accounts and SA passwords. For this Friday's poll, how do you handle them?

So what's my pickle with this title?

First I know that many of you wear both hats. I've done it and despite all the "best practices" that say to separate the developer from the production guy, it doesn't always work. Heck, if your company agrees to hire a DBA are you going to complain you really need two so that one can handle each side?

One person mentioned that you should definitely have someone else do your job. In this way you not only have a backup, but a built-in auditor. If someone does your job well, then they can see if you've been skirting rules or even flat out breaking the law doing something like embezzlement.

That was a interesting point and something I hadn't thought of. If you constantly avoid a rule required by SOX, HIPAA, or some other type of regulation, it's entirely possible that someone doing your job for a week or two might find this out.

Most of us in the IT world make a good living. I know that some younger people might not realize it, but think (or ask) what your parents make, and how quickly salaries have risen in IT. Not that we don't earn it, but we're well paid, at least in the US as compared to many other fields.

These things aren't always exclusive. You can build/implement an effective solution, which could also be the best solution, but if there's a conflict, I've always looked to be effective. By this I mean that I try to get the business purpose completed and move on to the next task.

I'm as bad as anyone, and not having had an HR department for 4 years, I've rarely taken a day off in that time. Including weekends! Sure I've had some vacations with the family, but more often than not I've had a laptop to check on work, get the newsletter out or something else. Last summer I went to Mexico and I think that was the first time I'd been unwired for more than 2 days in 4 years.

How many of you are bulldogs? Do you stick with a project that's not providing business value, especially as costs rise? It's not something that I necessarily see many development, or even implementation, groups doing. Once a project is funded, it usually goes to completion or until it fails; I rarely see someone kill a project at 40%, 50%, or 60% completion.

I've had quite a few people ask why to bother with this. It's not so much that I think software development needs to have a model, but rather that I think there is value in looking at the way things are done in the IT world against the way things are done in other professions. Maybe we can do better, perhaps we can improve our processes by looking at what works with other groups, and potentially we could avoid mistakes other professions make. So today I talk about engineering and software development.

I've heard that it referred to many times that doctors "practice medicine" and that it's as much an art as a science. After all, while we are all similar biologically, there are unique differences in many of us, and so many of the diseases and other work doctors deal with are very similar so a really good doctor doesn't just play the odds, but tailors his treatments to the individual.

At first I scoffed at his idea, but the more we debated, the more it made some sense. After all, very few of us outside the law really understand the nuances of how to build legal arguments (which often don't make sense), and we can't specifically state what final product we expect from a lawyer.

So today, my plan is to actually compare software construction to building construction, and talk about the common and uncommon points as well as encourage debate. Over the next few days, I'll pick a few other professions and we can debate those as well. Perhaps we'll come up with a good model, or perhaps not, but I'm sure many of you will view software design differently at the end.

Bill Gates talked about Windows 7 being released next year, and while most people dismissed that as vaporware, I'm seeing more and more news about the next version of Windows, which could (hopefully) put Vista to bed soon. Or at least, next year.

At the D6 conference in California this past week, there was a demo, showing some of the "Surface" or multi-touch capabilities that will be included in Windows 7.

Multi-touch is intesting and let me tell you that if you haven't played with it, it's cool. I recently got an Ipod iTouch, and I definitely like the ability to flip through things with my finger. However it's not necessarily overly exciting or useful after the first few minutes. I would like to have the ability to move things around on the screen with my fingers, click links, zoom, contract, etc. on my desktop, but I'm not sure how much I'd use it. A mouse felt really natural to me and made me more productive immediately, but moving from keyboard to screen and back? Not sure for us IT folks. For my Mom, it would be cool.

Database Administration is a tough job, it’s difficult to explain to others, not well understood by many employers, and very under appreciated. These awards are our chance to recognize DBAs as well give them a little limelight and reward, and showcase their efforts for the community. This will hopefully be an annual event and give some recognition to those in the profession for the work they do.

In the past updates, I've seen many people suggest diesel cars as an alternative to hybrids. VW is bringing back the diesel Jetta to the US, which is a great idea. I don't really like the Jetta, and wouldn't have purchased one, but when VW stopped bringing them to the US, I thought that was a mistake. The cost of diesel is high now, $4.10-4.20 in Denver v 3.60 for gasoline, so there's a 12-14% difference in MPG to even things out. However a 60mpg diesel is a great buy for now.

I often see a knock on Microsoft that they don't innovate at all and just buy/steal/whatever someone else's ideas and products and then market them extremely well. There's definitely no doubt that they market well, but I'm curious about what you think about innovation. So:

I haven't heard if this will be in SQL 11, but I'm sure that at some point we'll get some of this technology in the platform. It's something that many people would like to see and I can see the marketing people having a field day with the idea of "Matrix technology" embedded in SQL Server.

You are. IT's listed as one of the hottest IT jobs in 2007 and the US Bureau of Labor expects that DBAS demand will grow 37% from 2006 to 2016, a much higher than average growth rate than other occupations. Average is seen as 7-13%.

SQL Server is getting easier to use, but there's apparently still a high demand for DBAs. I realize that many of the statistics would include Oracle and DB2 (and other) platforms, but I'd think that with all of these platforms growing, there's be a pretty good demand for people across the board.

There are definitely times that I think the people that build the system stored procedures and other code we can see don't really work with SQL Server. It seems that best practices aren't followed and things aren't even well coded.

But most of the people I've met working on SQL Server seem to be really, really smart guys. They care about what they do and I'm somewhat amazed at times how passionate they are. I can't name everyone, but I'll point out one that really struck me.

This article lists 8 ways to bridge skills gaps in IT, and the first one is training. That's important and really requires some type of committment from the company. It's not something you find money for, but something that you should make a part of every worker's job. It's a big reason that I started End to End Training with Andy Warren and Brian Knight. We believe that training is important, and that it needs to be practical education for your job, not for a test.

I really like the idea of finding workers, grooming them, giving them a great environment in which to work, and trying to keep them around for 10 years. In today's mobile workforce that can be tough, but I still believe people are people and many of them would prefer to stick with the company they know if they're treated well. Just look at how many people stick around when they're not treated well. You know that you could keep many more with just a little effort.

That's not the case as I've seen large tables lock up applications as the changes are made. However I've not seen a lot of testing on how performance is impacted by different changes. Ken Kaufman made some tests that I found interesting when looking at column changes.

Most of us don't start work with a brand new company and get to build the systems from the ground up. I once started a company (this one) and even then we ended up using some existing code and had to make do with what was there. We often have to work with developers or DBAs that have been making changes to the system in their own way, and we have to fit in.

I started at a small company one time that had been around for about 4-5 months, and they had a small e-commerce system in place. I was hired as the first DBA for the existing 3-4 developers and was amazed by how much stuff they already had in place and how difficult it was to implement just simple controls to make things more repeatable and manageble for everyone.

That's a great quote from this article on data encryption and I think it's a great one to keep in mind. The actual process of encrypting the bits is probably the easiest part, which is why I think encryption is not that widely deployed.

Key management is a hard process to understand, much less implement. I've always been more concerned about that part, especially as we move to complex, multi-part keys such as public encryption. Managing the keys, ensuring they are distributed to the correct person, they can be revoked, they can be replaced if lost is a major issue.

That's a great quote from this article on data encryption and I think it's a great one to keep in mind. The actual process of encrypting the bits is probably the easiest part, which is why I think encryption is not that widely deployed.

Key management is a hard process to understand, much less implement. I've always been more concerned about that part, especially as we move to complex, multi-part keys such as public encryption. Managing the keys, ensuring they are distributed to the correct person, they can be revoked, they can be replaced if lost is a major issue.

Obviously the followup question is then "why not?" There were various reasons given, but most of them didn't really resonate with me. As I listened to the various answers being given, there was one main thing that struck me about the CLR.

OK, occasionally I'm not doing anything, but I'm often trying to actually get a few things done. This topic, however, has got me a little stumped and confused, so I'll put it out there for you all to debate and convince me one way or the other. Or convince someone else. Or maybe even get Microsoft to think about what they are doing.

Whether that's a good idea or not, other industries, as mentioned here, are using RFID to track the movement of large numbers of discrete items that get moved from place to place.

And that means lots of data!

In any case, it appears that Microsoft is thinking ahead and looking to keep some of the Yahoo talent and not just buy the technology and properties. They're considering setting aside a large amount of money to ensure that they retain employees.

I saw this article on 10 Ways to Help Keep Your Job Under Control and I was thinking it would have advice, but it's really a bunch of software utilities that mostly seem to do with either email management in some way, project management, or just general organization. If you're thinking you need something to help manage your life, I'd flip through the list and see what you think.

It was at the end of the rant that I saw this: "Why I really care is that Microsoft is vacuuming up way too many programmers." and "...between Microsoft and Google the starting salary for a smart CS grad is inching dangerously close to six figures and these smart kids, the cream of our universities, are working on hopeless and useless architecture astronomy..."

I saw a few announcements last week about issues with IIS, which morphed into reports about how this was a SQL Injection issue and not anything to do with IIS.

I think Aaron Bertrand has it correct, this is a SQL Injection attack, but if you were more careful about building your applications, this wouldn't happen.

I hate to keep beating the security drum, but until people build in security from the beginning, and prevent SQL Injection attacks, this type of thing will continue to happen.

One of the main topics lately seems to do with database sizing. I see questions on how to estimate space, why is the log so large, how do I shrink a database (you shouldn't), and other similar questions.

So many very large databases are warehouses of some sort. The Winter Corporation has a Top Ten program that tries to survey companies and find out who has the largest database. The latest results I've seen are for 2005, so I expect that things have grown since then. In that survey, the Land Registry had a 23TB OLTP database (DB2) and Yahoo had a 100TB data warehouse (Oracle).

While stockpiling some dead trees the other day at Borders, I noticed a display for the Sony EReader, which is an e-ink based electronic reader. It was interesting, small, and easy to read, but the reviews mainly complain about the horrible selection of books. Glancing through the list I see it's relatively small list of titles, heavily weighted in the McGraw-Hill area for technology.