Database Underground | Sean McCown » My Talk with Microsoft

December 27, 2006 | Comments: (0)

My Talk with Microsoft

TAGS: SQL Server

OK, if you guys remember, I told you that I'd be talking to MS about the tools they provide for DBAs, since they don't seem to be working out very well. Well, I think it was friday I finally got on a call with them, and here's basically what came of it.

Their research has made them choose a certain path, and they do recognize that they didn't get it exactly right. However, they made a sincere argument that they want to get it right. They swore to me that they're dedicated to making tools DBAs can actually use. We got into a couple specifics that get on my nerves, and a couple things that have actually crippled me; they were very receptive.
You have to understand though that whenever you compile a feature list, you have to balance effort vs return. It's simple math really. Should we spend 3mos and 10 devs writing this piece of code that 5 people are asking for, or that same amount of time writing something that 500 people are asking for? Do you write a really cool feature into the product, or do you fix what's already wrong with it? These are decisions every software vendor has to make for every release cycle. And it's a fine line to walk. From what I can tell, MS at least tries to do things right. They pull in customers and run them through the product and give them a set of tasks to perform. They then write the GUI based off of these results (and others I'm sure).
My big question now is since the GUI isn't living up to what it should be, is the test flawed, or is the general DBA population filled with idiots? I don't think they're pulling in MVPs for these studies. Maybe they are, I don't know, but it's always been my view that if you write for the real professionals, everyone else will fall in line. Don't write for the LCD, write for the ones who really know what they're doing. If you think about it, it's the smarter way of doing things. Which do you think is a better way to feed an industry... to write to the masses who don't know what they're doing, and force the real pros to work like them, or to write to the pros and show the masses how they should be working? I would think that writing to the pros would be better for everyone. Sure, the masses won't get it at first, but as they get better, they'll see that things are better this way. There's always the risk you run of scaring away your user base though. A certain amount of the population will run to the vendor tools that dumb everything down for them. That's fine, let them go. If they ever pick up a book and learn something they'll be back.

All this has me thinking though. What are the qualities of a good tool? This isn't necessarily DBA specific either; it goes for any kind of tool. I would say that the really good tools make you feel comfortable no matter what your skill level. They grow with you. If you're a novice, you can get your job done. As you grow in your career, there are things to discover in the tool that will make things easier for you. Yeah, that's it. A tool has to grow with its users. It has to have a good mix of GUI, menu, wizard, command-line, scripting, reporting, extensibility, etc. It all has to be there. Early in my career I found an MVP and asked what tool he used to admin SQL. He said he used Query Analyzer. I then had him give me a few of the things he felt that it gave him over anything else, and he gave me a list of about 10 things, most of which I'd never heard before. So I started using Query Analyzer and even though I was a little young in SQL to really get the appeal, I slowly started to see what he meant. In short, whatever the real experts are using is what everybody should be using. And if the those experts aren't endorsing your tools, then you don't have good tools. I talked to many of the MVPs at PASS, and none of them are very happy with the DBA tools for SQL Server.

Here's what we ultimately decided. I'm going to have a monthly call with MS to discuss issues with the tools. So, if you guys have any feature requests, feel free to send them to me and I'll champion them if I think they're worth anything.

So send them to me at sean_mccown@infoworld.com. We'll have other topics of discussion as well I'm sure, because we did discuss a couple other problems in other areas.

Ok, so that's the report.

Posted by Sean McCown on December 27, 2006 07:12 AM

RATE THIS ARTICLE: