My G+ Profile
« May 27, 2004 | Main | June 01, 2004 »
May 28, 2004
.NET at DABC
Dave Fletcher has a report on his blog about some of the technical choices by Utah State IT departments. He reports Brad Brown at ABC is pushing .NET while other departments are moving toward more Linux (which, with the Mono project could still be using .NET, of course). I think this is instructive.
Brad runs one of the most heavily retail environments in State government. He's got stores and a warehouse. Most of his infrastructure is built on Windows because the business driver, supporting retail operations, pushed them in that direction. In that environment, it makes perfect sense to invest in .NET on the back-end. There's tremendous leverage there.
I've long held that large organizations are going to have a difficult time saying "we're a .NET shop" or "we're a J2EE shop." Most IT organizations are going to be both. The beautiful part is that it just doesn't matter. With SOAP and WSDL, I can integrate a .NET app just as easy as I can integrate a J2EE app. So, I think Dave's right about standards, but I think the State can move beyond thinking about a standard on development platforms and get to a true reference profile that says "no matter what development platform you use, here's the standards you'd better be prepared to support."
An example of why this works is TCP. Back in the day, there were big arguments about what hardware to buy based on what network you used. Now we buy any hardware and know it will talk to the network. TCP solved these problems at one level in the network stack. Web services does the same thing further up the stack. We have the freedom to stop talking about .NET vs. J2EE vs. Whatever. That's a good thing.
02:09 PM | Recommend This | Print This
Do It Yourself IT
Doc is talking, over at his new IT Garage, about Clayton Christensen's Innovator's Dilemma and asking how open source fits into the equation. Doc's rif is that open source is an example of demand supplying itself. That's a different model than the traditional "vendor builds and sells products--IT shops buy them" model that we're accustomed to. Doc calls it DIY-IT for "do it yourself IT." Hence the "IT Garage."
To get the DIY-IT model, however, you have to understand how its different from rolling your own accounting package, which is something we've mostly, thankfully, moved away from. In the DIY-IT model, "yourself" doesn't mean "your IT shop" so much as it means "users." Open source is about the users taking responsibility for meeting their own demands and being empowered (through distributed development tools and processes) to do so effectively.
At the CTO breakfast this morning, we were discussing the pace of change in IT shops these days and how that's necessitated a move from the old-style notion of developing ERP apps in-house to buying packaged apps. Perhaps its the other way around--the trend has enabled and increased the pace of change. We're not talking about vendor induced churn here, but needs brought to the IT shop by the business.
I see DIY-IT as another step along that path. An agile business can't afford to wait for some vendor somewhere to meet its demand. It must supply its own demand quickly, without the expense in time and money of doing custom development. Open source gives business the power to get needed systems without having to custom build them. I believe that feature, more than price, gives open source its power.