Is SOA A Bloated Dollar Train?, Code Generators, And Starting Down the MCSD .NET Road
Rocky seems to think that SOA has the potential to be a "big train wreck." I generally agree with his statement, but I think it's already become a huge buzzword with little meaning behind it. It seems like more people (architects especially) are using the word "service" somewhere in their corporate meeting vocabulary every day (I'm also guilty of this myself), and it's a little disconcerting.
I'm finding it interesting that there's more talk about code generators. Personally, I'd like to move it up a notch and look at more model-driven approaches. Most developers seem comfortable with a compiler writing their CIL (with the runtime generating the x86). It seems like the next inevitability with be modeling applications and letting a tool generate the code that is compiled into CIL that the runtime munges in x86. Of course, the middle layers can always be optimized, but I'm wondering when the majority of code with be generated and not written. That's when the real fun will begin. Granted, people will argue that a tool can't write more efficient C# code than what they can write, but I bet they haven't even tried to write an application in CIL, and...well, why would you? That's the mundane part that's highly automatiable (if that's even a word ;) ). It's just a matter of time before people will just have fun writing code via modeling tools and the resulting code is just secondary.
Finally, I'm starting down the MCSD road...again. I got it with the pre-.NET tools a couple of years ago, and now I'm going to update it for .NET. My first test is next week Thursday. The only reason I'm doing this is that it benefits my employing company's partner status, and therefore it benefits me ;). To be brutually honest, I wouldn't even consider doing this as I find MCSDs rather annoying and not worth the time. After doing some initial studying and practice tests, I wondering why there's so much stress on data binding! I've never used in on any .NET project I've been on yet, and while I realize data binding has been vastly improved from pre-.NET to .NET, I still don't find it hugely beneficial. When I took the Java certification years ago, I felt like it proved that I knew the basics of Java. MCSD tests don't prove much to me.
* Posted at 04.08.2004 08:16:38 PM CST | Link *