What's a software developer to do...

What? You care about the results of your programming? That's so ... quaint. Are you old?

Nah, that's only a sign of experience; you can start worrying a lot, though, if you ever start thinking how some idiot might take your program and misuse it!

That goes for all programming, but applies most particularly to simulations of supernova explosions and financial risk modelling.

On a more serious note, there is an alternative you might try before you get that pointy haircut: You could write software at an engineering company; software driving, simulating, or controlling whatever machines the company happens to be engineering. Engineers tend to be pragmatic about the 'how' of programming, to be focused on the 'what', and to wear any ideology lightly.

Judging by what's annoying you, you've been hanging out with programmer-programmers too much.

Hah. Yeah, I don't have to worry about that as I seldom build anything useful. Right now I'm building an application used to treat travel agents like slave labor. It tracks every second of their time. Who wants to work like that? I can't even believe people do. Simulations of supernova... much more exciting.

You are probably right, Thomas. I'm a little bit stuck in that the technologies that I have the most experience in are not widely used for such purposes. Not that I'm not a quick study, just that I'd probably have to start at a lower rung on the salary scale. Well, that might be ok.

How's your freedom treating you?

Whatever happened to just letting travel agents earn their commissions?

Or is that an old-fashioned notion?

Yeah, these might not be the traditional travel agents you're thinking of. These are people who work reservations off a prioritized queue and who get bonuses based on how many and how fast. Think of it as something likely based in India...

Just out of curiosity: What technologies do you have the most experience with?

Freedom continues to treat me very nicely. Especially now that the sun is coming out and it's finally getting a little warmer.

Java web-based applications mostly, and the various application servers and frameworks that come with that. I've been doing some recent web client work in Flex, which is a nice change, but not exactly floating my boat. In general, I think Java is a rather inefficient language for engineering problems, but I could be wrong.

Thank goodness for the sunshine. I'm going to enjoy that on my furlough days.

Oh, we did use Java for some engineering work. Granted, it's indeed kind of slow, and kind of a resource hog. On the other hand, it's not bad enough in this department to be a show-stopper -- especially as long as Moore's law remains friendly to us. Plus, its a good language for prototyping, and provides hooks for embedding C code in case you need computing speed for number crunching and the like.

So if and when you're looking for a job in the quaint old let's-get-something-done industry, you may well find the market friendlier to your skills than you think.

Gosh, back when the world was much younger, I was a machinist just begging to be told what they wanted, without them trying to tell me how to get there.

I used to build the boring stuff for the people who didn't care how it was built as long as it was built by impossible (to do well) deadlines in the day, and work on stuff I liked at night.

After a few years I had enough apps built on my own that I started a business. The main reason I prefer to work for myself is so that we can do the applications well. I really hate having to cut corners and enjoy now being able to do better work.

I vote for starting your own project that you enjoy if you have the time.

No time right now, but I think that's generally where I'm headed. Either that or back to school.

Roger, one of the reasons I considered donning the pointy-haired wig (hat tip, Thomas) is that I think I'm well equipped to separate the how from the what, and to let the people responsible for the how deal with that question, as well as "when". You can hand a pile of features and a deadline to developers, but you'll only get one or the other in return unless you know how to negotiate.

Thomas, thanks for that tip. I think I will look into it. The market isn't exactly flaming hot, right now, but you never know what might happen.

Of course you see that mathematically it's identically equivalent to simulation of a thermonuclear explosion? I sometimes dream about computer at Los Alamos operating on granularity of femtoseconds - femto. You can watch each individual neutron leisurely hitting a nucleus in that timescale. Tell your travel agents about it, they'll be delighted - nothing in biology works that fast (far as I know, which isn't very far).

The good news is we can paint our cars any color we like in a femtosecond:

http://nextbigfuture.com/2008/02/femtosecond-laser-can-change-any-metal.html

Thomas - you would know exactly, but my impression is that Moore's law has hit some still vaguely defined asymptotic limit; whence the multicore processors with hundreds of cores coming our way.

Duck - not to worry too much about the "what", most programs don't work much of the time. Many of the alleged navigation programs are overrated. This from today's news >



> and that's after 2 other NATO submarines crashed into each other in the middle of the Atlantic, another US sub demolished a Japanese fishing boat while coming to the surface in the middle of the Pacific, yet another crashed into some underwater mountaintop, also in the Pacific, and on, and on.

You'd think with all the satellites, onboard radars, other onboard and onshore equipment, listening devices, just plain eyesight, something would work, but nah!

software developer can do many humanities work for our world.Nice topic.

Whatever happened is that software engineers started building applications that let people find the cheapest airline tickets and hotels with the need for travel agents.

When is the last time you spoke with a travel agent?

Let me jump on the other side of this discussion. The "how" is absolutely important. The "what" gets you functionality and usability. But the "how" gets you maintainability, extensibility and scalability. It is possible for smart hard working programmers to make something work without worrying about pesky details like design and architecture. The problem is after the initial work gets done, each new feature or requirement or problem becomes more and more difficult to fix.

The hot new things are hot mostly for good reasons. Dependency injection is awesome if you care about making things reusable and scalable. Aspect oriented programming is very helpful to make break complex problems into simpler little problems.

I enjoy creating new functionality with cool features and usable UIs. But I also love architecture and thinking about how to use the latest technology to make well engineered code that will be easy to maintain and extend.

This means engineers need to continually spend time learning about the new thing. But it is all worth it when you see how cool the new annotation driven controllers in the new Spring 3 framework.