>More than 20 years after the invention of the IBM PC (and the Apple Macintosh, for that matter), why is it still so difficult to build robust applications?
>
>MS would have developers migrate to .Net. On top of learning a new language (for pretty much everyone except C++ programmers), learning a large class "framework" seems to be required. It looks far from "easy" and is likely to leave a lot of developers behind. Superficially it looks like Kevin McNeish has an interesting niche with his "Mere Mortals for .Net" framework; I don't know how much it helps out learning the .Net classes. And, Mono aside (I'm not sure how "compatible" a non-MS runtime can ever be), .Net is Windows-only.
>
>Not having used Java, my impression is it's no piece of cake either and the available tools are not as "developer-friendly" as typical MS offerings (from the few comparative reviews I've read).
>
>LAMP (Linux/Apache/MySQL/Perl [or PHP or Python]) seems increasingly popular but my impression is it's mostly for Web based apps, may not be too suitable for desktop apps. Technology is relatively new and tools not as advanced as MS or Java.
>
>I like that VFP has multiple 3rd party frameworks available. Their judicious use can really boost productivity. But, for those cases where technology matters, it's not mainstream.
>
>In general, it seems that the disconnect between functional descriptions (e.g. receive input, calculate/verify, store data, retrieve data) and their implementation, is greater than ever.
>
>This industry has had a large cadre of brilliant people working in it for a long time. Why does it seem that development is getting more complex rather than easier?

Al,

Just some general thoughts. None of the statements herein are directed at anyone in particular. They're just generalizations based on my observations.

I've always believed that programming is about solving problems, not languages or methodologies. I think part of the reason that things haven't gotten any easier is that we're being asked solve increasingly more sophisticated problems as the technology has advanced.

However, in solving these problems, there are numerous pitfalls along the way. Any one of these can make the task considerably more difficult than it should.

Failure to either understand or to not get accurate requirements of what needs to be done is a potentially disasterous pitfall. It can lead to solving the wrong problem or implementation that doesn't completely address the one at hand.

Using the wrong tool can create problems. There are times when I need to use C/C++ and times when I need VFP.

Not employing sound design principles can make things terribly difficult. Using bad methodology can likewise make things harder.

Doing or failing to do proper testing to assure the accurancy of the results.

There may be more. These are just the ones that come to mind off the top of my head.