Most companies are very savvy with regards to the tools developers use, and any moderate business should have a good practices policy that outlines who/why/what/when etc - niche products like VFP are explicitly excluded from these lists.

These were the sorts of reasons that caused the company with which I am associated to move part of its work to NET starting in 2002. I'd regard that as the "generalized excellence" category I described earlier. It seems we agree about that. But I also mentioned providers of niche expertise or product that *itself* carries additional value for business apart from pure development. People in that situation report that the focus on tool may be less, meaning that the imperative to respond to vendor-declared obsolescence may be less. I've experienced both scenarios.

With regards to following the herd, we're not pushing the boundaries of science, we're developing reliable and robust applications, and for that reason that's exactly why you should run with the herd; to use VFP, purely out of personal preference, when .Net or C# will have a longer life and be far easier for a business to find support for, is, in my opinion, very unprofessional indeed.

That's another way of saying the same thing again, this time implying that we are all just developers, not innovators, and that people who disagree with you may be putting their personal preferences first and being unprofessional. Charming. ;-)

As I said, I don't see it as a big deal. People who need to run with the herd must have figured that out. People who have more room to move still need to remain educated so they can respond appropriately. There's also the frenetic pace of change that punishes you if you move too soon. Example: the fate of Linq to SQL. People who liked that approach and ran with it are dismayed to hear that it is being deprecated already. Those who hesitated are in a position to watch with interest to see how MS is going to roll L2S into the EF. If they do a good job, that may be an excellent long-term mechanism for smart data munging. Certainly Anders H has said that he expects data-centricity like VFP's in NET within 5 years, so if the choice is to move now and probably rewrite later or wait and see and rewrite when the dust settles, who can say from this distance which decision is best?

We as professionals should be serving the needs of our clients, not our own needs, and if that means recommending the use of .Net (or any other 'more popular' language), then that's exactly what we should do.

Of course; and as professionals, none of us would seek to impose a contrary straw-man position against which we can argue, would we? ;-)