I agree. I've never seen any of Lynn's products offerred anywhere, AFAIK.
While I agree with the last point, i.e. that there may be less-costly ways to do what any specific application of the 8255 does, I don't agree with the relevance of the notion that it "was not designed or organized to work well with the 8051/2 architecture." In fact, it was not equipped with internal address latches, as were the 8155 and 8156 RIOT chips, or the 8185 SRAM, which WERE designed to work with that architecture, though they were originally intended for the 8085, the external bus of which works in more or less in the same way as the 805x's, but aside from that, it works as well as any other general-purpose peripheral would work.

That's what's troubled me about the discussions surrounding the 8255 in this forum. Just because the first page of the datasheet doesn't say "Designed and built for use with the 8051" doesn't mean it won't work suitably. Naturally, as MCU speeds have increased, there are concerns about access timing. Naturally, since the 805x doesn't have a wait line, it's necessary to have the MCU operate at a rate compatible with the 8255's access time, and that varies considerably. However, the standard (steam-driven, as Erik puts it) 12 MHz 805x, which is the slowest version I've ever encountered, works just fine with the oldest, slowest 8255 version that was ever made. I haven't investigated which 805x versions, at which clock rates, are capable of generating a nRD or nWR pulse width at least 250 ns long, but I'd guess many, particularly those in use by students and experimentors, actually do.

Further, I haven't seen "All of the requests for this product " so I can't address that matter. What I've said is that if Lynn were to provide a 25 ns version capable of driving 24 mA Iol/Ioh at its outputs, it would be useable with most 805x's. It would undoubtedly outsell the Intersil or Toshiba versions, if it could compete with them in price. I'd consider it, since it would then provide the necessary performance to work as an external peripheral with most 805x's and provide the output current necessary to drive most external-world applications.

The principal benefit the 8255 offers is its flexibility. If you're designing a set of hardware for a specific application, yes, you can do that better and for less money with a CPLD or discrete (SSI/MSI) logic. However, if you're designing a "general purpose" board for use by students, hobbyists, experimentors, use in "one-of" applications, you don't know what the specific requirements are. If you have a versatile device such as the 8255 or 6821, you're offering their inherent adaptability on such a board. That's something that serves a wider purpose than just saying it could be done better.

I've yet to see a way to do what an 8255 can do for less cost, retaining the versatility that the 8255 offers. Keep in mind, the 8255 provides as much output drive as nearly any 805x's port bits. You can seldom do "useful work" with those outputs without external "help."

I've no idea what Lynn's packaging options are, but, as he's working in the replacement market, I'd assume he'll be using standard packaging for the device in question.

I've also no idea what business a student or hobbyist building one of something would want with a PCB. If he's using a PCB, particularly if he's using a PCB and asking about an 8255, I'd guess it's one he's bought, and it already has accomodation for the 8255 on it. PCB's are for mass production. They're ridiculously costly in units, and often require multiple respins before they're properly configured. If he's got that sort of money, he should hire a consultant.

This question should be addressed in the thread to which you refer. However, it does warrant some self-examination.

Most of the self-made millionaires I know made their fortunes not by applying "cutting edge" technology to a project that interested or amused them, but by thinking about what they could do to improve the lot of others. They did that by finding 20th-century, typically pre-war, technology to problems that were being addressed with what was essentially 18th-century technology. Many such problems still exist. If one spends time thinking about what will benefit others rather than what will amuse himself, he's got a chance of doing something great. If all he can think about is his own personal pleasure, then perhaps what would be most appropriate would be a long walk on a short pier, in shark-infested waters. At least, then, he'd be of some use.

As for the self-examination, one should consider whether he concerns himself more with getting and having a particular job than with doing the job. It's common, but if his main concern at work is not doing, rather than keeping, the job, then he's just a drag on the race. Too many people are concerned only with their own amusement and pleasure, and not with the net benefit they create. If their primary function is mental, and/or physical masturbation, we can do without 'em.

RE