Well, it's worth considering. Most of my clients have terminals that they use to "talk" to other hardware of one sort or another. A few have PC's, but the physical location of a desktop is often an issue. Sometimes the office personnel just don't like sharing. I have to make the system accessible via means already present in the shop, if that's possible. Keep in mind that many of my app's are machines used in manufacturing. These have, in the typical case, been in production use since the late '70's, and we're just keeping them in use until it can't be done any longer, by which time the owner will probably retire. That's why I'm focused on DIP-40 and PLCC-44 packaging for my "upgrades." I can drop in a DS89C4x0 and immediately improve performance and increase on-board code space.

In some cases, the main problem is that the external EPROMs are too slow, hence, have to be replaced by faster versions. Since, back in '79, the biggest readily available EPROM was a 2732, in a 24-pin package, and since those were 250 ns or slower, it's often an issue "haywiring" the board to accept a bigger part. If I need an ICE, I still have the old Intel MDS with the ICE-51, though I haven't used it in over a decade.

Since a terminal, normally not an ANSI type, but a TVI9xx, typically, that's what I use, and that's what I intend for the client to use when he's doing his own work on the system.


That's a reference to an earlier comment in this thread. I'd be interested in how this is done. I've often wondered how this is done in the 805x context, since it hasn't got a software-generated-interrupt instruction. Can you point me to a detailed description?


No doubt! Not all free tools will work in all applications. I'd like very much to put together a complete debugging tool set that my clients can use once I'm no longer available, but I doubt they'd be willing both to pay for such a tool and to learn to use it, considering few of them are much younger than I.


Well, we're in agreement on that. I contend that there should be no development-related "tracks" in the code to be delivered, not even one byte.



Fortunately, I've not yet been forced to deal with this headache. It would trouble my clients in these long-term manufacturing installations to see me show up with a notebook PC and an external ICE, since they're not easily motivated to buy something and then learn to use it. I'm not going to be around forever, nor are they, and they just want their businesses, which have been limping along for a decade or so, using the old technology as best they can, knowing their manufacturing will eventually move to Asia where shop labor costs <<$10 a day rather than >>$20 per hour. Justifying what we old-timers think of as a significant outlay and an associated (painful) learning curve, to someone who figures he'll be out of business in a year or five, is not so easy.

Keep in mind that, here in the U.S, we are operating in a declining economy, where, in 30 years, most of the capital will be held by a tiny minority, ~100 or so, of currently well-connected persons, and the rest of the populace, particularly those currently focused on fun, fun, fun, and material toys, etc, will be working for 4th-world wages because they need to eat. The remainder, unable to get those low-wage jobs, will be out in the streets with their knives and guns, hunting one another for food.

RE