|I was using a DIY programmer and the strangest thing happened. I've seen
other people mention similar |problems with other programmers as well.

|The unit would work fine with a couple of 16C84/16F84 PICs I purchased a
few months ago, but would |not work with newer chips I purchased recently
(same speed and type). It would return a failure error.

|I tried a couple of programmers which were posted on the web, they worked
fine with ALL PICs. |Perhaps there has been some type of change to the
device... << File: ATT00002.htm >>

I have on occasion had this sort of thing happen, though never with '84's.
Generally, it was caused by either:

[1] Not realizing that someone else using the machine had run Windows and
    shelled out to DOS so I was no longer at a "real" DOS prompt.

[2] A wire than had come loose on my early band-aid-and-chewing-gum prog-
    rammer.

[3] ** An apparent change in PICs such that they no longer tolerated VPP
    rising before VDD. **

Note that when in-circuit programming 12C5xx's, the latter problem can
pose some "interesting challenges" since the VPP can't rise above VDD
until VDD is itself reasonably high (near +5V) but VPP needs to rise
over 7 volts within a very short time of VDD's hitting about 2 volts.
This in turn necessitates having VDD rise quickly, which can be a bit
hard if there's much capacitance on the board.