HPIB was a very sound interface design that was clearly much ahead of its time. It endured generations of instruments and digital electronics technologies.
Designing interfaces for HPIB was a breeze, with clearly defined addressing, interrupt and talk/listen paradigm.
Of course, the addressing space is rather small, fixed addressing for nodes is limiting, and auto discover address resolution protocols are much more flexible, but HPIB networks still work very reliably.
A piece of sound electrical engineering from the 60's that still inspires my designs of embedded systems of today.
Yes, I have used HP-IL!
When I joined HP Scientific Instruments Division, several of our lab instruments had HP-IL.
It was a very clever 2 wire differential signaling that simplified lab equipment networking of disk drives, printers, chromatographs, analyzers, to build test setups and analytical systems.
I worked as a systems engineer for petrochemical applications, amongst other things, and I used the HP5890 chromatograph with the HP7673A sampler robot, with chromatography integrators.
It had no-brainer D-shaped connectors that made easy for anybody to setup the network, with automatic plug-and-play address discovery.
However, one sales manager once managed to connect the cables inverted, by connecting the D-Shaped upside-down. It was a nightmare to discover why the network was completely disrupted.
Since then I've learned not to underestimate the users capacity of misusing perfectly designed interfaces.
I never used it myself, but I thought HP-IL (HP Interface Loop) was pretty cute when I read about it. It's a token ring topology designed for medium speed communication between various controllers -- including battery-operated programmable calculators -- and test equipment. The links are transformer-coupled, so you don't need a common ground. The transformer turns ratio matches 5V logic with relatively weak current to low-voltage high-current differential signals for transmission on the links. Connectors and cables have gender so you can't hook it up wrong [*]. Nice piece of electrical engineering, I thought.
[*] after reading Jonny's comment below, I'd like toinsert "without a concerted effort" :-)
If some of you folks think GPIB is a dinosaur, I wonder if there's still anyone around who even remembers HP-MSIB (modular system interface bus)? This was introduced by HP out of Roseville just as GE was trying to define the CASS test system for the Navy, and HP had used MSIB for several of the instruments in the RF/microwave test rack. I can barely recall it now but I do recall there used to be a mechanism (I guess one of the instruments in the chassis served as a kind of protocol "bridge" so you could get to various MSIB components via GPIB) and I was brought in under contract on a team down in Huntsville AL that wrote drivers in 68000 assembly language to allow multiple components on the GPIB and the MSIB to be composed into "virtual instruments" which could be accessed in a custom variant of the ATLAS language. In many ways it was the ultimate "thankless task", GE was WAY behind schedule and wound up closing the Huntsville facility almost before the solder had cooled and transitioning the project to the "production phase" at Martin Marietta. It turned out that the entire PREMISE of "virtual instruments" actually DOOMED the scope of the project before it even got off the ground because what it actually did was to demonstrate that you could really extend the value of the capital you had poured into your efforts generating test programs in ATLAS by assembling instruments that met the test scenarios almost "on the fly" so the need for large complex racks with high capital spend was obsolete before the project was even finished, and unfortunately for GE the Navy realized it as well! Ah yes, the "good old days"?
Do Windows 7 and 8 have GPIB support? It's hard enough to get parallel port support for them.
I know that our company has moved toward using USB to GPIB adaptors rather than plug in cards so maybe those work. Or at least the person who did the setup made them work....
We have some older equipment in some test setups where we use labView test scripts to run a test using GPIB. We've got dedicated computers and data loggers for those setups so we don't have to go to the pain of reconfiguring. Most of these setups are thermal chambers which are still in almost constant use even with the new chambers we got in the last couple years...
Ha! I recently did a 10 minute presentation on GPIB for my degree. It was shocking to discover some students had never heard of GPIB before!! I did however end the presentation with the suggestion that VME and VXI could be taking GPIBs place in the industry though.
The problem, of course, is that no one ever upgrades either hardware or software. How many times have I had to add/replace a piece of equipment on a 10-15 year old rack that's being run by an equally old Sun Sparcstation (or similar)? Too many.
Update software? Not likely - requires a complete replacement of the machine and likely a major re-write of the code.
Replace all the stuff with newer stuff that gets along? Why would they do that, it's too expensive!
grrr... I've been spoiled by pug and play. I don't much use USB for the reasons you state, and on the rare instance I've had an ethernet cable go bad, it was a lot easier to replace.
Even an ether to serial converter box is prefereable to GPIB... To me, anyway.