Dilbert strip

Terminology interferes with the point.

Consider a PAYG telco system to debit a customer's account. Internally it is simple, but there's of lot of graubly cruft imposed by the quirks in an external system produced by another company, probably to meet a part-implementation of an ETSI standard.

Yes, the obvious partitioning ought to occur, but in practice it often doesn't. A standard term for this kind of thing is "leaky abstraction".

There are, of course, pure hardware equivalents.

And the "Uni" was using electron (ion?) machines to create one-off student ICs.

I was at CCL at the time :)

ould use them to probe a complicated integrated circuit and see where it wa s going wrong.

Alec Broers was writing tiny stuff with electron beams at Cambridge at the time - as he had been in various place for some years by then.

The Cambridge Instruments EBMF 10.5 electron-beam microfabricator that I wo rked on (along with lots of other people, going back about 10 years) was to o expensive for Cambridge University, but we did have one at IMEC in Belgiu m, and they were great at finding things that it wasn't doing quite right . ..

We sold another to AWA in Sydney (where it's still working, though the comp any that owns it now has a different name) and the guys that sold it were c huffed to find out that the fab in which is was housed (along with a lot of other expensive stuff) had been built by Lend-Lease where my younger broth er had negotiated the contract. It wasn't something that I'd known...

That is my standard mode of operation. Before one line is drawn on the schematic there is a module spec in an infant stage.

Same for the layouter later. But most of all, you cannot hold a reasonable design review without a module spec.

Name doesn't ring a bell. ISTR a professor with a middle-Eastern name, but that might be a faulty neuron firing.

Universities are good, useful "hostile environments" :)

e:

nd

could use them to probe a complicated integrated circuit and see where it was going wrong.

the time - as he had been in various place for some years by then.

he spent a while as vice-chancellor of Cambridge University (1996-2003). I' d run into him before that.

Haroon Ahmed. His graduate students were very proud of what they did, and w ere known to have told me that their three transistor amplifier was just wh at I needed for the EBMF 10.5.

I worked on (along with lots of other people, going back about 10 years) wa s too expensive for Cambridge University, but we did have one at IMEC in Be lgium, and they were great at finding things that it wasn't doing quite rig ht ...

Absolutely. We valued them for that. They were also full of graduate studen ts who would dismantle our gear, put it back together wrong, and complain t hat it didn't work - the one example that I ran into was working at DEC in Scotland, and not only did he get me and a service engineer to travel up to Scotland undo the consequences of his curiosity, but he also bitched about the machine at academic get-togethers after that. It was very much a first generation electron beam tester, with lots of stuff to bitch about, but we had done quite a lot of work to undo his curiosity-driven fiddling, and in a way that should have allowed him to avoid the mistakes he made first tim e around.

Yup, that's the name.