And half a decade earlier Tony Hoare (later of CSP fame) wrote a full Algol-60 compiler in 4K of 39bit words.
That taught me how to think in ASM (as opposed to the 39bit word machine!), and the 6800/6809 how to think neatly in ASM.
The Z80 taught me how to design a machine that looked good until you tried to use it - at which point you found you were using its 8080 subset.
I /never/ understood the attraction of the 8086, especially not the salesdroids protestations that the 80286 and 80386 were /simple/ compatible extensions.
Wouldn't it have been too competitive with their big iron?
But IBM did take a 68000, and modify its microcode so that a /pair/ of them emulated a low end 370. Bizarre.
That's what you do when you know who you want to hire and they need an H-1. By law you have to advertise the job for N days, so you take the guy's CV and write the whole thing into the job requirements. Whadda ya know, no Am erican meets the requirements, so you get to hire your foreigner.
Not I--I've written code for a DG Nova, but never for a VAX or PDP. I used to really like the HP 9816, though, which used a 68000 with a super-useful dialect of BASIC that made instrument control an absolute breeze. (My HP 35
665A is about equivalent to one of those, besides being a DSA.)
The Nova, like the HP2114, was really a 16-bit PDP8. Page/accumulator oriented, pretty klunky. Dec was going to do the same thing, but reason won and the PDP-11 happened instead.
Over-architected CPUs - like HP3000, i32, Itanic - have collapsed of their own weight, and their own slowness.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
The salary was something goofy like 80k. My girlfriend's ditzy 29 year old roommate with a bachelor's in communications/PR makes more than that (working for the state government, natch)
The department is called something like uh, "INS" if I remember correctly though so I imagine it's not the most personally rewarding work for most people
You can do a LOT with very little (e.g., the Reading Machine ran in
48KW of core -- *real* core). But, you can't hire the ITT mindset for those sorts of jobs.
[IIRC, the boot loader for the KRM was 16 words of code... maybe 32?]
The Z80 (and the x80's that followed) was actually a very capable processor. Things like the index registers were poorly understood by the "ASM mindset" crowd (hint: point IX/Y at a *struct* and their value is obvious). Once you appreciated the value of that addressing mode and learned to leverage it in your algorithms, you could do quite a lot in a far more structured manner than more "direct" addressing modes.
[And, the structure of the opcodes made it relatively easy to write embedded runtime debuggers and self-modifying code]
The alternate register set was also a win for fast ISRs (if used with care).
But, the 64KB address space (code+data) ultimately limited what you could realistically do with it. As memory was "less easily" shared (vs. the 68xx simple clock), you had to resort to more complex kludges (ultimately, the '180 parts -- which made compilers harder to adopt).
The 68K's threw away this simple elegance (bus arbitration being more complicated -- esp in light of the absence of fast programmable logic or affordable LSI memory controllers)
The 80x86's were the result of Intel failing to rethink their earlier design decisions and opting to "try for a sequel"... the same malady that plagues hollywood writers, etc.
Zilog, of course, was one of the more prominent examples of this failing mindset.
In too many endeavors, an unhealthy prejudice in favor of preserving existing products/efforts/designs has significant consequences for where future work can go...
And that's precisely where the Z80 drove me up the wall with my problems. I had linked lists of structs, where one of the struct elements pointed to the next struct.
In the 6800 you could chain down the list with one instruction per link, something like ld x,[x+0].
You cannot do that with the Z80's IX and IY registers, unless the pointer is in the HL register. And you can't transfer IX to/from HL, except via A. And you can't do the struct offsets in the HL, except manually, destroying HL in the process. Summary: everything took 5* as many instructions as "necessary".
Orthogonal instruction set? Don't make me laugh; just look at the very sparse opcode encoding map, and weep.
The Itanic collapsed because the hardies punted the difficult problems to the compiler writers - and those problems had been intractable for three decades. They presumed they could make a breakthrough; they couldn't.
PDP-11 had 8 registers, one of which was the PC and one was the stack pointer. You could do any arithmetic, or index, or autoinc/dec operation on any register.
NEG PC
was perfectly legal, but of course crashed. All sorts of fun stuff ensued.
Indexing off the PC let one write relocatable modules.
There was the "land mine" instruction,
MOV -(PC), -(PC)
which copied itself below itself, and then re-executed.
68K and 6800/02/03/09 looked a lot like a small PDP-11.
The original PDP-11/20 was all TTL, on two big boards, something like
520 chips, no microcode. It was a classic DEC async hairball, full of one-shots and RCs and tapped delay lines.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I'd never have doubted you/her/her roomie. (Sorry if anybody up there was triggered by my assuming that someone described as "ditzy" who lives with a presumably cis-gendered heterosexual female is, in fact, biologically fema le and self-identifies as such.) ;)
The H-1 guy can maybe negotiate a higher salary. I very much doubt that eve n our current administrative tyrants check up to see what his actual salary winds up being.
The ash-heap of the Bay is full of "social media" startups who were utterly convinced they were going to be The Next Big Thing and operated (and spent!) as if their aspirations could manufacture reality.
They didn't leave much but Aeron chairs and foosball tables. JL can probably tell you about it
Though it's difficult I try to remember that the most extreme on both sides of the party lines tend to scream the loudest. Twitter feminists "triggered" by "manspreading" ranting on Twitter, yes you can find 'em here, but they're not terribly popular outside of certain deep hive-mind populations; usually of white-as-white-can-be 20-somethings from families who spent too much money on 'em.
Even my GF, who did her masters degree in gender studies, admits that many of the most extreme are probably driven by diagnosable mental illness. Many of the women I knew in college in the 1990s who were into real-world activism drew away from the "movement" as it went up it's own rear end on the Internet and started to become more of a trade union for histrionics than anything of substance.
The number of flags burned in the 1990s to protest Clinton at liberal arts colleges was certainly impressive as well.
Gotta admit though that life frankly sucked hard for many women and gays during much of the 20th century at the very least. Alan Turing didn't commit suicide because everyone was so nice to him.
Most women in the state-funded social services and education in MA work long hours at tough jobs for low pay, same as anywhere else.
I would hope so, it looked like a monster of a CV.
In my day that was de rigueur, educational support AND family health insurance.
I prepaid our first child, Boston Lying In Hospital, the other three were courtesy of Motorola's health insurance. ...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| STV, Queen Creek, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I'm looking for work... see my website.
Thinking outside the box... producing elegant solutions.
Maybe the candidates don't like being patronised, or think about the consequences of working for somebody who needs the basics explained to him, and provide the kind of answer that ends the interview quickly.
My attitude to job interviews was that I need to get the candidates talking, preferably about technical stuff, and asking questions that belittled their grasp of electronics wasn't a winning strategy.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.