UNIX co-creator Ken Thompson is a… what user now? Last weekend, the SCALE conference came back from the pandemic with a bang: Ken Thompson as keynote speaker. In the Q&A at the end of his talk, Thompson made a surprising confession. Bell Labs researcher Ken Thompson was one of the developers of MULTICS, the ancestor of and inspiration for UNIX. He also developed Space Travel …
Lot's of otherwise smart people like the walled garden and some very good tools. Good thing that many of us couldn't afford their offerings or they'd be much more popular.
I actually started coding in "B" (https://en.wikipedia.org/wiki/B_(programming_language)) after starting in assembler and FORTRAN on the GE architecture. It's been fun watching the evolution of these sets of languages and hardware capabilities for so many years.
Thank you, Ken.
It is expensive only if you buy it new. It is well made stuff so decent second hand kit is available. I’m typing this on a used MacBook Pro. I have recently bought a reconditioned iPhone from eBay for less than £200. It plays better with the laptop than my Android and when I finish transferring stuff from the Android I will log out of my Google account with a happy feeling. They have become Evil.
My “spare” MBP that lives in the sitting room is a 2009 model, recently upgraded to Monterey using OpenCore Legacy Patcher.
I’m not sure I’d use it for music production but it’s fine for casual office use and gentle coding.
But how its that for upgrading? I still have my old Lenovo from about 2012, W530. Still going fine. The A is popping off the keyboard but its easy to open the laptop to replace the keyboard. I've replaced the HDD with an SSD that didn't cost a fortune and I've put in loads of RAM. Switch to my partners Apple Macbook Pro that she doesn't use much now, and trying to get a new HDD for it and it costs a bomb. Unfortunately she's locked into everything Apple so won't break away.
I bought a second hand SATA SSD for my 2012 Macbook Pro. Works perfectly. Accepted though that this is harder on the news ones but it's been a while since I've seen a Windows laptop with an SSD that wasn't soldered on.
What Windows laptops are you looking at? There are a tone of them that don't have the SSD soldered on. And then we have Framework. If I could, I'd buy one, I really want one. A modular laptop were everything is replaceable. The really stupid thing is, if Apple did such a thing and at a decent price, they'd probably get people like me actually buy some of their kit. But we know they are never going to do that.
I've got one as it's useful for the fairly rare occaisons when I need to test something on MacOS - it dual boots and runs Windows most of the time.
Don't think I'll be getting another one, because the need to test on MacOS has mostly disappeared now, and the lack of a TPM means that they can't run W11 (or not without a level of workaround that I don't want to do)l.
[Author here]
> It is well made stuff so decent second hand kit is available.
100% this.
I am typing on a Retina iMac, my most powerful computer. I bought it from a friend -- one who still works at Red Hat; I make no editorial comment about that -- used, for $LOTS but also a lot less than quarter $NEW_PRICE.
It runs mostly FOSS apps, currently on 10.14. I have no payment method on my Apple account, because it is 28 years old and the company didn't accept online payments then. You couldn't have a payment method. Nowadays, iTunes throws errors over this, but I only use it for streaming BBC radio and listening to my own MP3s so I don't really care.
I also have G4 and Core i5 Mac minis, a Macbook Pro, 2 beige G3s, a Mac Plus, an iBook G4 and an SE/30. And a Retina iPad and iPhone 6S+.
Most were free. The mini and iPhone cost me, but used, so very little.
I don't personally use their laptops. I dislike the keyboards and trackpads. My fiancee uses the MBP.
Apple kit is good, but buying it new is a mug's game.
A few years back I acquired a Mac Mini Server from work, it had spent the previous 7 years acting as a file server for a suite of iMacs.
Last year I traded it in for a new M1 Mac Mini, the shop gave me £200 for the 10 year old Mac Mini Server, which I'm not about to grumble about, considering I didn't pay for it in the frst place.
When I checked back later in the day they'd priced it up at £280 and it sold in under 24 hours.
I think the Mac Mini is the perfect entry into Mac OS, not stupidly priced, and built to last at least a decade...
> It is well made stuff so decent second hand kit is available.
That used to be true. Maybe. Based on these experiences and experiences from other people Apple quality is at or slightly below the quality you would expect from a HP laptop.
I've owned/used several Macs. Used to administrate a large number of early OS X macs at a college. They were solid. Then I bought a Power G4 ibook, which failed due to bad solder joints on the GPU. Recently was required to use Mac for work and first one had keyboard issues. Newer one has been crashing because of the taskbar failing. Which has been solved because the trackbar is now failed completely. No more trackbar, no more crashing.
As of right now I have to say that my discount $254 Acer throw-around laptop has taken a huge amount of abuse and has carried on quite a bit more reliable then my powerbooks that have lived their entire life sitting protected on a corner of my desk.
The big advantage for things like Raspberry Pis is that except for the MicroSD they are pretty simple and distraction-free. If you don't need the capacity offered by a modern PC then they are perfectly fine.
Reminds me of the good ole days[tm] on alt.folklore.computers when people would argue about some fine point of C semantics, and DMR would weigh in with a flat statement about it.
Which of course led to the question: "What makes you think you know about it?"...
Something similar happened at a USENIX conference many years ago. Somebody from M$ gave a talk on their shiny new POSIX subsystem that even included a fully functional Korn shell, ksh. As I said it was a long time ago...
A guy went to the mike and asked questions about various defects in this ksh implementation. The M$ guy blustered and bullshitted for a while, eventually saying the questioner didn't really know the Korn shell. "Who are you anyway,?" he asked. "David Korn" was the reply.
Reminds me of the good ole days[tm] on alt.folklore.computers when people would argue about some fine point of C semantics, and DMR would weigh in with a flat statement about it.
Caused quite a careful check of the email headers when you'd suddenly get a long email from DMR discussing a posting you'd made about some aspect of how the kernel did things and he'd explain how and why they did stuff originally and he'd be really interested in how the version you were working on did it differently. Fascinating bloke to converse with and sadly missed. Staggering to think what they achieved with so little.
long ago at work I had to set up and install a system that would used by 100s of thousands of people. We also used it internally and so I was told I was not allowed access until I had completed the training. Even though I had more experience with it than everyone else in the company combined I was forced to go. I eventually did go and spent most of the time asking questions that the trainer couldn't answer
I eventually did go and spent most of the time asking questions that the trainer couldn't answer
I'll bet you're a lot of fun at parties.
Humiliating the trainer like this just to give your ego a boost was a pointless waste of time. It certainly wouldn't have been appreciated by the others who were on that class and presumably needed to know how to use the new IT system.
As a child I wrote Spectrum (Z80) assembler and was published in a magazine. When I went to Polytechnic the Assembler Programming (Z80) lecturer (M. L.) insisted that I attended his course. He was a PhD student and about 10 pages further in the book than the rest of the students. I was asked not to attend any further lectures after I asked too many questions in the first. Other students would pass on the coursework assignments and would hand in the answers. (In actual fact it was the answers for 6 students, each with their own style and idiosyncrasies e.g. common spelling mistakes and different commenting styles. It kept me well stocked at the student bar! The lecturer never said anything and they all passed.) What did annoy me was only getting 99% for the course - the bastard got his own back and docked me one mark for not attending his lectures :-(
A Good response! I saw the same things, learning languages myself as they appeared was never too bad after figuring out how to write 8080 code with a pencil originally and then teaching myself everything else. When Pascal appeared I started using it and then I ended up having to teach a college teacher how to use it. LOL, helping was not an unusual event and the two of us got everyone else up and running with Pascal that week.
When Apple computers first appeared the company I was working for bought them for all the salesmen and I had to get them up and running for them but have ever used one much myself. Originally computers would appear and "updates" were only optional so everything was easy. These days "updates" are happening all the time and making a mess of the user environment daily.
"Why do I drink coffee while reading The Register??? Another keyboard bytes the dust..."
Same here, but with tea :)
Seriously, which kind of bozo at Google decided he should go through a C test like all the others ?
Dude is probably famous in a very bad way, now ...
Whoever wrote that rule was probably long gone, and in their defence, they probably weren't expecting one of the authors of C to turn up.
The real idiot was whoever his boss was, who didn't just tick the "passed the C test, ok to commit code" box, manually.
If there's any trolling it's the "because I was sort of born into it" about running Apple. It's not surprising that he finds Debian/Raspbian an accessible replacement for Unix. The questions following the talk do reveal two things. One is that the trojan is no longer in the C compiler. The other is the answer to "Who ate all the Pis?". He's got so many v4s its not surprising the rest of us can't get any.
Does anyone have the real story on how Ken Thompson was born into Apple? Was he always using Apple hardware with some kind of Unix or plan 9? The question was about the OS rather than hardware too so can we assume he was always using an OS supplied by Apple? What did he use before Mac OS X? AUX?
As I've commented on here before I ordered an 8Gb Pi 4 on 20th December 2021. Current expected delivery is mid-October this year.
Yes, I know two years a long time to wait, but I comfort myself with the knowledge that people pay hundreds of thousands or even millions of pounds on new cars and wait an equal amount of time for delivery, but my Pi 4 will still be more reliable and last longer. Probably.
The trick to ordering an unavailable Pi is to take advantage of being able to transplant a "buy it now" button from any other Shopify page to an "out of stock" page. Your order goes through and then you're top of the list when they release new stock to that supplier. I may have done this twice last year, although I'd deny it if challenged in court ;-)
I'm retired, but not (quite) as old as ken. Maybe it isn't quite a joke - I use an iMac and iPad for most day-to-day stuff, and have 3 RPs: An original Pi; a 2GB Pi4; and a Pi Zero 2W which are for background use and fun. I refer to the newer ones as my "server farm".
If you don't mind the lack of ports, the pi Zero is "astonishing". I originally used it as a Pi-hole but now it's for casual development work - A rough calculation indicates that its VAX MIPS/VUP performance is nominally ~5,000 times that of a 11/750 that I used in the 1980s. A multiuser web/database test with Bullseye 64 bit Light and a SQLite database on a microSD card gave ~14 inserts+70 updates/second.
I still run Windows XP and 10 in Parallels on the iMac to run stuff that I wrote before I retired, but if I replace the Mac they won't be installed.
AT&T ultimately missed a trick there. The Design and Implementation of 4.x BSD by McKusick, Marshall et all became the standard reference for anyone studying Unix internals.
Even recently I've been known to recommend them - there are guides on the innards of Linux but it is a) more complex b) changes rapidly and c) is getting a bit grubby really - large parts of it are well overdue for a refactor if not redesign.
"The Design and Implementation of 4.x BSD by McKusick, Marshall et all"
Marshall McKusick is one person, not two. He became one of the top guys in BSD when Bill Joy left to start Sun Microsystems. That was at the tail end of 4.2BSD.
That's not to say Marshal wasn't an important part of the BSDs prior, but Bill Joy was the designer and implementor through 4.2,
Yeah. AT&T really did miss a trick there.
The book you mentioned by McKusick, Karels et al was published 3-4 years after Maurice Bach's "The Design of the UNIX Operating System". At that time he worked at Bell Labs and was a colleague of Dennis Ritchie, Ken Thompson, Brian Kernighan, etc.
Both books are outstanding. The BSD book(s) are better on kernel internals detail because they weren't encumbered by AT&T's lawyers and licensing stupidity, as Bach's was. On the other hand, Bach's book is better at explaining kernel fundamentals. IMO they complement each other.
The reasoning should be obvious.
The man created an operating system still in use after almost 60 years. Sure, it has evolvved, but it's still very relevant. So, when he expresses an opinion on an OS, I'll listen to him.
Thank you, Dr. Thompson, for all you have done for computing.
But just like Gates's OS, Apple didn't really want to be seen as just another *nix variant until they had milked the proprietary OS angle for a few decades.
Personally, I'd love to see a large jump in the evolution of these server/desktop OS's that takes the best and leaves the accumulated cruft behind.
It acquired its OpenGroup Unix certification in 2007, a mere 6 years after public launch, Mac OS X having little technology in common with Mac OS 9 and earlier.
(though a bit longer than 6 years if you were to include the NeXTStep origins, but I’d argue that would be misleading because a lot of work was done by Apple specifically to be capable of passing the certification. Even then, they never attempted to deny its BSD underpinnings either).
This post has been deleted by its author
Back in 2000, Thompson didn't really like Linux (too rough around the edges for his kind of work) ... but as of an interview in 2009, Thompson was using Linux. See page 479 of Siebel's Coders at Work.
According to something I read from Russ Cox, Thompson was indeed using a Mac at go ogle, but he used Plan 9 from user space on it.
Oh whoa I didn't know that! Apparently I've long been using more of Plan 9, and far more meaningfully, than when I just installed the reimplemented and ported bits of it from Linux repos out of idoe curiosity. And I had no idea! Crazy to think how some people's influences on the world can seep in so widely and deeply.
"that he designed the C language."
Not everyone might be as proud of that :)
Personally I love the language especially once the C89 was standardized.
I think of C as a very sharp tool like an adze - you can, like Odysseus, build a boat or more easily take your foot off.
Wonder why he doesn't run Plan9 on his Pi?
Remarkable piece of hardware. I was given a 2b (return for a favour) which I mounted in a small plastic food container with holes cut for the hdmi, mouse and keyboard cables. Out of curiosity I plugged the usb cable for the Pi's power supply into the usb port of the old TV I intended to use for the display - I was suprised that there was enough grunt for the Pi to boot which then extremely quickly reached multi-user. Makes a nice self contained Linux PC - had to use a usb wifi dongle if I couldn't use ethernet.
Anyone who actually read Thompson's Turing Award paper would know that he did not include a Trojan in the C compiler. He merely discusses how he could have done it. That's the entire point of the paper: how can you trust the software you use? Still well worth reading today.
Well, not quite. From https://www.tuhs.org/pipermail/tuhs/2021-September/024485.html "Ken did it locally, but a vigilant person at PWB noticed there was an experimental compiler on the research machine and grabbed it. While they weren't looking for hidden stuff, they probably were trying to find what was new in the compiler." More interesting history in the same thread.
I can't believe that no other commentard has mentioned their experiences of MULTICS (although I fully expect to be Ninja'd in that respect)
Back in the heady days of the early '80's my university (Brunel) had a system that us undergraduates could use as we saw fit. There were rows of teletypewriters with no screen but spewed fanfold lined paper, and a handful of early CRT terminals (VT52?) and even a card reader for the postgrads.
It was completely oversubscribed but we discovered that free terminals could be found in the wee small hours after the uni bar closed. This was excellent training in becoming a functional alcoholic, but thankfully none of the code I wrote back then is available for peer review ...
That said, we had very little actual work to do, so I bought the small MULTICS handbook and amused myself by fiddling around on the command line. All good background for when I took delivery of my own Sparcstation several years later.
Being an engineer, I rarely throw things away, but I did bin that handbook during a house move. I've regretted it ever since, not just for the nostalgia, but it would have helped me recreate my university computer centre experience with nothing more than a Raspberry Pi.
Time for beer.
Shocking that Ken Thompson is using a form of Linux?
I mean, it might have been shocking that instead of something UNIX-related, he was using a Macintosh.
I was expecting the "shocking" revelation that he used the operating system that actually lets people get work done, for which a large number of applications are available. Microsoft Windows. I could see people being shocked about that.
Millions? Try billions, over several generations, with many more to come.
Considering how important unixoid systems are in the worlds infrastructure, and how influential C is, was and will continue to be, not to mention UTF8, regex, Golang and god knows how many other contributions to computer science, this man and Dennis Ritchie have easily influenced the lives of billions of people, and for the better.
The fundamentals of OS X were built on solid ground. No problem with them. But the direction of the unfixable hardware and Apple Store are understandable reasons why one might want to move away.
Honestly Raspbian, hardware performance aside, is plenty good enough for a lot of routine uses. So, my question is, why not?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
