Funny...
it never goes quite that well for Electric Usagi!
But really, what did he do? Took a processor card, couple of memory cards, a power supply, and a backplane, and faked a disc interface? Should I be surprised that it worked?
Former Microsoft engineer Dave Plummer took a trip down memory lane this week by building a functioning PDP-11 minicomputer from parts found in a tub of hardware. It's a fun watch, especially for anyone charged with maintaining these devices during their heyday. Unfortunately, Plummer did not place his creation in a period- …
This post has been deleted by its author
For someone who ran a spamming outfit masquerading as a third party utilities company (SoftwareOnline.com), sure.
Should I be surprised that it worked?
In general, yes - unless you're absolutely sure of the provenance of the bits and pieces there's a good chance that either some critical bit is missing or that something is simply broken. Having said that, he chose to start from a fairly promising point. Unlike the earliest PDP-11s, whose CPUs were made from multiple cards with their own backplane, the KDF11-B in this 11/23+ is a single QBUS card containing DEC's second microprocessor implementation of the CPU and so has a lot less to possibly go wrong. And having two memory cards increases the chances that at least one will work.
The QBone is probably cheating a bit, but I/O devices are typically large/rare/mechanical so acquiring them working can be a significant challenge and expense.
Yabbut... I could perhaps have been more explanatory in my original post.
In Dave's video, there is no mention, not even in passing, of any investigatory work that he might have done prior to building the thing. We don't know how many capacitors he had to replace, how many missing chips had to be sourced - but the implication was that he found everything complete and ready to go. Which kind of implies further that they were either original spare parts, or were removed from a working system. My expectation would have been that electrolytic capacitors from the era left unpowered would be potential short circuits by now, and that the original power supply might be a bit risky (but didn't he use a modern supply?)
I suggest that at the time it was built, it was _not_ built down to a price; it was built by engineers who knew their jobs; could read data sheets; and expected that their designs might have to work for years if not decades. It seems unlikely that they would have chosen any parts (capacitors are an obvious example) other than the best they could; cost was not really an object.
What I was saying was that it is a compliment to the original designers that their boards worked after this time - we don't know how long - rather than a surprise that Dave could stick it together.
Not just the electrolytics. Neglected PDP power supplies of that era were notorious for having RIFA capacitors as mains filters, which crack & let in damp over time. The first time they are powered on they explode.
I've just fully-recapped such a PSU, electrolytics included, after it sat unused in a cupboard for 25 years. So far, so good...
Not just the electrolytics. Neglected PDP power supplies of that era were notorious for having RIFA capacitors as mains filters
1) Yes, no reason to expect electrolytics in a PDP power supply.
2) But it's not a PDP power supply anyway, it appears to be a generic 5V-12V power supply of the shelf.
1) Yes, no reason to expect electrolytics in a PDP power supply.
Not sure if you're being serious about that, I counted 28 electrolytics to replace from 1μF to 2200μF, including a couple of physically big ones (2x 220μF @ 450volt and 2x 1000μF @ 250volt). The exploding RIFA ones are just standard mains filtering film caps, 0.22μF and 0.47μF, but must be special safety (X-rating) ones.
2) But it's not a PDP power supply anyway, it appears to be a generic 5V-12V power supply of the shelf.
For Plummer's rebuild, yes. It doesn't have all the supervisory and control circuits, nor the fan output. The original PDP supply was also bought-in, from Astec. It weights a couple of kg, and working ones now go for £500 - £1000 depending on condition, so easy to see why he used a generic one for his test.
This post has been deleted by its author
> The QBone is probably cheating a bit, but I/O devices are typically large/rare/mechanical so acquiring them working can be a significant challenge and expense.
Doesn't TNMoC use a Raspberry PI to emulate a room full of washing machine-sized discs on their ICL 2966?
Still impressive that the original OCP and IOC cabinets still work :)
I have to throw in my cynicism too. Dave is accused of exaggerating his achievements at Microsoft, and has kept Mum about the accusations about running the scummy adjacent softwareonline dot com. His bio’s read a lot like the boastful BS that Tommy Tallarico wrote about himself and successfully made the “truth” until a minority corrected the record.
He also left Microsoft specifically to run his scammy company. He really is a piece of work who made a lot of money off of poor, innocent people. I'm fairly critical of The Register because of its failings in this regard. I'm also very critical of The Reg's community that openly supports that behavior of both The Register and Plumber. Nobody should platform these kinds of people, but here we are.
It is actually kind of amusing, since I watched the video about 20 minutes before seeing this piece.
I did think the QBone was an excellent dodge... who the hell thought to make a board to adapt a BeagleBone into a PDP-11 bus?? That has got to win a prize for most esoteric niche product ever.
Our fave scavenging suave Dave saves from tech grave fave computer that someone had clave, restores at man cave so that it boots and behaves, has result that many oldies crave. The keyboard slave at our fave scavenging tech Aves riding emotional wave raves in brave article they gave, hopes forum knaves will behave, forced to resort to poetic stave.
I'll now leave, and have a shave.
Who called him an idiot?
It's nothing against Plummer specifically, I just have a deep disdain for YouTube and YouTubers (not to mention TikTokers) and the mini personality-cults that seem to follow them around. Read your post again, doesn't it sound just a little sycophantic?
El Reg has dozens of articles promoting Dave's YouTube channel for some reason. I would have just as much cynicism if El Reg plugged any other YouTuber as much as this.
There's something about the standard "YouTube Video Thumbnail" with a gormless face and some colourful text and arrows to draw your eye and make you feel "astonished", that just makes me want to put the keyboard through the monitor.
"There's something about the standard "YouTube Video Thumbnail" with a gormless face and some colourful text and arrows to draw your eye and make you feel "astonished", that just makes me want to put the keyboard through the monitor."
If you follow any of the more sensible and technical YouTubers, they commonly bemoan the fact the HAVE to do that, and use stupid superlatives in their video titles to get YouTube to actually list them well up the list and get seen. It's all SEO tricks they have to keep adapting to. And as with all arms races, it only ever gets worse. It's the "Best ever wash", "Whiter than white" soap powder wars repeated ad nauseum. I use TarTube[*] to track my favourites and it's not unusual to see a video pop up again with a new, more descriptive title a week or so after release once they no longer need to game the algorithm for a specific video.
* only as a "what's new" thing, I still watch on YouTube, don't download them. I don't subscribe to channels or even have a YouTube/google account, but at l;east watching them on YouTube (which does mean some level of Google tracking, makes it properly into the site metrics which downloading may not. Not really sure how that works. And I do support a small number of my most favourite channels on Patreon too, which also comes with the other benefits, like not having to watch on Youtube at all.
I'm similar to you for Toobs of Ewe stuff ... But I VPN it and kill all cookies after each session. I've got a few channels saved in a couple bookmark folders. In a so-called "private browsing" session I right-click a folder, open all, eyeball each for new content. I only bother with proper videos, I have no time for "shorts"-attention span theater. NoScript and uBlock do their respective jobs as required. I don't do patreon, but I'll buy merch occasionally (ballcaps, mostly ... I go through dozens every year around here).
"I have no time for "shorts"-attention span theater."
Make sure all your bookmarks end with /videos/ and you'll never again be bothered by shorts, live streams or otherwise (unless the channel owner puts that stuff in the main "Videos" section. Same applies to getting channel updates using TarTube :-)
I happily devour a few bits of "tech youtuber" content - mostly the chill, take-this-odd-thing-apart sort of videos with no egos present.
Sounds like you're watching the wrong channels imo. Of course, if you want brash and somewhat over the top there's always Linus Tech Tips.
Just lurking, because it's interesting, but... I find it all a bit ironic, as I think I'm one of the least "YouTube" YouTubers out there. Have you seen my thumbnails? They're hardly LTTesque. I do videos on PDP-11s and assembly language, and there are still gatekeepers who find it too mainstream?
At the end of the day, you're getting my best masking performance. If I could do better, I would... and will. And since I can barely stand to watch anything from more than about a year ago, I'd say it's evolving over time.
The implication that I've exaggerated my Microsoft accomplishments is a bit weird. Yes, I made a video about start menu code that it turns out didn't ship at the last minute, but I wrote the code, I've confirmed that with the team lead, and as soon as I was made aware, I (a) wrote about it on Twitter, (b) updated the video description, and (c) wrote to the Register directly to let them know. It's not clear to me how it should have been handled differently. What should I have done in light of this honest mistake where I just didn't have all the info when I hit Publish?
Even in that Start Menu video I'm careful to give extensive credit to the Win95 team as the original authors and designers within the first minute. I'm guessing that it still irked SN.
If there's something I've taken credit for that rubs people the wrong way, let me know specifically, and I'll answer to it. But I don't like the "innuendo" thrown around like that!
And other than things that were "just me" initially, like Task Manager, Format, ZipFolders and the pinball port (which I'm always careful to disclaim as a port) I don't think I've taken "credit" for anything, especially since I was just part of a larger team, and am pretty clear about that.
To quote myself from one of my videos, "Success has many fathers, and I'm just the fun uncle...". People often comment that it seems I worked on everything, and part of that is I was one of the devs that ported the entire shell, so I touched a lot of it. And I had to fill in a lot of missing blanks like Format, and they were all in pretty visible places. So I had the privilege of working on a lot of stuff that people have been before, which is handy for a YouTube channel to make a connection. But I sure don't claim it's any better or worse than someone who spent a decade working on the unseen heap manager. Just easier to write about!
As a trivial example, did I ever write a shipping kernel API? Sure. Was I a kernel developer? Hell no... And I worked on the *prototype* of Windows Media Center but not the shipping one. I added CD-ROM caching to MS-DOS but it already had disk caching. And so on. I'd like to think I'm fairly careful about that stuff!
This is TehIntarWebTubes ... it doesn't matter what you write (video, whatever), nor where you place it. SOMEBODY will bitch about it ... usually because the are jealous that they have never been in a similar situation, and likely never will be.
Thus it ever was. Ignore it. Stick around and comment where you see fit ... we can use more perspective from people like you around here.
And relax, have a homebrew.
Dave Plumber ran a scam company that tricked people into downloading fake antivirus and security software using online ads, and then the software installed adware and nagware onto users machines. Aside from that, his former coworkers consistently report that he way overemphasized his impact and success at Microsoft.
> usually because the are jealous that they have never been in a similar situation
True, I haven't made millions off of innocent people. I may not have the chops to get noticed around here then. /s
> we can use more perspective from people like you around here.
I think we can do without, thanks.
Where I started working had a PDP 11/34 and a Data General Nova complete with huge 19" rack-mounted 5MB HDD, with removable platters. So I remember them well, even though I never actually used them.
My own IBM PX/AT card & software replaced the Nova's job, and some Sun workstations replaced the PDP 11/34, about 2 years after I started, but I remember them fondly.
….make life so much easier than the cramped, heavily shielded enclosures PDP’s had in their first lives. The thought of inserting and especially removing cards without donating copious amounts of blood is almost incomprehensible to this veteran.
Come across “one?” Try PDP-11/20 (operating an instrument at a mountaintop observatory), multiple -11/34s, commanding, receiving data from, and analyzing those data from instruments on a spacecraft, an -11/40 used by another project, that borrowed me for some grunt work, and a schizophrenic -11/70 that ran IAS during the day and Unix at night. Probably others this geezer can’t recall right now.
Unless you have modified the RSX-11S Sysgen process you have not lived on the edge.
I was working on a project destined for the MOD in the 1976/78 timeframe where we used RSX-11M and RSX-11S systems (civillian kit but ruggedised)
We had to add some code to the boot instructions to init some special devices. That code was put into the Kernel RAM area that was used for buffers etc after the boot was done.
RSX-11S was a RAM only version of RSX-11M. It was a lot of fun working on that bit of the project.
Those were the days when every byte counted. None of the 100Mb apps. Disk were RK05's or RL01's (2.5Mb or 5Mb)
Geez, I'm getting old.
This is bizarrely coincidental- or, arguably, not. When I saw this article it triggered my memory of a previous Register story and a quote contained within it. I couldn't remember whether it was the PDP-11, PDP-8 or something else that quote referred to, and I felt compelled to look it up.
Anyway, the story was about the fact that despite C being widely-considered as close to the metal and- in effect- a "portable assembly language", this wasn't really the case nowadays, since the underlying computational model of modern processors no longer reflected the 16-bit minicomputers C was designed around.
And the quote *was* indeed...
"Your [ present-day ] computer is not a fast PDP-11".
So, yes, C was designed with the PDP-11 architecture in mind thus, in a vague sense, it could be argued that you're still programming a simulated PDP-11.
"C was designed with the PDP-11 architecture in mind"
Not true. C was first implemented on a PDP-7, with the compiler being written in PDP-7 assembly language. It was a bastardized amalgam of BCPL and ALGOL, and had absolutely nothing to do with the PDP-11. That came later (along with a hefty pinch of PL/I just for spice). If you look, you can still see the shadows of these ancient underpinnings of the language in today's C dialects.
I was going by the quote and ready to stand corrected. However...
Having checked the Wikipedia article, it does appear that C itself- as opposed to its predecessors which (as you note) it was heavily based on- *was* designed with the PDP-11 in mind whether or not it first appeared on that machine specifically:-
: Thompson modified the syntax [ of BCPL ] to be less wordy and similar to a simplified ALGOL known as SMALGOL. Thompson called the result B. [..] However, few utilities were ultimately written in B because it was too slow and could not take advantage of PDP-11 features such as byte addressability.
: In 1971, Ritchie started to improve B, to utilise the features of the more-powerful PDP-11. [..] He called this New B (NB). [..] A new compiler was written, and the language was renamed C.
That said, I think it's safe to say that the original quote would have applied whether or not it referred to the PDP-7 or the PDP-11; both, it's safe to assume, are some distance from a modern PC.
Actually, I'm not sure how important- or even meaningful- it is to be "close to the metal" on a modern x86 PC anyway, since all modern "x86" CPUs are essentially compatibility wrappers around a non-x86 core and break down and convert those instructions into a completely different architecture internally (with that internal format not intended to be used directly, not should it be, even if it could).
I first used an 11/45 for an Operating Systems course. We each had a disk pack to play with. I have no recollection of what we were required to accomplish. :-)
Nice, simple, instruction set. Wrote the core of a Forth interpreter(?) in PDP-11 assembler.
A PDP-11/60 was my first non-mainframe code target for a compiler. I treated it as a stack machine, and proceeded to make it fail on my code. Turns out its microcode didn't like some of the instructions I generated. Then got permission to use the 11/45 and it worked!
In my case, SWMBO calls their designated space the machineroom/museum/mausoleum/morgue depending on mood and context.
The Heath H-11 and attendant kit was in a corner of the living room for a few years. Now she's up here in the office, keeping an eye on me. Yes, she still works, but is rarely fired up anymore. Makes dragsters sound quiet in comparison ... That's the computer, not the Wife.
Yes, I know the H-11 is LSI-11 based, and not an actual PDP-11 ... but it's close enough for government work. Literally.
Probably wanted a system that no version of Windows could ever run on. ;)
The PDP-11 had a nice small instruction set which was ideal for students learning assembly and OS principles. The actual DEC hardware was pretty Soviet by comparison with today's kit as was every other vendor's.
I recall working out, on paper, how to build a PDP-11 processor from TTL components from the Fairchild catalogue first by hardware instruction decoding then considering using an Intel 4004 to interpret the 11's instructions and produce microcode to control the TTL and then pretty obviously an u-instruction pipeline / cache could be introduced or letting the compiler produce the u-code directly. :)
A few years later the Motorola 6809 which I could afford was not too dissimilar to an 11 (and with a separate D/I space supported 64k+64k) from which I learnt a great deal. The M68000 was even closer but not then affordable. :(
"The PDP-11 had a nice small instruction set which was ideal for students learning assembly and OS principles."
IMO, the DEC kit from that era is still unmatched as a teaching environment ... IF you are trying to teach actual computer and networking concepts, and not how to fondle a slab or get frustrated by Redmond or Cupertino's marketing driven options.
[Reg FOSS desk here]
> the difference between a microcomputer, a minicomputer and a mainframe
Mainframe: filled rooms, cost so much only big companies could afford them. Originally, non-interactive, batch-based. No "users" but handled stuff for thousands of staff.
Minicomputer: small enough for a department to afford, desk-sized (or desk-side), CPU built from boards or many chips. Interactive, shared by dozens of users. (Hundreds at my uni, but the machines -- a cluster of 2 DEC VAX 11/780s -- struggled under the load: 10min for a DIR command.)
Microcomputer: single-chip CPU (i.e. a microprocessor), affordable by and used by just one user.
> the difference between a microcomputer, a minicomputer and a mainframe
You can race some minicomputers down the hallway - microcomputers and mainframes not so much
The MicroVAX coffee table edition was the fastest - good turning circle but difficult to steer except in straight lines - especially when sitting on it ....
I'm surprised nobody's brought up the obvious ... I/O
Mainframe: All of the I/O ... Need more? Easily added. Virtually unlimited. Spendy.
Mini: Some I/O. Can usually add quite a bit more, but silly hardware tricks possibly required. Not so spendy.
Micro: Just enough I/O for one user to get by on. On a good day, if you squint. Cheap as (UART) chips.
> the difference between a microcomputer, a minicomputer and a mainframe.
1950s IBM (and all 1960s movie/TV) was mainframe.
Digital Equipment Corp WAS minicomputers. Others dabbled in minis but DEC did it all, and the accessories and software and storage and networking. Many others pioneered or me-tooed. My dad was sent by RCA to Pittsburgh to study steel-mill automation. Low-end DECs were more controllers than computers but there's much value in that. Sadly RCA was already over-extended with management cruft and lost talent.
While there were some spiffy 8080 and 680x microcomputers, Apple, IBM, and Compaq defined the market.
Microcomputer - five digit price tag, or occasionally four digit if you wanted to go cheap.
Seriously? I'd argue that you could slice a digit off both cases and it'd be more accurate.
Were microcomputers *ever* five digits by the 80s, other possibly than high-end workstations?
The 1977 introductory price of the Apple II was apparently $2638- that's $13,700 in *today's* money- for the fully-populated 48K model. But that was a huge amount of memory back then (the 4K base model was less than half that). That's still only five digits if you take inflation into account, and I'd assume the price of RAM- and that version- had fallen drastically by the start of the 80s.
By the early 80s, "cheap" microcomputers were comfortably in three- not four- digit territory, with the lowest-end models even breaking the $100/£100 mark (e.g. the ZX80 was under £100, the ZX81 was £70 fully-assembled (the kit was even cheaper)).
Arguably, if we're talking about capability, 1959's IBM 1401 was the first widely available micro computer. The lease on a minimal configuration would set you back about $30,000/yr, in 1959 dollars. By 1961, a typical configuration would set you back about $87,000/yr ... or you could purchase one outright for about $370,000.
Not sure what you mean by "microcomputer" there? As far as I'm aware, it always meant one based around a microprocessor, and those weren't even invented until the the early 70s (or late 60s if you include a not publicly-known design for the US military).
Wikipedia confirms that the IBM 1401 was a mainframe.
(And OP did specifically mention *80s* prices)
Yes, seriously. Entry level for a business micro in the mid-'80s was £2-3,000, anything really well spec'ed would very easily break into five figures by the time you'd added memory, storage, a printer and several thousand pounds of software. We were selling stand-alone DTP packages for £15,000 or so, back in the happy days of 40% margins on hardware.
Three digits for a micro meant a self-build CP/M machine or home computer that no serious business would even acknowledge existed, let alone contemplate using (with the exception of the Apple II, which had a surprisingly long life in labs and other coal-face bits of big business).
That only really started to change with the advent of the Amstrad PC1512 in the second half of the decade, and that would never be considered by any significantly-sized business. Very handy for small businesses running on pirated copies of Lotus 123, though.
GJC
The first PDP-11 I ever saw was in a county auditor's office, sitting in its own private 10x12 throne room. It was running the entire county - finance, accounting, payroll, annd real property taxation. As state auditors, we were in awe. And to make it even more fun, we got to start a couple programs on it. We'd ask for information to do the audit and it was handed to us right now because of the PDP-11.
About 6 years later when I was a county auditor, we moved from a Burroughs B6000 to a B60 to a MicroVax and then a MicroVAX III.