Posts by Peter Gathercole

It's fun making Studio Ghibli-style images with ChatGPT – but intellectual property is no laughing matter

Re: There are very few genuinely un-intelligent people out there.

With the subject matter being animation, one is reminded of the passengers on the Axiom in Wall-E.

I generally prefer the earlier Ghibli films such as Nausicaa of the Valley of the Wind and Laputa Castle in the Sky, but for a quizzical view of life, Only Yesterday was very engaging (although I did not really enjoy the Daisy Ridley dub), and Whisper of the Heart are both very good.

I was wowed by the animation in The Wind Rises. It is so beautiful, and the earthquake and the following panic scenes were just unbelievable when considered that they were hand-drawn.

But for a story with a real impact, you should watch Grave of the Fireflies. It's not my favourite, but the subject matter and the way it is treated gives you a real jolt. Definitely not a children's film.

The most important experimental distro you've never heard of gets new project lead

It does not have to be just ~/bin. I normally have a ~/lib, if only to hold local terminfo entries that I use to make things (like the new Windows Terminal which is not fully xterm compatible) work.

Re: There's no one way to rule them all!

I edited the above, and actually introduced an inaccuracy.

What UNIX has had for so long is shared-text binaries, which is where only one copy of the code part of a program (the text segment) exists in memory for all running instances of that program. I don't know exactly when this first appeared, but it was in the first version I used, which was Edition 6 in the late '70s.

What I later added to the post was the reference to the page load on demand feature, which could only have existed in UNIX since on demand paging was added to UNIX, about the time of BSD 4.2, and carried into SVR3. This was still a long time ago, but may not (quite) pre-date DOS.

There's no one way to rule them all!

It's also the case with things like snap and flatpak, that each package will load it's own copy of potentially shareable libraries, because they each run in their own isolated environments, so even if two packages use exactly the same version of a particular library, they will each load their own copy.

In some respects, having page load on demand shared-text (it's a UNIX thing that's been around for a long time, longer that Linux, Windows or even DOS has existed) binaries is not that much of an overhead, especially in things that may have multiple copies running (like a shell). It may even actually have a performance benefit over packages with diverse and unique sets of many shared libraries, as a single binary will load what it needs in one operation, rather than chasing around the filesystems for lots and lots of shared libraries that haven't been loaded by other programs.

Probably the best compromise is a mix of shared very common libraries, and static linked unusual library, with the API to the common libraries strictly controlled so they don't change. But that is really pie-in-the-sky for a modern Linux distro.

But given the quantity and speed of persistent storage and memory, using historical rules to justify efficient program start is a bit of a moot point.

Users hated a new app – maybe so much they filed a fake support call

Re: It's the nuts

The time I was asked by a customer to go to an oil rig to sort out a software problem (which was almost certainly something they'd done, not a defect), I was told that even if I fixed the problem immediately, I'd have to remain on the rig for several days until the next shift change, as they would not put on a special helicopter to get me back any quicker.

I was not happy, and my manager was not happy (and my wife was LIVID), so my boss said I would go, but only if they would pay my external daily rate for the whole of the time I was off-shore, and quoted some outrageous daily rate, plus the inconvenience allowance and costs.

I don't know how much a helicopter would have cost to lay on for a special journey (but it was probably more than the costs a quoted by my manager), but it was decided that they would configure another server and take it out, and bring the one with problems back onshore so it could be fixed more easily.

M365 Family users wake up to notice 'Your subscription expired'

Re: No worries

If you keep to software in the repositories of the relevant Linux distribution, then you really don't have to worry about installing them. Just use the package manager, and provided you're pointed at the correct repo, there should be no dependency issues.

For software that is not in a repo. Try looking for a flatpak or an appimage, or if using Ubuntu, a Snap. These should be self contained, and although they may be heaver on resources to run, they should just work, again without dependency issues.

If you are actually using a tarball or installing from source, then you're probably not that worried about doing a bit of googling!

Re: And that's why...

Telemetry that they admitted multiple times they were using to see what their users were doing.

Musk's DOGE muzzled on X over tape storage baloney

Re: 70-year-old technology

Both of these were the height of technology when UFO was made. IIRC, the largest UK government data system in the late '60s was the DSS in Longbenton in Newcastle, and that was doing all of it's work with tape-to-tape batch runs. They didn't switch to disk processing until much later.

Although there were similar looking data tape drives, the ones shown on UFO were actually multi-track audio or video tape machines mounted vertically to make them look like computer tape drives.

But it was interesting to see that neither Moonbase nor SID had any tape drives visible anywhere.

It was also interesting to see that whenever a computer had anything to output, it was almost always printed on paper of some sort (ticker tape or cards, or sometimes continuious paper). This was some time before the general introduction of visual display units in real life, so the props. department wouldn't have known what to mimic. The best they came up with was radar-like CRTs, which were in use at the time with computers (Spacewar, anyone), but what is shown is obviously analogue.

Miss Eland's (Ed Straker's secretary at the film studio) keyboard also amuses me somewhat. It is such an impractical device for a typist that I wonder why they even tried.

Space 1999 was a little more realistic for the time, but not much. People could be seen toggling switches on wall mounted computers, rather than using keyboards, all over the place, something that was never going to happen in reality. There appeared to be one keyboard in the main mission, but that was mostly it.

Re: Minor correction

I think that you jest, but I'm sure it would be possible to put some data exfiltration mechanism into a drive, but the actual mechanism of sending the data back might be problematic. I guess you could build in a mobile data hot-spot to run through the (possibly non-existent) hidden data channels in the Huawei 5G mobile infrastructure, but since that was ripped out in the UK, you'll probably have to have mobile data hardware and a valid eSIMs in the devices, or put enough intelligence to allow the device to go out through any network on the computer the device is sitting in.

Re: Minor correction

I agree with what you say to a point.

The issue I have is that you're not really just talking about a number of hard drives. You're actually talking about hard drives and the hardware to attach them to a system. If you look at SAN storage devices, the drives are not the most expensive part of the systems.

I know that there were hardware devices that put hard drives in cassettes so that they could be attached and detached from shared controllers, and I seem to vaguely remember a hard drive library with a robot to move the drives around. But it is a product that faded into myth, and Commvault "PnP Disk Libraries" are not what I'm referring to. And I know that many systems have disks in removable carriers so that they can be removed from a system without powering the system down, but I've not seen anything to automate that type of device movement.

By comparison, both removable optical media and tape have a passive media with little to no active components. This should make them cheaper, allow the drives to read/write them to be better engineered, and also lend themselves to automated loaders or libraries.

Re: Only $1M?

Grrrr. DASD - Direct Access Storage Device. Must check before hitting Submit!

Re: Only $1M?

Country specifics.

I was actually deliberately resisting the urge to type disk rather than disc, because of being on the RHS of the pond (and vainly trying to maintain the nationality of The Register). IBM used to call it Direct Access Data Device (DASD) and many of the long-term IBMers that I came across never used either disk or disc, unless they were referring to floppies.

Disc would have been used in the UK, were it not that most computer companies originated in the US (at least those that survived for any time) so most documentation written in English would have come from the US unaltered.

As an example, documentation from Ferranti, English Electric or ICL (early UK based computer manufacturers) seems to have used "disc" most of the time.

Re: Minor correction

Those capacity values for LTO tape you quoted are for the raw capacity. All LTO tape drives have data compression that is 2-2.5x and does not slow down data speeds, so your 18TB on an LTO9 actually becomes more like 40TB in actual use (depending on the type of data being backed up). This is what we achieve in real use.

Re: Minor correction

I think that you have to quantify this.

If you want to buy 2400' 800bpi reel-to-reel tapes, you may struggle, and you would definitely not find new drives for these.

Like most things, tape has families and generations, and older tape/drives do become obsolete. Just off the top of my head, I think that the following tape types would be difficult to read now. 1/2" 2400', DecTape1/2, QIC (all densities), DLT (all densities), 8mm, 4mm DAT, IBM 3480, all StorageTek cartridges, and both types of Magstar tape.

Even the current poster boy of tape, LTO has compatibility which only go back two generations. Your fancy LTO9 tape drives, which are very much still available, can't read LTO6 or before tapes (and I can't remember the matrix, there was somewhat of a break in compatibility around LTO7/8 IIRC).

So the tape medium may well, and does, last decades, but you have to remember to maintain enough of the correct types of drives and systems to be able to read them. And this often also requires systems to attach them to. Even tradditionsl SCSI, which you could probably find drives for most of these tape types is getting really rare on modern day computing hardware.

But I would still trust tape over most writeable optical media that is not speciffically intended for archival purposes. You just have to be prepared to keep it working.

Re: Only $1M?

In the early days of magnetic media, the magnetic material was indeed iron oxide. It's stable and was very cheap. If you have any cassette tapes around, the tape is most likely to be brown for this reason (although dark grey is a colour associated with chrome ferochrome and metal tapes). It's difficult to tell the colour of LTO tapes, because you basically never see the tape any more as it's totally enclosed in the cartridge, and you only really see it if the end tab did not lock back in to the cartridge when it was ejected, or the tape or cartridge is broken. If you take apart a modern hard disc, the surface is most likely to be silver.

As data densities increased, more esoteric materials were used to decrease the size of the magnetic domains.

I have to admit that your distinction between disc and platter seems a bit arbitrary. They're both round and flat. The term 'platter' only really started being used when computer DASD/hard drives/discs started having more than 2 surfaces, requiring more than one actual disc. At that time, the media was often still called a disc or disc pack, but the individual circular pieces of metal/glass were called platters, so a 'disc' was often made up using multiple platters. And remember, when disc drives started being seen on computers, more often than not, the disc/disc pack was removable, and the disc drive was the separate device used to read the discs.

The industry moved to glass from aluminium for the platters when the track density increased enough that the thermal expansion of the metal became a problem, and when the spin speed increased to the point where aluminium would distort due to the high speeds and accelerations. You used to be able to tell when an aluminium pack contained a disc was bent or distorted by how it sounded when it was spinning.

Re: Nothing to beat it on $/Gb basis.

Sinclair Microdrive?

NASA doubles odds of Moon hitting near-Earth asteroid

Re: Space 1999 - The Reboot!

As I understand it, the US TV executives wanted a completely new series, and a new series that was significantly different from UFO. Gerry was really wanting to make a second series of UFO, but was told that the funding would not be forthcoming from ATV, as they needed to sell it to the US.

Gerry then proposed a moon based story line, but the answer came back that there should be no Earth based episodes, so the pitch was changed to have the moon blasted out of Earth's orbit.

Another requirement was that there had to be US actors in the main lead roles, which is why Space 1999 ended up with Martin Landau and Barbara Bain in the cast. They were not terrible in their roles, but I feel that they were a bit out of place in a mainly UK cast.

At the end of the first season, the US wanted another change, and wanted to use US writer Fred Freiberger to write episodes. Apparently Gerry was reluctant, having seen the general output of US TV series of the era, but said that if there had to be a US writer, they had to also be a producer, so as to take some of the blame if the second season tanked. This is the main reason why the second season feels so different from the first, with more men in monster costumes rather than the slightly more thoughtful story lines in the first season. Oh, and the per-episode budget was cut, as well as many of the first season characters disappearing with no explanation. Most of these changes were due to Freiberger.

I was really not hugely fond of the second season, and I am not surprised that there wasn't a third.

Re: Space 1999 - The Reboot!

I think that sequence was edited out of the version that was originally televised in the Saturday morning kids slot. It came as a bit of a surprise to me when I first saw the episode on DVD, and I had watched the series right from first broadcast, and I'm sure I would have remembered, even if I was just 10 at the time.

It seems that the dark nature of the story lines was less of an issue to the TV channels than a little bit of titillation.

Legal clock ticking for Microsoft over alleged software license abuses

Re: Looks like...

Which is all good, but if DRM gets in the way, it doesn't matter if you have a justified right to use the stuff or not.

Re: Looks like...

All of these "% of annual global turnover" fines will never be applied, for all they are enshrined in law. They are a toothless penalty.

20% will be a very large part of the annual operating profit of any of these companies that they apply to, and if they refuse to pay it, what happens next?

Well firstly, the US Government would get involved on the behalf of any US companies, saying that they are unfair. (especially with this current administration - more tariffs, anyone?)

Secondly, they would be appealed and appealed to the very end.

If all this did not get a resolution, how would a government force a company in a different legal jurisdiction to actually pony up? The operating arm of the company in that jurisdiction would have nowhere near the value of the fines in assets that could be seized. The only weapon they have is to bankrupt and ban their operation in the country concerned, and honestly, in this case, can you see Microsoft being refused a license to sell products in Europe or the UK, bearing in mind that pretty much all of business and the governments themselves would stop dead without Windows or Azure (and don't believe that there isn't a kill switch - I'm pretty certain that they will be able to disable the Windows entitlements for any internet connected system, especially Windows 11 or later)?

To avoid disaster-recovery disasters, learn from Reg readers' experiences

Re: once upon a time at the pub

This is why you use a storage management system which automates the copying of media. For example, a sufficiently designed tape management system will have at least two on-site copies at any time, and at least one off-site copy, which will either be a physical tape movement, or some remote copy.

A system such as IBM Storage Protect (aka TSM) can manage all of this for you if it is set up correctly.

Re: Same area of town

I worked at a UK bank that had a very good disaster recovery plan, and even tested important parts of it regularly to see whether they could bring up selected systems at the DR site (but did not actually run the service from there).

The only problem was that they didn't have a plan to repatriate any of the services in the case that it was a temporary (but long enough to invoke DR) outage! In fact, once the services were running at the DR site, there was no plan to reconstruct or resynchronise the primary, or any tertiary site for further protection.

It was a standing joke that if DR were to be invoked, almost the whole infrastructure team would move the services to DR, and then had their notices in, because they could not see how to return from DR.

Windows 11 adds auto-recovery, kills offline setup loophole

Ah. Smoothwall.

For it's time, Smoothwall was quite suitable for protecting a home (or even a small business) network without requiring too much effort. I originally put it on a redundant Pentium 100 system which had no CPU fan and a very quiet power supply, even before we had always-on broadband. Using the dial-on-demand feature with a flat fee internet provider (at the time Virgin.net) on a dedicated 2nd phone line enabled me to distribute the Internet firstly to all of the systems in my 'office' (which turned into a gaming room as soon as the kids got home from school), then through thinwire Ethernet to the rooms close to the office (I could have used twisted pair, but I had the cables, terminators, T-pieces and NICs scavenged from bins where I worked as they got thrown out), then twisted pair as computers and laptops got RJ45 Ethernet connectors and I got TP hubs/switches, and then via WiFi to the parts of the house that that would reach.

Was pleasantly surprised when we had ADSL installed, that I could just replace the modem that provided the Red network with a second NIC connected to the router and re-run the network setup to change the Red definitions, and the whole house had access to broadband (at least where the WiFi would reach) with minimal changes, way before ISPs provided wireless routers.

I sometimes regret that the community edition of Smoothwall fell into an unmaintained state, although I did contemplate trying to contribute to help keep it alive, but I didn't have the time.

The passive aggression of connecting USB to PS/2

Re: slap a keyboard ...

On the PS/2 Model 80s that I was using (running AIX/PS2) back in 1990, they were both IBM beige.

Re: *!$#&@ Bluetooth

Irn Bru or Tizer (never got to the bottom of which one it was, the kids wouldn't own up to who did it, but it was orange and sticky) definitely killed one of my model Ms. IBM did not pot a sealing bead on the membrane of their Model Ms (Unicomp do now), so liquid ingress was possible, It dissolved the tracks on the membrane, and more got pulled off when trying to peel it apart to repair the tracks. Whether throwing it in the dish washer before trying a repair would have helped, I will never know.

I've also had a Unicomp one fail as a result to Coke being spilled on it for similar reasons (don't know how it got past the sealing bead, but it did), but at least I could get a replacement membrane shipped to repair it.

I've also had the control board of a USB Unicomp one fail as well.

Model Ms are not indestructible. But they can be very long lived if treated right.

I often whether the non-contact Model F keyboards with the capacitive sensors would have lasted even longer, but I suspect that most of those were scrapped because they had the 5 pin DIN and PC/XT protocols that made them difficult to use on later PCs (plus they were, by today's standards, a very strange layout). They did feel amazing to type on, however.

Re: Weight?

I stand corrected.

I've just put my 1990 1391406 model M on the bathroom scales, and with the coiled cable it's exactly 2.15 kg (although the cable contributes quite a lot to the overall weight).

In my defence, the ones I've been dealing with more recently have been Unicomp model Ms, and the backplate as well as the plastics are much lighter.

The 3279 keyboards were something else, however.

Weight?

I'm not sure that even original IBM Model Ms were that heavy.

Of course the metal keyboards on an IBM 3279 mainframe terminal may have been.

Re: Polling vs Interrupted keyboards.

Also, if you want to use suspend with wake-up-on-keypress, you have to make sure that a USB keyboard is plugged in to a port that has power even when the system is in standby.

Re: *!$#&@ Bluetooth

For wireless keyboards and mice that have a USB dongle, the dongle looks to the system like a USB hub with a USB keyboard and mouse attached. The system is completely unaware of the devices being wireless. Although the wireless protocol in these devices often use Bluetooth hardware, most do not use the Bluetooth software protocols (and thus can't be used with the built-in Bluetooth support that many systems now have). Proper Bluetooth devices are, of course, different.

But I sometimes use old systems (often HP ex. Windows terminals that I redeploy as low-power, always on Linux servers) that do not by default enable USB in the BIOS during startup. When you come across one of these systems set to default, it is impossible to use a USB keyboard to enter setup to change the default to enable the USB during startup.

Of course, changing this default setting allows you to use USB keyboards, but it's very much a chicken-and-egg situation. And you have to remember to not set the BIOS back to factory defaults!

As a result, I always keep a keyboard with a PS/2 connector around somewhere. The main one (that I will never get rid of unless it fails) is my IBM Model M.

Also, cue the historical "Keyboard error, press F1 to continue" message that always amused me!

50 years ago the last Saturn rocket rolled out of NASA's Vehicle Assembly Building

Re: Shameless plug

I think with Apollo 13, the lack of bureaucracy was essential, bearing in mind that time was so very critical at every point after the initial incident. After all, the problems kept coming until they were back on Earth.

I'm sure that even in the coast back phase, where the astronauts were just trying to stay alive and not freeze, while not using too much oxygen, that the activity at the Cape and many other sites was controlled chaos, with people at all levels contributing any useful ideas that they had.

Microsoft patches patch that broke USB printing in Windows 11

Re: Printing!

This is exactly why I have at least one of my always on Linux systems in the home network. It accepts IPP and formats it for some of my older printers, although one by one they are dying.

But setting it up gave me some interesting problems. It seems that on more recent Ubuntu versions, it became the default to set a remote printer up so as to double-render a print (once in the client, and then again in the server), which at best printed gibberish, and at worst resulted in no print. I'm sure that CUPS used look at the headers on the print stream to decide at the server end whether a print was already rendered before passing it through the rendering engine again, but no more, it seems.

Printing!

Actually, printing is one of the areas where you really do want the OS to get involved.

You obviously don't remember the DOS days where you had to install the correct printer 'driver' (actually an escape sequence database) in EVERY SINGLE SOFTWARE APPLICATION that you wanted to use. And don't get me started on proportional font metrics and support!

An abstracted print model where the application uses a common print format that the OS knows how to manipulate to achieve the desired output by having a printer driver installed in the OS for your printer was a leap forward in device compatibility, and the rasterised WDDM (and CUPS on UNIX-like OSes) model finally made WYSIWYG an effective reality.

Although I can see the benefits of getting rid of the requirement of how printers print even rasterised pages by putting PDF interpreters into the printers, I don't really like like IPP and IPPS. It seems like a good idea, but sometimes the colour model used in the PDF generating code does not match the printer properly (I had CUPS print greyscale on a colour printer not that long ago for exactly that reason), and sometimes the exact device capabilities (such as double-sided or automatic booklet printing) that some printers offer cannot be enabled without some fine-tuning of whatever is generating the PDF. This seems like a backward move, putting some device support back in the applications.

And when they finally pull the old device models out of the print system (as the CUPS developers want to do), driving old non IPP/IPPS printers will become a lot more difficult!

Feds drop bomb on Multiplan in legal war over healthcare 'price-fixing' algorithms

Re: Multiplan?

IIRC, Multiplan was also available for MS-DOS.

When the Poly that I worked at in the '80s started teaching use of packages to it's Business students, the packages were Multiplan, Word and R:Base, when the rest of the world seemed to be using Lotus123, WordPerfect and dBASE II.

I think it was due to the educational discounts that Microsoft was prepared to offer (they had a license to ship R:Base in Europe, as they didn't at that time have any basic database program).

Tesla Cybertruck recall #8: Exterior trim peels itself off, again

Re: Tesla has issued its eighth Cybertruck recall

The difference is whether you're prepared to let the nearside of your car get scratched, if only minor scratches. People who actually live in these areas will generally drive older cars, where the minor damage from hedges and stone chips will just be seen as the hazards of driving.

Here in Somerset, the country lanes are not that wider than Cornwall, and I can tell a local driver from one just visiting by how they drive on minor roads. Visitors will try to keep equal distance from both sides, and will be a real pain in the neck if you ever meet one head on, especially if they're unwilling/unable to reverse to the passing place that is just behind them.

I drive a 17 year old Land Rover (although one of the more car like ones), but don't mind the hedges, because what is a few more scratches!

Museum digs up Digital Equipment Corporation's dusty digital equipment

Re: Progress

I'd bee surprised if you found many Sun systems running RFS (unless you were working at AT&T). It was an AT&T STREAMS based remote filesystem that came as an add-on for SVR3.2 (although I had it running on R&D Unix 5.2.6 on an Amdahl mainframe, but that was nearly SVR3 anyway).

At the same site, we had also started using SunOS 4.0.3, which was nominally SVR4, but even then we had to put the R&D extensions over the top in order to get it to play with the RFS, Although it had STREAMS, the RFS components still had to be added to SunOS, and I don't think Sun provided it, as they wanted everyone to use NFS.

Weeks with a BBC Micro? Good enough to fix a mainframe, apparently

Re: Loud printers

Pretty certain that IBM and other band/chain printers were aware of where all the characters were on the band, and could trigger more than one hammer at a time if characters on the band happened to be in the right location, and also had more than one set of the characters on the band. But you're right, printing the same character in every location on a line would not cause the triggering of all the hammers at once.

Band printers generally made a fast continuous chattering sound, except when the engineers ran their test patterns (often with the acoustic cover lifted), which made much more noise, probably deliberately, by printing worst case patterns during the test.

@ComicalEngineer

I can think of many Fortran 77 constructs that don't map well to BBC Basic, even Basic 4 on the Master and 65C02 second processor.

Either you wrote excessively simple Fortran, or you re-wrote significant parts of the program.

IIRC, it was possible to use the built in *EDIT command on the Master 128 to edit BBC Basic programs and re-tokenising them when you saved the file, making bulk text substitutions a little easier, but it still would have been quite a lot of work.

But I agree on your sentiments about the BBC Micro and it's surrounding ecosystem.

Oh, and Pr1me systems were generally regarded as super-minicomputers in the same league as contemporary VAXes, not mainframes.

Re: Interested in the type of 'mainframe'

Hmm. Don't agree about AIX not being UNIX. It is, by certification, quite definitely a UNIX, and if you ignore all the IBMisms that have been added, you can treat it as a UNIX system without any difficulty.

If you look at what makes a UNIX UNIX, it is mainly the API and the userland command set. AIX has all of this, being POSIX 1003, UNIX 98, UNIX 03 and UNIX V7 certified (about the only system that has UNIX V7)

Under the covers, well what is UNIX anyway. AIX from version 3.1 has a kernel that is written mainly in C (AIX 1.X for the 6150 PC/RT supposedly had much of it's VRM and microkernel re-written in a PL/I subset, but AIX 3.1 for the RS/6000 was mostly a re-port of SVR2, with bits of BSD grafted on and IBM enhancements - I know, I saw quite a lot of the code when I did an AIX Internals course back in 1990).

Since that time, IBM has enhanced AIX quite considerably, although I have no idea how much of the current code base is written in languages other than C.

What you can say is that compared with, say, Solaris, which has retrofitted technology from other UNIX and UNIX-like operating systems (specifically parts of the GNU tool set) into the base operating system, AIX is quite archaic in feel. It does not offer many of the GNU "--" flags on commands (but much of this is available as non-warranted add-ons from IBM). But AIX DOES implement what the UNIX standards require, it's just that other UNIX and UNIX-like OSes moved on beyond the standards.

One think that people (like you) point to is the changes in the way that AIX is administered compared to other UNIXs. But the management of UNIX systems was never included in any standard, and from my 45+ years of using and managing various UNIX systems from Edition 6 through SVR4 to the later versions of Solaris, HP/UX and Digital UNIX (and AIX, of course), there is and never was a common way of system management across different vendors versions of UNIX. Sun, HP, DEC and all the others, they all did it differently.

But one thing I would point out is what UNIX branded OSes are still under active development? Solaris - no. HPUX - also no. Digital UNIX - can you still buy it? MacOS - yes, but how may people use it as a UNIX system, it's only really a desktop OS nowadays. UNIXWARE/SVR5 - well maybe, it's a little unclear what XINUOS are actually doing with it. And everything else is either NOT officially UNIX (think all of the non-certified BSD derivatives or GNU Linux), or has fallen by the wayside.

I know that Liam has a different view of firstly what is UNIX (no, the fact that Inspur K-UX and EulerOS were at one time certified against UNIX 03 does not make all Linux systems certified), and secondly whether AIX is actually still being developed, but AIX 7.3 was release a few years ago, and both 7.2 and 7.3 are still having patches provided, as well as new features added as new Power systems are delivered. Expect a new TL for both 7.2 and 7.3 when Power 11 systems start being shipped.

Ummm.

Huh? Wordstar on a BBC Micro? Did you have the Z80 second processor?

You probably meant WordWise.

I preferred View myself.

Interested in the type of 'mainframe'

Then, as now, a real mainframe like 360/370/30XX systems (we are talking the time that the BBC Micro was new), did not have s single power switch (unless you include the EPO in the machine room), and even if you did end up powering down the CPU complex, applying power probably wouldn't have brought it back up again!

Mainframes had quite convoluted start-up sequences which often included resetting the disk string controllers and the communication controllers, and then using some arcane incantations through the MOC in order to boot the system. They were normally curated by a dedicated team of operators as well.

IMHO, he was probably talking about a System/34, or System/36 departmental computer (not a mainframe, but regarded as a midrange system), which were quite often installed in an office environment and managed by non-specialist staff.

But the whole thing about word-processing on any IBM system prior to the Displaywriter and/or PC (and maybe the 6150) was the screen-at-a-time terminal interface (3270/5250) that IBM used to use for their midrange and mainframe data systems. This would have been enough to drive a typist migrating from a real typewriter absolutely crazy until they had got used to it!

When I joined IBM in 1989 along with quite a few other new start IBMers recruited from the industry, this style of terminal was still in common use, and the number of people who had been introduced to computers using VAXes, UNIX systems, other mini's, PCs and home computers who struggled with the screen at a time nature of 3270 type terminals and PROFS (the most common IBM mainframe office suite for s370 and later mainframes) was almost everyone around me.

Ubuntu 25.10 plans to swap GNU coreutils for Rust

Re: License @Rich 2

The problem with not using any license at all for something any larger than trivial is that it begs someone to slap a copyright on your code, and then hit you with a copyright infringement, which you then have to fight through the courts.

Even if you can prove that the code was prior-art when the the copyright was registered, you still need to defend the claim, and this can cost LOTS of money.

This has happened to people publishing their own music on social media, particularly YouTube with their obnoxious take down claims, so don't think that it can't happen to your code.

Speed comparisons

I'm extremely interested in seeing how the speed of the Rust versions compare with the originals. We've seen lots of claims about speed comparisons, but real world examples of commands that I use on a regular basis is of much more immediate interest to me.

Dell discloses monster 20-petaFLOPS desktop built on Nvidia's GB300 Superchip

Desktop?

I have to admit that the picture does not seem to be a desktop, at least by my definition of a desktop. I would expect a desktop to be at most, the size of an IBM PC-AT, which was itself a large desktop.

It's probably best described as a (large) mini-tower, or a deskside system. Having one of those on a normal desk would not leave much room for monitors, keyboards or mice.

Earth's atmosphere is shrinking and thinning, which is bad news for Starlink and other LEO Sats

Re: Starlink = Starcrash!

The modern Thunderbirds is a show that was made relatively recently using computer animation. It was produced by an Australian TV company. I know that Gerry Anderson was not super impressed, but as he had sole the entire rights to ATV, he has no control over any of the Thunderbirds IP (which is why the atrocious live action film was allowed to be made).

I think that the rights are slowly reverting back to the Anderson Estate, and there are companies called Anderson Entertainment (set up by Jamie Anderson) and Century 21 Films Ltd. which between them are gathering together the IP to make it available in various forms again.

Re: Starlink = Starcrash!

Strangely, the first TV showing of UFO was put in the Saturday morning slot, just like the earlier shows. I don't think the schedulers actually realised that there were ideas and even scenes that were not really suitable for young children.

At the time, I think we had just transitioned from ATV to London Weekend Television (for the Saturday morning slot), so there was probably enough turmoil for them to not have realised the tone shift in UFO.

I'm not sure that Southern (which we could also receive, and had different scheduling as did all of the Regional ITV companies) ever carried UFO in that slot.

I do remember coming across it in a 10:30 or 11:00 PM slot sometime in the mid '70s. More recently, it had a long set of repeat runs on Forces TV on Freeview, but I think that channel has stopped broadcasting (even though it was a SD channel, you needed a DVB2 receiver to pick it up, but I think it was broadcast to British Forces Overseas as well).

Re: Starlink = Starcrash!

The only problem with Fireball XL5 was that the show was about 24 minutes long (made to fit a UK commercial TV 30 minute slot), and the first 4-5 minutes were the same loooooong launch sequence of XL5.

So you ended up with about 18 minutes of story (taking the closing credits into account).

I bought the collected set on DVD, and the repetitiveness of the opening sequence starts to grate if you're binge watching the series, and the chapter marks aren't really that helpful.

Stingray actually had shorter opening credits, so they don't have the same effect, and by the time they had to fill a 48-51 minute slot with Thunderbirds, they had got the trick of the forward flashes of the programme to fill out the length, although the stock shots of the launch sequences also managed to fill out the time.

So far, I have the complete Fireball XL5, Stingray, Thunderbirds, Captain Scarlet, Joe 90 and UFO. I don't really think I will bother with Twizzle, Torchey, Four Feather Falls or The Secret Service, but I may pick up Supercar on a whim if I see it. If I have time on my hands after I retire, I may pick up Space 1999, although I'm only really interested in series 1, I didn't really like the direction of series 2.

Ah, the memories of Saturday morning TV in the '60s when they used to be shown.

Re: An increase in space junk...

Don't forget the white jumpsuits and string vests!

Why was the real 1980 so boring.

Official HP toner not official enough after dodgy update, say users

HP were in this game. HP LaserJet 1000 (the original one, not the re-branded Samsung one) was a WinPrinter soooooo dumb, that it only had enough firmware to bootstrap a load of the real firmware down the parallel printer port (it didn't have a native USB port, but did come with an HP badged USB to parallel adapter). This required you to have software on the PC it was attached to to provide the firmware (honestly, how much does an EEPROM cost, even then!). Once loaded, the image was rasterised in the PC and sent as a bit mapped image to the printer.

There was, and this was a real surprise, a set of Linux print drivers produced BY HP(!) that enabled the Linux PC to provide the firmware, but it was a real faff setting it up, as it was provided as source in a tarball that had to be unpacked and compiled!

I bought one at a car boot sale for a few quid, and the person selling it said it wouldn't work with Windows 7 or later (which is why he was ditching it), which I took as a challenge to get it working from Linux. Came as a real surprise when I had to compile and install actual software rather than just the relevant PPD files to get it working, but get it working I did.

Used it for a couple of years attached to an equally obsolete Asus EeePC 701 (with extra memory and running from a microSD card) running Ubuntu as a network attached printer until I needed the space (I found I had amassed about 8 printers of different types, and eventually had to rationalise [because of the wife complaining], which resulted in several working but obsolete printers being sent to the tip).

The IT world moves fast, so why are admins slow to upgrade?

Re: Bucket, Paint, Brush, Corner?

I run systems that are clustered in an environment which is appallingly difficult to arrange full system outages. They were designed to be clustered from the outset, and generally speaking it is quite possible to do most types of patching by moving the services around the systems that are running them. For most types of maintenance, the down times for the entire system for infrastructure work like OS patching is normally measured in a few minutes per year, mostly as very brief interruptions as cluster moves take place (although I can't say the same for other components in the entire environment).

But this only goes so far.

You will eventually come to a point where the clustering software itself requires patching, and there are certain break points where you have to have the entire cluster down to patch/upgrade this software. I've also had situations where certain OS operations cannot be completed while the cluster is running (an example was when the cluster software plugged into an OS monitoring system, and would not allow that system to be patched without stopping the cluster).

These situations are actually few and far between, but eventually you have them. But the longer you run the systems, the more likely you are to hit one of these break points. The design of the environment I mentioned above is now 15 years old, and has been patched to keep it as current as possible, but the end of the road of this state is neigh!

Junior techie rushed off for fun weekend after making a terminal mistake that crashed a client

AS/400

AS/400s aren't as old as that. Launched in something like 1988, a long time after MS was set up (1975 or 1981 for the incorporation date), and even after some versions of Windows were produced!

Of course, the AS/400 was the follow on replacements for the System/36 and System/38 systems, so you could say that they pre-dated MS (maybe... I've not checked the actual dates!).