Posts by Peter Gathercole

Re: Census data

My wife has a hobby of genealogy, and she's actually quite good at it. One of the things she has to cope with is different spellings of unusual (and some not so unusual) names, when one census taker had a different guess at how the name was spelt from the last or following census.

Trying to track people who move from one place to another when their names may not have been recorded correctly can be a real problem.

You also have people who switch their first names around at different times, give known names rather than their christened names, or deliberately want to disappear and try to give different information in different places. It's all a bit of a game, really, and one that a lot of amateur genealogists (and some so-called professionals) get wrong, and then post incorrect information onto Ancestory and FindMyPast, poisoning the record!

My wife does not count a link until she can find at least two separate pieces of evidence linking people together.

Re: Ohhh...real shops!

I've got a particular problem. Years ago, I bought a mobile phone on a contract for one of my kids.

They're kind of wedded to the number on that contract (even though it's actually well out of contract), but it's in my name, even though they've taken over paying for it.

So we want to move the number to another contract in his name. It's proved difficult over the phone, because it seems it needs a three way call with the phone provider's call center, and we both have to prove who we are, which seems too difficult for the call center operatives to get their heads around.

One time, before the lockdown and while he was visiting, we decided that it might be better to visit the company's shop, so we both loaded ourselves up with ID, and travel to the nearest shop (~25 miles).

We walk in, and explain to the young person there what we wanted to do, and after his face paled for a few seconds, he said "No, I can't do that here, you'd be better off calling the call center."

Made me wonder what the purpose of the shop actually was.

Re: I miss IBM keyboards

Yes, it's quite true with the IBM Model M. There's a big steel plate forming the bottom of the plastic-membraneX3-rubber_shim-steel sandwich inside the plastic case. It's not just for weight, it's actually the bottom layer, with the rubber sheet and the the lower membrane immediately above it, and is also curved so that the rake is not just the height of the keycaps, but a formed by the plate.

There's videos on YouTube about bolt-modding the keyboards which show you exactly how they are put together.

With the Unicomp keyboards, the plate is of a lower gauge than the original IBM ones, and I get the impression that the plastic of the external case is either thinner, or of a less dense plastic as well (it's certainly more flexible when it's apart). This makes the whole keyboard, though still heavy by today's standards, lighter than a real Model M.

Re: Dishwashers

The previous model F keyboards (on the 5150 and other IBM systems) operated by capacitance change. Model M's used a reduced cost mechanism compared to the model F, and are definitely contact based.

The rocker mechanism is plastic in the model M (I've had several of them apart, and they are definitely plastic), and operates when the spring buckles, and presses the two layers of the membrane together, in the same way that the rubber dome of cheaper keyboards do. I've even operated the membrane without the rocker.

Believe it or not, the model F was even more clicky (well, actually more clunky) than the model M. I thought the feel of the model F was absolutely amazing (better than the later model M), but unfortunately, the layout was quite eccentric, although there was not really a standard keyboard layout back then. The model M shipped with the 5170 PC/AT really defined that (although the physical key positions - although not the character layout - of the DEC LK201 keyboard was very similar).

Re: Smoking

With both of the rebuilds, when I dismantled the keyboards, the sticky mess that was left after the drink had also eaten into the tracks on the membrane.

The first rebuild, a real Model M, I did not soak the keyboard (I had not heard of the process), just went straight in to the stripdown, and found the membranes stuck together. Unsticking them ruined the tracks, and even though I tried to use conductive paint, I could not get all of the keys working, even after several attempts.

The second one, a Unicomp, I knew about the soak method. I tried it with the whole keyboard, and then tried with the plastic interior out of the case and with the complete key set of keys and spring/rocker assemblies, but it still didn't return the keyboard to an operating state.

Fearing the worst, I cut the plastic rivets, and found that the Unicomp membrane was actually mostly sealed with a mastic bead, which prevented the water getting in in sufficient quantity to clean up the gunk (the previous IBM one was open on all sides, so probably would have worked). The drink had seeped into one of the holes of the rivets.

But even opening the membrane, and cleaning it well by hand was still not enough. After bolt-modding it and putting it back together, I found that two keys would not work. So I stripped it down again, and with a meter found that one track had been apparantly eaten into by the drink. Resorting to the conductive paint again, this time I managed to restore the tracks, and got all the keys working.

I think that the IBM board failed because of the higher resistance of the painted tracks, and the difference in electronics (PS/2 vs. USB).

The bolt-mod actually made the feel of the keyboard better (some of the plastic rivets had already broken - the keyboard was my son's, who has a machine-gun like typing style).

Re: Dishwashers

I believe that one of the models of Cray supercomputer used to immerse some of the boards in a non-conductive, inert liquid to aid heat dissipation. I can't remember what it was called, but I understand it was green.

Re: Dishwashers

I would suggest not using any dishwasher detergent or tablets. They tend to have mild abrasives and slightly corrosive components in them that may leave a residue inside the keyboard or damage PCBs.

I would probably not use anything, just let the water do it's stuff, but it might depend on how hard the water is in your area.

Re: I miss IBM keyboards

I have bought two Unicomp keyboards. They're definitely good, and quite robust (one of the Model M's I referred to above as having been repaired after exposure to carbonated drinks was a Unicomp one), but they are quite a bit lighter than my real 1992 Model M. Also, some of them appear not to have the 2 part key caps that a real model M had.

Also, the moulds are wearing after all this time, so they're beginning to look a bit lower quality.

I would still recommend them for people who want a positive typing experience, but I would say that some of the individual key-switch keyboards are now quite good as well (Model M's are actually membrane keyboards, the feel comes from the buckling spring mechanism). The one I'm using at the moment is a SUMVISION Acies ten-key-less keyboard (my Model M is too big to cohabit with another keyboard on my working-at-home desk), and significantly cheaper than a Unicomp Model M.

The only real problem is that the key legends require you to have the keyboard lit to be able to see them, and they're a bit garish with the lights on. I might just replace the LEDs with white ones to see whether that makes any difference.

Re: Smoking

Yes. Sticky carbonated drinks will kill even a Model M (2 attempted rebuilds, one worked, one failed).

Re: Not Cheap

Wow. When I was buying IBM Model M's for X-Stations and IBM 43Ps, we were only paying about £105, and I thought that was expensive!

Mind you, buying a Unicomp Model M is about the same price now, although in real terms, they are much cheaper because of inflation.

Re: keyboard condoms

My pet hate was when people would come up to my screen, and poke it with greasy fingers, or even worse, a pen.

Back in the day of matte ground screens on CRTs to prevent reflections, greasy marks were very noticeable, and ballpoint pen ink was very difficult to get off.

Tends to happen rather less with LCD screens now. I guess that either they're further away at the back of the desk, or people realize that they are more fragile.

Talking about fragile, during an office move a while back, a significant number of LCD monitors were ruined because people stuck labels to the screen part of their screen with sticky tape, and then ripped the surface when trying to peel the tape off!

Re: Experience is the best teacher

I agree with the above AC.

I've always tried to make my changes as complete as possible, and include the possible impact and the backout plan, and in general I've rarely had a problem with getting my changes through the change board.

Most of the time, the purpose of the change board is to make sure that adequate thought and planning has gone into the change, and if you achieve a good track record for accurately describing and implementing the change, and being able to cope when problems occur within the description of the backout process, then you generally have a relatively easy ride through the board. If problems happen that weren't anticipated, then there will need to be a wash-up afterwards, because it probably means that something wasn't considered.

If a change needs further scrutiny during the board, it probably means that you've either overlooked something, or that it is affecting something of critical importance, and people want to be absolutely sure that everything has been covered. These are valid outcomes of the board, although you may see it as obstructive if it prevents you doing the change. In that case, work harder to make sure it doesn't happen again.

This can come down to be a matter of trust, and you have to earn that trust,

It also had a number of instructions that made writing compound arithmetic operations easier, in particular the barrel shifter.

Re: Retro-Wreckers

IIRC, when the 68000 was launched, there was going to be a problem with support chips for it, as Motorola concentrated on getting the processor out of the door.

To compensate for this, they built in the ability to use support chips from the successful 6800 and 6809 families of their product set, which only supported an 8 bit data bus.

I'm sure that it was this ability that allowed Motorola to offer the 68008 at all. They just disabled the 16 bus hardware out (I'm pretty certain that the original 68000 only had a 16 bit data bus, the wider address bus parts came later) and stuck it in a smaller package.

It is quite possible that the silicon came off the same production line, but were in some way test failures, but this is just speculation on my part.

Re: "Some dealt with the RAM pack with..."

That would have been Personal Computer World. Chimps were used on the covers of all of the issues covering Sinclair Computers from the Z80 to the QL. IIRC, the chimps on the QL cover were wearing bowler hats.

Re: "Some dealt with the RAM pack with..."

My solution was to separate the keyboard from the rest of the system (well, add an external keyboard, I guess).

I took a Tandy membrane keyboard I bought in one of their 10% off per day sales, opened it, scratched and repainted the conductive paint traces to match the ZX81 keyboard, and extended them to include the space bar. I connected this to the ZX81 wlth a ribbon cable that ran through a slot I cut in the front of the case. I took a photo of the ZX81 and cut it up to put key legends on the keys.

Fitted into a case built from thick styrene sheet, with a switch for power, and it worked a treat. Once you isolated the keyboard from the rest of the system, everything became much more stable, especially if you had any additional expansion cards (I had the Quicksilver AY-3-8910 sound board, together with a second modulator to put sound through the TV). It was not as elegant as the system in the article, but it worked very well.

I never got round to putting Quicksilver's Hi. Res. graphics card in, but I did find a neat trick to implement programmable graphic characters with an extra 1K of static RAM on the ROM side of the R7-R14 data bus isolation resistors and by manipulating the I register.

Re: Hats off

If the pandemic had hit in the '90s, chances are the phone network would have melted down, not the proto-internet. Business would have been conducted over the phone, by fax, and for a few people, buliitin boards and email, with an even smaller number of people using dial-in serial sessions.

For people doing traditional office jobs, the best they probably could have hoped for would have been a PC at home, and file transfer to and from the office, if they could have worked out the way of getting to their Netware and SMB file shares. And fewer people would have had PCs at home already, so the PC supply chain would also have melted, at least if businesses could fund the high cost of business class PCs at the time.

Whether people would be able to do what they did in the office would be a moot point. I suspect that only a small number of people would have been able to continue working normally.

Ironically, I probably could have done most of my normal work from home in the early '90s, but at that time I would probably have been one of the people called into the office to keep the IT systems running.

One of the things I had worked out in the early '90s was how to drive an X11 session through the modem and firewall infrastructure that we had at the time (it was mainly done for computer shows and remote support, but I used it myself when I was at home), but I suspect that it would have melted down with more than a couple of sessions, even if we could have provided systems that could run capable X11 servers at home. At the time, I had an IBM 6150 (AIX) at home, which was a little more than a PC (and much larger), but ran X11 and Slip reasonably well

I was not a particularly early adopter of home internet, not getting a PC capable of connecting to an ISP until 1996, and even then it was at 14.4Kb/s, which was also about the time I also started using Linux.

Re: Elite

The disk space was actually more restricted than 100K.

The original disks (if you bought one) was marketed as a "Dual 40/80 track disk" that would work on a single stepped 40 track drive, and both single and double stepped 80 track drives.

IIRC, it did this by fiddling with the number of tracks that were recorded, and having 6 of the 80 tracks (or three of the 40 tracks) near the beginning of the disk that were non-standard formatted, which allowed it to work out what the type of disk drive was for. I think that the track number was actually written into the sectors, allowing the program to seek to a track, and read the track number, which then allowed the program to work out what type of drive it was.

This meant that the program actually fitted into 37 of the 40 tracks.

As a side effect (or maybe it was the original intent), this made normal disk copiers fail to copy the disk.

Again, if my memory serves me, these 6 tracks were formatted with 128 byte sectors rather than the normal 256 byte sectors.

I remember that Acorn User published an advanced track-by-track disk copier that directly accessed the 8271 controller registers to implement a 'universal disk copier', although it assumed that the disk format was uniform across the whole disk (it checked the sector size and number of sectors on track zero, one and two, and then applied the format to the entire disk).

I took the basic 8271 programming procedure from that article, and used it to analyze an original Elite disk, and then wrote my own track-by-track copier that read the format and data of each track, formatted the same track on the new disk with exactly the same sector sizes, and then wrote the data. It was horrendously slow, because it only worked a few tracks at a time because of the limited memory in the system, but it created 100% playable Elite copies (and I believe that it would have coped with any valid or even some invalid disks, as long as they kept within the capabilities of the 8271).

The 8271 was an obsolete chip before Acorn chose to use it in the BBC micro. and it had severe limitations. The WD 1770/2/3 controllers were significantly more capable, but weren't compatible. They did appear in later BBC Micros (I think in the Electron and either the B+ or the B+128, and all later systems). Some vendors who provided 1770 controllers for the original BBCs also built a 8271 emulation into their adapter and DFS that allowed programs that directly accessed the 8271 (like the original Elite disks) to work, but not all did that, meaning that old versions of Elite would not work on BBC micros with some vendors 1770 disk controllers. I don't believe that the Acorn 1770 DFS did, making it a poorer (even if it was an original Acorn design) version than some of the third party ones.

Of course, later versions of the Elite disk would work with WD1770 or Intel 8271 controllers, and I think that Acornsoft actually offered a disk-swap for people who upgraded their machines.

Re: One thing is for sure @Pascal

Are you really thinking it's going to take at least 9 more years at least to achieve this? Or did you mean the 2020s?

Flexible use of SFTP

There are full filesystem interfaces that use SFTP as a transport layer. It's quite impressive, although I would guess that it's not the most efficient means of transferring files.

All things like this are possible over more simple protocols. Anyone remember Archie?

Re: Better state of mind for everybody: I do not touch other's computers @The Basis

I thought you were stealing my lines.

I do look after systems with 100s of processors (at least in this role, the last one was around 40,000 Power 7 CPUs) that do not run Windows and run large environments. My employer provided laptop runs Linux, as do all of my own systems, and those of my Wife.

Unfortunately, I do have to run Windows on a remote-access system, just enough to get Putty running through Citrix gateways (the RAS system is built around Windows), and my kids have Windows gaming rigs, but I don't actually look after any of those systems.

None of this is a lie. It can be done. And it is possible to gloat a little when people say "You know about computers, can you help me with my PC" and I turn around saying that I know very little about Windows.

Re: I despise that thing

The rubber nipple can be removed, leaving a plastic stub that is below the level of the surrounding keys.

It's not particularly aesthetically pleasant, but it gets it out from under your fingers.

Me? I struggle with laptops which don't have a trackpoint, and I disable any touchpad on a laptop if I can.

Re: How is this possible? @me

I feel so stupid. Algol 68. It was Coral 66.

Must get more sleep to get the ol' memory working better!

Re: How is this possible? @Vincent

There is nothing wrong with having different operators for test and assignment. It's only the lax style that was encouraged by languages such as Basic that made programmers think that it's a good idea to use the same operator.

Having two different operators just makes it very clear what is happening when you can use tests and assignments in arithmetic operations. I admit that you have to know how TRUE and FALSE are represented in the language to benefit from this, and you can get some very hard to read statements, but having explicit delimiters can benefit bot the compiler and the programmer.

There are a huge raft of programming languages where test and assignment are represented by different operators. You just have a particular bias.

The semicolon (or other delimiter - Python is picky about end-of-lines themselves) at the end of a statement allows complex mult-line statements to be adequately delimited. It allows you to do complex things without having to assign values into variables, allowing the compiler to automatically optimise the storage and discarding of intermediate results that otherwise would need variable assignment. And this allows the compiler to pick up when nesting of blocks is incorrect by counting the levels of block delimiters and comparing that to the statement delimiters.

Sure, the compiler can work out where it thinks the end of a statement should be, but having used teaching languages that attempt to correct missing end of statement delimiters, I've seen the bizarre and sometime complex problems that this can cause. No, I'm firmly on the side of the making a language as concise as possible.

Just imagine the problems if you tried to write a free-format natural programming language. How often does English allow ambiguous instructions? There's a whole profession (lawyers) who make their living trying to make sense of English!

Re: Gone but not forgotten

It was Eadon.

Funny, I've just tried to find his comment history, and it looks like it's been expunged from The Reg. comments history. All of the comment trails lead back to a "This post has been deleted by a moderator".

I remember I had some run-ins with him, but none as memorable as the ones I had with "Kebabert"

Thinking back, there's a lot of previously frequent commenters who have disappeared. Is the readership of The Register aging that fast, or are people just losing interest?

It would be interesting to have a chart of the top 100 commenters every year since The Register started keeping stats, and follow up on the people who no longer comment. Maybe we should sent the neighbors round to check that they're OK!