The 2038 computer bug

[Danny M2ZParticipant @dannym2z]

Just like the millenium bug was hyped this new vision has surfaced. 2038 bug

Just like the millenium bug I presume that some people shall try to profiteer from the hype, but then I wonder how many operating systems shal be 32 bit in 18 years time ?

* Danny M *

[Danny M2ZParticipant @dannym2z]

[Dave HalfordParticipant @davehalford22513]

Ones in museums.

[Nick Clarke 3Participant @nickclarke3]

Most computers nowadays use an internet time server that sets the time and unless unconnected to the internet, do not use the internal clock. PC clocks are notorious for drifting and even a small network will usually have a common time across it.

Time servers connect to an atomic clock, or connect to a computer directly connected to an atomic clock and so do not depend upon the UNIX time counting up – all that is necessary is for these servers to output the time information in a format that can be understood by the client computers. This can surely be changed with updates during the life of a system.

I am not so certain that there will be few 32bit computers in use in 2038 as many microcontrollers and embedded devices may well still 32 bit, however I am fairly certain that few laptop/desktop/tablet computers will be 32 bit, and in fact I doubt if many current laptop/desktop/tablet computers will still be in use at all then.

Windows 10 defaults to the Microsoft time server but you can change this to another should you wish to. Your ISP may also offer the option of a stantardised time.

[jimmy bParticipant @jimmyb]

I didn't worry about the last one. 18 years is to far away to think about!

Jim

[not done it yetParticipant @notdoneityet]

Will this affect my Acorn BBC computer?

[Brian WoodParticipant @brianwood45127]

I'm with Jimmy b—-I expect it will be something I won't be around any longer to have to bother myself about! Besides, the speed of development in this field will have very probably eliminated that kind of concern and exhausted the finances of the buying public struggling to keep up with the latest trends.

Brian

[Peter G. ShawParticipant @peterg-shaw75338]

Yay, 18 years to go. Good grief, I'll be er, damn near 100, certainly mid 90's, that is, if I'm still here which is very unlikely.

Let's assume I am here. And that I am still sufficiently mobile. And sufficiently strong enough to manhandle pieces of metal. Will I still be using my Warco 220 lathe? Possibly as it's a reasonably simple (like it's owner) machine. And my Sieg X2 clone milling machine? Now that might survive, but the track record of the electronics would suggest not. And then there's the NuTool vertical drill. Basic simple machine using a fractional horsepower single phase motor, no fancy electronics here. I rather fancy it might well be clapped out as it isn't that good now and it's only 26 years old. What about my computers? Ah, now there's a thing. I have a 12 year old Toshiba main machine, a slightly newer Toshiba backup machine, and a rescued from on its way to the tip Advent, all being laptops, and all running Linux Mint and all showing some physical damage. Perhaps they'll be operated by brain waves?

But much more to the point, my grandsons will be 43 and 41, whilst my granddaughters will be 31 and 22. Perhaps I'll have great grandchildren. Nice to think so, but I rather think that any great grandchildren will be wondering who is that drooling old man in the corner!

Enough, must stop rambling on,

Peter G. Shaw

[Peter SansomParticipant @petersansom44767]

This has long been know. I remember reading about it in the 1990's. The issue is caused by the Unix operating system calculating time by using the number of seconds since the the first Unix computer was turned on in 1970. It can be mitigated my the OS kernel correct the clock when the 32bit time integer rolls over. Remember that we started with 8bit computers, then 16bit. 32bit only came in the windows world wit the 80386.

How many of you have a 80286 or older 8086, so don't worry about.

If you are worried about this remember there is a calendar correction due in about 2000 years time, extra day if I remember, read it many years ago.

That is someone elses problem, if humans are still here.

[V8EngParticipant @v8eng]

Now let’s see: the end of the computing world was nigh at the turn of 2000, that didn’t happen so I am not going to worry about 2038.

I shall be about 93 so guess the main worry might something other than that.

Activists of various types will likely have driven us back to the dark ages long before then anyway.

[Brian GParticipant @briang]

Posted by jimmy b on 12/07/2020 10:45:37:

I didn't worry about the last one. 18 years is to far away to think about!

Jim

I first hit problems due to Y2K in 1984, so 18 years isn't that much farther away! In my case, long term planned-maintenance orders for spare parts to be delivered early in the 21st century were generating requirements 100 years early, and whilst we negotiated to remove the orders from our system and hold them manually, the MRP system attempted to reschedule supply orders back to the early 1900s (sometimes future order suggestions were generated as well because "stack to stock" lead times went back before 1900, which put them in the 1980s and 90s, as strictly speaking Y2K didn't just affect the millenium but any rollover of 2-digit years), so that all of our production planning, financial forecasting and purchase expediting was thrown into chaos.

In a later job we started planning for Y2K about 1995 and had new systems in place by the end of 1998 (the supplier of our original software was no longer in business, so we took the opportunity to move to new production, inventory and accounting systems) so we were able to avoid the financial hit that we would have taken if unable to process orders or invoice customers. I suspect that a lot of Y2K spending, whilst it was necessary to avoid the problem, would have taken place anyway as part of organisations' normal upgrade processes.

Brian G

[Brian OldfordParticipant @brianoldford70365]

It's interesting to read peoples' comments about how little the "Millenium Bug" affected people. The real reason was because of the hard work put in before the date by IT professionals testing and updating systems to avoid those problems.
I worked in IT at that time and was on standby that night. Because of that hard work I earned a nice crust without a single phone call.

[Neil WyattModerator @neilwyatt]

Posted by Brian Oldford on 13/07/2020 15:12:00:

It's interesting to read peoples' comments about how little the "Millenium Bug" affected people. The real reason was because of the hard work put in before the date by IT professionals testing and updating systems to avoid those problems.
I worked in IT at that time and was on standby that night. Because of that hard work I earned a nice crust without a single phone call.

Reminds me of some friends who set up a company called Dynamic Data Technology.

DDT – kills all known bugs…

Neil

[Brian OldfordParticipant @brianoldford70365]

Posted by Neil Wyatt on 13/07/2020 15:20:25:

Reminds me of some friends who set up a company called Dynamic Data Technology.

DDT – kills all known bugs…

Neil

I'm unsure whether to laugh or groan.

[Speedy Builder5Participant @speedybuilder5]

+1 for Brians comment. There is no way a 2 digit date was going to work correctly on Jan 1st 2000. In fact it stopped working long before that for forward order dates, end of guarantee periods etc. Young uns can't comprehend how expensive disc storage was back then, the tricks we had to use to minimise data sizes, the processor speed (Or lack of). I remember that a 16Kbyte core memory board cost about £8,000 in 1975 and our machine was a 32Kb one but in spite of that, we ran the complete stock control, BOMs, purchasing and spares and rectification system for CINCINNATI Milacron UK. Our total disc storage was 48Mb which was production, development, operating system.

[SillyOldDufferModerator @sillyoldduffer]

As if Covid-19, Brexit, and Global Warming weren't enough to worry about! Due to time_t being a signed 64 bit integer on Ubuntu I now only have until 15:30:08 on Sunday, 4 December in the year 292277026596 to fix my computers. At the rate I finish projects, I shall have to ask for an extension – that's less than a million billion days!

Dave

[Danny M2ZParticipant @dannym2z]

Posted by Speedy Builder5 on 13/07/2020 16:24:42:

+1 for Brians comment. There is no way a 2 digit date was going to work correctly on Jan 1st 2000. In fact it stopped working long before that for forward order dates, end of guarantee periods etc. Young uns can't comprehend how expensive disc storage was back then, the tricks we had to use to minimise data sizes, the processor speed (Or lack of). I remember that a 16Kbyte core memory board cost about £8,000 in 1975 and our machine was a 32Kb one but in spite of that, we ran the complete stock control, BOMs, purchasing and spares and rectification system for CINCINNATI Milacron UK. Our total disc storage was 48Mb which was production, development, operating system.

Lol, you were lucky.

In the early '70'sI worked with 1965 technology magnetic drum memory systems that were used to compute the parabolic flight of a projectile, a mixture of valves and an early version of a transistorised computer used to drive me crazy debugging the faults, but when it worked the operators were very happy.

The drum had enough memory to store azimuth, elevation and bearing data points for a single radar tracked projectile which were then integrated to plot the origin or the destination using fixed constants for gravity and air drag. We were told (1974) that a replacement magnetic drum unit cost $120,000

On a good day, the people firing the projectiles could be located to within 30m at 10km range so were not happy when a quick CB return fire of airburst arrived over their location.

That was a long time ago but it's factual and it's history.

* Danny M *

[larry phelan 1Participant @larryphelan1]

Peter,

My Nu Tool drilling machine was bought in 1983 and is still going strong, might think about getting new belts for it some day !

But you know what they say "If it aint broke, dont fix it "

Good advice !

[Peter G. ShawParticipant @peterg-shaw75338]

Larry,

Normally I would agree with you, but such is the way of our hobby that there is always a tendency to want to improve and my drill could certainly do with some improvement, eg to reduce the sideways slop in the quill. In fact, I've a sneaky suspicion that I have seen someone else with exactly the same problem and which was written up for either ME or MEW. That solution was to make a cut in the housing and then clamp it up thus reducing the sideways slop. At least, I think that was it, although I do remember the cutting bit. I must admit that the writer was a braver man than me!

Cheers,

Peter G. Shaw

[Howard LewisParticipant @howardlewis46836]

My company overcame the Millenium bug by resetting all the dates on the computer to 1972. The only problem that we saw was that looking up engine history, the screen showed that it had been built, tested and despatched before the design drawings had been made!.

A quick mental calculation gave us the real dates.

IF there will be a problem, a solution will have been found by then. Otherwise I shan't get my telegram from the Palace!

Howard

[Author]

Posts