I’m wondering if all the electronics inside modern rolling stock is going
to be the ultimate limiter of lifetime. There was a recent article in The
Register about DB advertising for a Windows 3.11 administrator to keep some
in train display stuff going. There’s going to have to be a very big spares
inventory to keep things going over a 30 to 40 year lifetime, plus the need
to maintain the knowledge of how it all works. All the things that look
terribly modern now are going to be very old fashioned in a few decades
time.

On 01/02/2024 12:53, Tweed wrote:
> I’m wondering if all the electronics inside modern rolling stock is going
> to be the ultimate limiter of lifetime. There was a recent article in The
> Register about DB advertising for a Windows 3.11 administrator to keep some
> in train display stuff going. There’s going to have to be a very big spares
> inventory to keep things going over a 30 to 40 year lifetime, plus the need
> to maintain the knowledge of how it all works. All the things that look
> terribly modern now are going to be very old fashioned in a few decades
> time.
>
>

Regrettably that is the way of the world now. I am sick and tired of
computer magazines telling me to ensure that I scan photos etc "to
ensure that I can keep them for the future" or other inane comments. I
have real photos taken 50-60 years ago which are still good. Even if the
floppies did not corrupt over time I would still have problems looking
at corrspondence from 20 years ago as the software was obsolete years
ago - remember Tasword anyone?

--
Grumpy

On Thu, 1 Feb 2024 14:35:20 +0000
ColinR <rail@greystane.shetland.co.uk> wrote:
>On 01/02/2024 12:53, Tweed wrote:
>> I’m wondering if all the electronics inside modern rolling stock is going
>> to be the ultimate limiter of lifetime. There was a recent article in The
>> Register about DB advertising for a Windows 3.11 administrator to keep some
>> in train display stuff going. There’s going to have to be a very big spares
>
>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>> to maintain the knowledge of how it all works. All the things that look
>> terribly modern now are going to be very old fashioned in a few decades
>> time.
>>
>>
>
>Regrettably that is the way of the world now. I am sick and tired of
>computer magazines telling me to ensure that I scan photos etc "to
>ensure that I can keep them for the future" or other inane comments. I
>have real photos taken 50-60 years ago which are still good. Even if the

I've got photos of when I was a baby in the 70s and unfortunately most of
them have faded badly despite being kept out of the light for most of the time.
Good quality film might last for decades but consumer stuff really didn't.

>floppies did not corrupt over time I would still have problems looking
>at corrspondence from 20 years ago as the software was obsolete years
>ago - remember Tasword anyone?

Use jpeg for pictures and .txt for any important documentation. You'll lose
all the formatting and diagrams etc with the latter but the text is the most
important thing.

In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
2024, Tweed <usenet.tweed@gmail.com> remarked:

>I’m wondering if all the electronics inside modern rolling stock is going
>to be the ultimate limiter of lifetime. There was a recent article in The
>Register about DB advertising for a Windows 3.11 administrator to keep some
>in train display stuff going. There’s going to have to be a very big spares
>inventory to keep things going over a 30 to 40 year lifetime, plus the need
>to maintain the knowledge of how it all works. All the things that look
>terribly modern now are going to be very old fashioned in a few decades
>time.

On the other hand, as you can get Raspberry PIs with onboard emulators
of 1980's 8-bit home computers, the idea you need inventory of old
hardware is false.
--
Roland Perry

In message <upga76$23ovp$1@dont-email.me>, at 14:35:20 on Thu, 1 Feb
2024, ColinR <rail@greystane.shetland.co.uk> remarked:
>On 01/02/2024 12:53, Tweed wrote:

>> I’m wondering if all the electronics inside modern rolling stock is going
>> to be the ultimate limiter of lifetime. There was a recent article in The
>> Register about DB advertising for a Windows 3.11 administrator to keep some
>> in train display stuff going. There’s going to have to be a very big spares
>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>> to maintain the knowledge of how it all works. All the things that look
>> terribly modern now are going to be very old fashioned in a few decades
>> time.
>
>Regrettably that is the way of the world now. I am sick and tired of
>computer magazines telling me to ensure that I scan photos etc "to
>ensure that I can keep them for the future" or other inane comments. I
>have real photos taken 50-60 years ago which are still good. Even if
>the floppies did not corrupt over time I would still have problems
>looking at corrspondence from 20 years ago as the software was obsolete
>years ago - remember Tasword anyone?

Yes, and I've probably got a copy of Tasword somewhere. Retro computing
enthusiasts would stick it on a thumb drive which emulates a 3" floppy.

I'm typing this on software dating back to 1995, and which does its job
perfectly well.
--
Roland Perry

Roland Perry <roland@perry.uk> wrote:
> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>
>> I’m wondering if all the electronics inside modern rolling stock is going
>> to be the ultimate limiter of lifetime. There was a recent article in The
>> Register about DB advertising for a Windows 3.11 administrator to keep some
>> in train display stuff going. There’s going to have to be a very big spares
>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>> to maintain the knowledge of how it all works. All the things that look
>> terribly modern now are going to be very old fashioned in a few decades
>> time.
>
> On the other hand, as you can get Raspberry PIs with onboard emulators
> of 1980's 8-bit home computers, the idea you need inventory of old
> hardware is false.

Emulating the OS is not the only problem to be solved. At least the
following issues still arise: hardware interfaces, interface protocols,
people who understand the old stuff (new hires aren’t going to want to base
a career on obsolete technology). Retrofits of new hardware and their
associated code will need to be tested and certified, especially anything
safety critical. Everything is solvable with enough money and I suspect the
aerospace industry works around the problem because there’s lots of old
aircraft of the same type to make it worthwhile. However relatively short
production runs of UK bespoke stock is a different matter.

Am 01.02.2024 um 15:35 schrieb ColinR:
> On 01/02/2024 12:53, Tweed wrote:
>> I’m wondering if all the electronics inside modern rolling stock is going
>> to be the ultimate limiter of lifetime. There was a recent article in The
>> Register about DB advertising for a Windows 3.11 administrator to keep
>> some
>> in train display stuff going. There’s going to have to be a very big
>> spares
>> inventory to keep things going over a 30 to 40 year lifetime, plus the
>> need
>> to maintain the knowledge of how it all works. All the things that look
>> terribly modern now are going to be very old fashioned in a few decades
>> time.
>
> Regrettably that is the way of the world now. I am sick and tired of
> computer magazines telling me to ensure that I scan photos etc "to
> ensure that I can keep them for the future" or other inane comments. I
> have real photos taken 50-60 years ago which are still good. Even if the
> floppies did not corrupt over time I would still have problems looking
> at corrspondence from 20 years ago as the software was obsolete years
> ago - remember Tasword anyone?

Correct. Protecting digital assets requires the following steps:

1) move them from proprietary formats to standardized formats (TXT, PDF,
JPG, PNG, DOCX) as soon as converters are available
2) store at least two copies in different locations. Validate the
readability regularly.

This is only minimally better for "consumer grade" than the past
consumer grade photos which faded by the decades with no possibility for
rescue.

I failed at step 2; I had to pay €100 recently to get my "unique" family
videos moved from the hard disk (dated appox. 2010) with broken USB port
to a new SDD disk.

In message <upgj9e$25b6p$1@dont-email.me>, at 18:10:06 on Thu, 1 Feb
2024, Rolf Mantel <news@hartig-mantel.de> remarked:
>Am 01.02.2024 um 15:35 schrieb ColinR:
>> On 01/02/2024 12:53, Tweed wrote:

>>> I’m wondering if all the electronics inside modern rolling stock
>>>is going to be the ultimate limiter of lifetime. There was a recent
>>>article in The Register about DB advertising for a Windows 3.11
>>>administrator to keep some in train display stuff going. There’s
>>>going to have to be a very big spares inventory to keep things
>>>going over a 30 to 40 year lifetime, plus the need to maintain the
>>>knowledge of how it all works. All the things that look terribly
>>>modern now are going to be very old fashioned in a few decades time.

>> Regrettably that is the way of the world now. I am sick and tired of
>>computer magazines telling me to ensure that I scan photos etc "to
>>ensure that I can keep them for the future" or other inane comments. I
>>have real photos taken 50-60 years ago which are still good. Even if
>>the floppies did not corrupt over time I would still have problems
>>looking at corrspondence from 20 years ago as the software was
>>obsolete years ago - remember Tasword anyone?
>
>Correct. Protecting digital assets requires the following steps:
>
>1) move them from proprietary formats to standardized formats (TXT,
>PDF, JPG, PNG, DOCX) as soon as converters are available
>2) store at least two copies in different locations. Validate the
>readability regularly.
>
>This is only minimally better for "consumer grade" than the past
>consumer grade photos which faded by the decades with no possibility
>for rescue.
>
>I failed at step 2; I had to pay €100 recently to get my "unique"
>family videos moved from the hard disk (dated appox. 2010) with broken
>USB port

USB ports on hard drives is fairly unusual.

Unless you mean a USB port on an external HDD unit (where every HDD
inside I've ever seen has a conventional interface on it). If the
latter you were well and trully ripped off.

>to a new SDD disk.

Assuming you actually mean SSD, having had several fail, I wouldn't
trust them further than I could throw them (into the bin, mainly).
--
Roland Perry

In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
2024, Tweed <usenet.tweed@gmail.com> remarked:
>Roland Perry <roland@perry.uk> wrote:
>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>
>>> I’m wondering if all the electronics inside modern rolling stock is going
>>> to be the ultimate limiter of lifetime. There was a recent article in The
>>> Register about DB advertising for a Windows 3.11 administrator to keep some
>>> in train display stuff going. There’s going to have to be a very
>>>big spares
>>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>>> to maintain the knowledge of how it all works. All the things that look
>>> terribly modern now are going to be very old fashioned in a few decades
>>> time.
>>
>> On the other hand, as you can get Raspberry PIs with onboard emulators
>> of 1980's 8-bit home computers, the idea you need inventory of old
>> hardware is false.
>
>Emulating the OS is not the only problem to be solved. At least the
>following issues still arise: hardware interfaces, interface protocols,
>people who understand the old stuff (new hires aren’t going to want to base
>a career on obsolete technology). Retrofits of new hardware and their
>associated code will need to be tested and certified, especially anything
>safety critical. Everything is solvable with enough money and I suspect the
>aerospace industry works around the problem because there’s lots of old
>aircraft of the same type to make it worthwhile. However relatively short
>production runs of UK bespoke stock is a different matter.

Whatever. I can tell you don't have a grip of the concepts I was
discussing. Or this time is it you who is trolling?
--
Roland Perry

Roland Perry <roland@perry.uk> wrote:
> In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
> >Roland Perry <roland@perry.uk> wrote:
> >> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
> >> 2024, Tweed <usenet.tweed@gmail.com> remarked:
> >>
> >>> I’m wondering if all the electronics inside modern rolling stock is going
> >>> to be the ultimate limiter of lifetime. There was a recent article in The
> >>> Register about DB advertising for a Windows 3.11 administrator to keep some
> >>> in train display stuff going. There’s going to have to be a very
> >>>big spares
> >>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
> >>> to maintain the knowledge of how it all works. All the things that look
> >>> terribly modern now are going to be very old fashioned in a few decades
> >>> time.
> >>
> >> On the other hand, as you can get Raspberry PIs with onboard emulators
> >> of 1980's 8-bit home computers, the idea you need inventory of old
> >> hardware is false.
> >
> >Emulating the OS is not the only problem to be solved. At least the
> >following issues still arise: hardware interfaces, interface protocols,
> >people who understand the old stuff (new hires aren’t going to want to base
> >a career on obsolete technology). Retrofits of new hardware and their
> >associated code will need to be tested and certified, especially anything
> >safety critical. Everything is solvable with enough money and I suspect the
> >aerospace industry works around the problem because there’s lots of old
> >aircraft of the same type to make it worthwhile. However relatively short
> >production runs of UK bespoke stock is a different matter.
>
> Whatever. I can tell you don't have a grip of the concepts I was
> discussing. Or this time is it you who is trolling?

Emulation of the software is the easy bit. Obscure hardware is troublesome.
How do you propose to connect some piece of hardware with a SCSI port, GPIB
or a BBC Micro 1MHz bus connection to your emulated system?

It may be possible to find a solutions, but the next problem is
certification. This thing is used in aviation/nuclear/medical applications,
and modification potentially means the whole system needs to be recertified.
Suddenly the costs go up by orders of magnitude.

It is not easy or cheap, and that's why keeping the old thing going is the
most expedient thing to do. Even if that means you need a cupboard full of
8 inch floppy drives.

Theo

In message <jGr*CHVBz@news.chiark.greenend.org.uk>, at 18:22:41 on Thu,
1 Feb 2024, Theo <theom+news@chiark.greenend.org.uk> remarked:
>> >> On the other hand, as you can get Raspberry PIs with onboard emulators
>> >> of 1980's 8-bit home computers, the idea you need inventory of old
>> >> hardware is false.
>> >
>> >Emulating the OS is not the only problem to be solved. At least the
>> >following issues still arise: hardware interfaces, interface protocols,
>> >people who understand the old stuff (new hires aren’t going to want to base
>> >a career on obsolete technology). Retrofits of new hardware and their
>> >associated code will need to be tested and certified, especially anything
>> >safety critical. Everything is solvable with enough money and I suspect the
>> >aerospace industry works around the problem because there’s lots of old
>> >aircraft of the same type to make it worthwhile. However relatively short
>> >production runs of UK bespoke stock is a different matter.
>>
>> Whatever. I can tell you don't have a grip of the concepts I was
>> discussing. Or this time is it you who is trolling?
>
>Emulation of the software is the easy bit. Obscure hardware is troublesome.
>How do you propose to connect some piece of hardware with a SCSI port, GPIB
>or a BBC Micro 1MHz bus connection to your emulated system?

With a relevant converter which should be virtually an off-the-shelf
item.
--
Roland Perry

Roland Perry <roland@perry.uk> wrote:
> In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>> Roland Perry <roland@perry.uk> wrote:
>>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>
>>>> I’m wondering if all the electronics inside modern rolling stock is going
>>>> to be the ultimate limiter of lifetime. There was a recent article in The
>>>> Register about DB advertising for a Windows 3.11 administrator to keep some
>>>> in train display stuff going. There’s going to have to be a very
>>>> big spares
>>>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>>>> to maintain the knowledge of how it all works. All the things that look
>>>> terribly modern now are going to be very old fashioned in a few decades
>>>> time.
>>>
>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>> hardware is false.
>>
>> Emulating the OS is not the only problem to be solved. At least the
>> following issues still arise: hardware interfaces, interface protocols,
>> people who understand the old stuff (new hires aren’t going to want to base
>> a career on obsolete technology). Retrofits of new hardware and their
>> associated code will need to be tested and certified, especially anything
>> safety critical. Everything is solvable with enough money and I suspect the
>> aerospace industry works around the problem because there’s lots of old
>> aircraft of the same type to make it worthwhile. However relatively short
>> production runs of UK bespoke stock is a different matter.
>
> Whatever. I can tell you don't have a grip of the concepts I was
> discussing. Or this time is it you who is trolling?

I’m sorry but I don’t understand the point you are trying to make.

Roland Perry <roland@perry.uk> wrote:
> In message <jGr*CHVBz@news.chiark.greenend.org.uk>, at 18:22:41 on Thu,
> 1 Feb 2024, Theo <theom+news@chiark.greenend.org.uk> remarked:
>>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>>> hardware is false.
>>>>
>>>> Emulating the OS is not the only problem to be solved. At least the
>>>> following issues still arise: hardware interfaces, interface protocols,
>>>> people who understand the old stuff (new hires aren’t going to want to base
>>>> a career on obsolete technology). Retrofits of new hardware and their
>>>> associated code will need to be tested and certified, especially anything
>>>> safety critical. Everything is solvable with enough money and I suspect the
>>>> aerospace industry works around the problem because there’s lots of old
>>>> aircraft of the same type to make it worthwhile. However relatively short
>>>> production runs of UK bespoke stock is a different matter.
>>>
>>> Whatever. I can tell you don't have a grip of the concepts I was
>>> discussing. Or this time is it you who is trolling?
>>
>> Emulation of the software is the easy bit. Obscure hardware is troublesome.
>> How do you propose to connect some piece of hardware with a SCSI port, GPIB
>> or a BBC Micro 1MHz bus connection to your emulated system?
>
> With a relevant converter which should be virtually an off-the-shelf
> item.

The converter needs to be tested, certified etc. Something that does the
job in a domestic environment is very different to working as intended, and
provably so, in a safety critical environment is very different. Even
defining as intended is hard once the original designers are long gone.

As an example, not railway related admittedly, say you have some kit
dependent on Token Ring, how do you keep that running for 40 years?

In message <upgp27$26f40$1@dont-email.me>, at 18:48:39 on Thu, 1 Feb
2024, Tweed <usenet.tweed@gmail.com> remarked:
>Roland Perry <roland@perry.uk> wrote:
>> In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>> Roland Perry <roland@perry.uk> wrote:
>>>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>>
>>>>> I’m wondering if all the electronics inside modern rolling stock
>>>>>is going to be the ultimate limiter of lifetime. There was a
>>>>>recent article in The Register about DB advertising for a Windows
>>>>>3.11 administrator to keep some in train display stuff going.
>>>>>There’s going to have to be a very big spares inventory to keep
>>>>>things going over a 30 to 40 year lifetime, plus the need to
>>>>>maintain the knowledge of how it all works. All the things that
>>>>>look terribly modern now are going to be very old fashioned in a few decades time.
>>>>
>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>> hardware is false.
>>>
>>> Emulating the OS is not the only problem to be solved. At least the
>>>following issues still arise: hardware interfaces, interface
>>>protocols, people who understand the old stuff (new hires aren’t
>>>going to want to base a career on obsolete technology). Retrofits of
>>>new hardware and their associated code will need to be tested and
>>>certified, especially anything safety critical. Everything is
>>>solvable with enough money and I suspect the aerospace industry
>>>works around the problem because there’s lots of old aircraft of
>>>the same type to make it worthwhile. However relatively short
>>>production runs of UK bespoke stock is a different matter.
>>
>> Whatever. I can tell you don't have a grip of the concepts I was
>> discussing. Or this time is it you who is trolling?
>
>I’m sorry but I don’t understand the point you are trying to make.

That emulators and port converters will be capable of solving a lot of
the problems you have raised.
--
Roland Perry

In message <upgpgh$26hu8$1@dont-email.me>, at 18:56:17 on Thu, 1 Feb
2024, Tweed <usenet.tweed@gmail.com> remarked:
>Roland Perry <roland@perry.uk> wrote:
>> In message <jGr*CHVBz@news.chiark.greenend.org.uk>, at 18:22:41 on Thu,
>> 1 Feb 2024, Theo <theom+news@chiark.greenend.org.uk> remarked:
>>>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>>>> hardware is false.
>>>>>
>>>>> Emulating the OS is not the only problem to be solved. At least the
>>>>> following issues still arise: hardware interfaces, interface protocols,
>>>>> people who understand the old stuff (new hires aren’t going to
>>>>>want to base
>>>>> a career on obsolete technology). Retrofits of new hardware and their
>>>>> associated code will need to be tested and certified, especially anything
>>>>> safety critical. Everything is solvable with enough money and I
>>>>>suspect the
>>>>> aerospace industry works around the problem because there’s lots of old
>>>>> aircraft of the same type to make it worthwhile. However relatively short
>>>>> production runs of UK bespoke stock is a different matter.
>>>>
>>>> Whatever. I can tell you don't have a grip of the concepts I was
>>>> discussing. Or this time is it you who is trolling?
>>>
>>> Emulation of the software is the easy bit. Obscure hardware is troublesome.
>>> How do you propose to connect some piece of hardware with a SCSI port, GPIB
>>> or a BBC Micro 1MHz bus connection to your emulated system?
>>
>> With a relevant converter which should be virtually an off-the-shelf
>> item.
>
>The converter needs to be tested, certified etc. Something that does the
>job in a domestic environment

Who mentioned emulators (and port converters) only working in a domestic
environment?

>is very different to working as intended, and
>provably so, in a safety critical environment is very different. Even
>defining as intended is hard once the original designers are long gone.
>
>As an example, not railway related admittedly, say you have some kit
>dependent on Token Ring, how do you keep that running for 40 years?

By building a relevant emulator/port_converter combination.

Token Ring isn't rocket science, and one of my colleagues built
thick-Ethernet S100 boards by hand out of discrete components in the
late 70's.
--
Roland Perry

Roland Perry <roland@perry.uk> wrote:
> In message <upgp27$26f40$1@dont-email.me>, at 18:48:39 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>> Roland Perry <roland@perry.uk> wrote:
>>> In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>> Roland Perry <roland@perry.uk> wrote:
>>>>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>>>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>>>
>>>>>> I’m wondering if all the electronics inside modern rolling stock
>>>>>> is going to be the ultimate limiter of lifetime. There was a
>>>>>> recent article in The Register about DB advertising for a Windows
>>>>>> 3.11 administrator to keep some in train display stuff going.
>>>>>> There’s going to have to be a very big spares inventory to keep
>>>>>> things going over a 30 to 40 year lifetime, plus the need to
>>>>>> maintain the knowledge of how it all works. All the things that
>>>>>> look terribly modern now are going to be very old fashioned in a few decades time.
>>>>>
>>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>>> hardware is false.
>>>>
>>>> Emulating the OS is not the only problem to be solved. At least the
>>>> following issues still arise: hardware interfaces, interface
>>>> protocols, people who understand the old stuff (new hires aren’t
>>>> going to want to base a career on obsolete technology). Retrofits of
>>>> new hardware and their associated code will need to be tested and
>>>> certified, especially anything safety critical. Everything is
>>>> solvable with enough money and I suspect the aerospace industry
>>>> works around the problem because there’s lots of old aircraft of
>>>> the same type to make it worthwhile. However relatively short
>>>> production runs of UK bespoke stock is a different matter.
>>>
>>> Whatever. I can tell you don't have a grip of the concepts I was
>>> discussing. Or this time is it you who is trolling?
>>
>> I’m sorry but I don’t understand the point you are trying to make.
>
> That emulators and port converters will be capable of solving a lot of
> the problems you have raised.

It’s much harder than that. Clearly possible for things like the passenger
info displays, much harder for safety critical items.

Roland Perry <roland@perry.uk> wrote:
> In message <upgpgh$26hu8$1@dont-email.me>, at 18:56:17 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>> Roland Perry <roland@perry.uk> wrote:
>>> In message <jGr*CHVBz@news.chiark.greenend.org.uk>, at 18:22:41 on Thu,
>>> 1 Feb 2024, Theo <theom+news@chiark.greenend.org.uk> remarked:
>>>>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>>>>> hardware is false.
>>>>>>
>>>>>> Emulating the OS is not the only problem to be solved. At least the
>>>>>> following issues still arise: hardware interfaces, interface protocols,
>>>>>> people who understand the old stuff (new hires aren’t going to
>>>>>> want to base
>>>>>> a career on obsolete technology). Retrofits of new hardware and their
>>>>>> associated code will need to be tested and certified, especially anything
>>>>>> safety critical. Everything is solvable with enough money and I
>>>>>> suspect the
>>>>>> aerospace industry works around the problem because there’s lots of old
>>>>>> aircraft of the same type to make it worthwhile. However relatively short
>>>>>> production runs of UK bespoke stock is a different matter.
>>>>>
>>>>> Whatever. I can tell you don't have a grip of the concepts I was
>>>>> discussing. Or this time is it you who is trolling?
>>>>
>>>> Emulation of the software is the easy bit. Obscure hardware is troublesome.
>>>> How do you propose to connect some piece of hardware with a SCSI port, GPIB
>>>> or a BBC Micro 1MHz bus connection to your emulated system?
>>>
>>> With a relevant converter which should be virtually an off-the-shelf
>>> item.
>>
>> The converter needs to be tested, certified etc. Something that does the
>> job in a domestic environment
>
> Who mentioned emulators (and port converters) only working in a domestic
> environment?
>
>> is very different to working as intended, and
>> provably so, in a safety critical environment is very different. Even
>> defining as intended is hard once the original designers are long gone.
>>
>> As an example, not railway related admittedly, say you have some kit
>> dependent on Token Ring, how do you keep that running for 40 years?
>
> By building a relevant emulator/port_converter combination.
>
> Token Ring isn't rocket science, and one of my colleagues built
> thick-Ethernet S100 boards by hand out of discrete components in the
> late 70's.

Understanding an ancient system when you are an engineer born after when
the system was designed (it came home to me the other day when a colleague
stated she was born at the time I started work at my current employer),
finding someone who even wants to take on the task of electronics
archeology, finding the original software/firmware specifications, checking
it all works to specifications, and then checking it works to the
undocumented reality and then getting it certified, are all expensive and
time consuming. I’d wager a lot of stock will go to the scrapper long
before the mechanical stuff is life expired because of obsolete
electronics.

On 01.02.2024 17:48, Roland Perry wrote:
> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>
>> I’m wondering if all the electronics inside modern rolling stock is going
>> to be the ultimate limiter of lifetime. There was a recent article in The
>> Register about DB advertising for a Windows 3.11 administrator to keep
>> some
>> in train display stuff going. There’s going to have to be a very big
>> spares
>> inventory to keep things going over a 30 to 40 year lifetime, plus the
>> need
>> to maintain the knowledge of how it all works. All the things that look
>> terribly modern now are going to be very old fashioned in a few decades
>> time.
>
> On the other hand, as you can get Raspberry PIs with onboard emulators
> of 1980's 8-bit home computers, the idea you need inventory of old
> hardware is false.

The ususal issue where this sort of problem arises is not just a matter
of running old software, though. The problem often is a result of some
custom hardware element such a custom interface card, that connected via
some long-abandoned connector type, for which drivers are not available
for anythign approaching a modern operating system, even if you could
get a motherboard with the relevant physical connector on it. It's no
use emulating some 8 bit or 16 bit OS on modern hardware if you can't
get that to run on anything that can physically connect to the relevant
inerface card (and run the connection appropriately, which may not be
the case as these sort of things sometimes were built around quirks that
were not offically part of the spec but were the result of the
hardware/software of the day).

A person I knew some time back was involved in the project to re-create
Prof Stephen Hawking's voice. When the Prof first needed a voice
synthesiser, he got what was state of the art at the time, but over time
the technology of voice synthesis moved on, and both the hardware and
software used for that sort of purpose were unrecognisably changed. When
the hardware that Hawking used began to become unreliable due to age,
this posed a problem. While the voice he spoke with was not exactly
natural sounding or all that much like his natural voice before his
illness, over the years it had become "his" voice. To paraphrase Lady
Bracknell, to lose one voice, Prof Hawkign, may be regarded as a
misfortune; to lose two looks like carelessness.

Long story short, they tracked down a retired person who had worked at
the company that developed the original synthesiser, and managed to put
together a machine with modern hardware and software that, after a few
iterations with the professor, was accepted as a satisfactory match. In
the event, the professor did not live much longer, and I do not know
whether he actually switched to the new solution.

Robin

On 01.02.2024 20:02, Roland Perry wrote:
> In message <upgp27$26f40$1@dont-email.me>, at 18:48:39 on Thu, 1 Feb
> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>> Roland Perry <roland@perry.uk> wrote:
>>> In message <upgj2n$25d3i$1@dont-email.me>, at 17:06:31 on Thu, 1 Feb
>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>> Roland Perry <roland@perry.uk> wrote:
>>>>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>>>>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>>>>
>>>>>> I’m wondering if all the electronics inside modern rolling stock
>>>>>> is going  to be the ultimate limiter of lifetime. There was a
>>>>>> recent article in The  Register about DB advertising for a Windows
>>>>>> 3.11 administrator to keep some  in train display stuff going.
>>>>>> There’s going to have to be a very  big spares  inventory to keep
>>>>>> things going over a 30 to 40 year lifetime, plus the need  to
>>>>>> maintain the knowledge of how it all works. All the things that
>>>>>> look  terribly modern now are going to be very old fashioned in a
>>>>>> few decades  time.
>>>>>
>>>>> On the other hand, as you can get Raspberry PIs with onboard emulators
>>>>> of 1980's 8-bit home computers, the idea you need inventory of old
>>>>> hardware is false.
>>>>
>>>> Emulating the OS is not the only problem to be solved. At least the
>>>> following issues still arise: hardware interfaces, interface
>>>> protocols,  people who understand the old stuff (new hires aren’t
>>>> going to want to base  a career on obsolete technology). Retrofits
>>>> of new hardware and their  associated code will need to be tested
>>>> and certified, especially anything  safety critical. Everything is
>>>> solvable with enough money and I suspect the  aerospace industry
>>>> works around the problem because there’s lots of old  aircraft of
>>>> the same type to make it worthwhile. However relatively short
>>>> production runs of UK bespoke stock is a different matter.
>>>
>>> Whatever. I can tell you don't have a grip of the concepts I was
>>> discussing. Or this time is it you who is trolling?
>>
>> I’m sorry but I don’t understand the point you are trying to make.
>
> That emulators and port converters will be capable of solving a lot of
> the problems you have raised.

There is a tradeoff inherent in this sort of thing. On the one hand you
can just try to recruit some old guy nearing retirement who has the
knolwedge to keep the old kit running. On the other hand, you can try to
build a new system that replicates the functionality on modern hardware.
THere is the added complication with the second option that the actual
software involved will be owned by the OEM supplier, not the end user
(in this case, Siemens, not DB), and they may decide they have no
interest in helping you out with this project, in which case it baloons
into not only emmulating and testing, but potentially also
reverse-engineering. In this situation, I can see why DB might feel that
the lower cost route is to find some old guy who still remembers how to
keep the old stuff running until the trains it is used in are retired.

Robin

Roland Perry <roland@perry.uk> wrote:
> In message <jGr*CHVBz@news.chiark.greenend.org.uk>, at 18:22:41 on Thu,
> 1 Feb 2024, Theo <theom+news@chiark.greenend.org.uk> remarked:
> >
> >Emulation of the software is the easy bit. Obscure hardware is troublesome.
> >How do you propose to connect some piece of hardware with a SCSI port, GPIB
> >or a BBC Micro 1MHz bus connection to your emulated system?
>
> With a relevant converter which should be virtually an off-the-shelf
> item.

So could you please show us converters for those three protocols which will
speak to an emulated machine and would be suitable for an industrial setup?

There are USB-GPIB converters but I'm not aware of any of them hooking into
emulators. SCSI to USB adapters haven't been made for 20 years; there are
some Raspberry Pi projects for pretending to be a SCSI drive to talk to an
Amiga or classic Mac, but not the other way around (interfacing some random
SCSI device to a USB PC). You can use a SCSI PCI card in a passthrough
setup to a VM, but you probably won't find DOS drivers for it. You can't
easily passthrough an ISA card to a VM. And there is no off the shelf
hardware for the 1MHz bus - best you can do is a card for an Archimedes
which hasn't been made for 35 years.

Theo

Bob <bob@domain.com> wrote:
> There is a tradeoff inherent in this sort of thing. On the one hand you
> can just try to recruit some old guy nearing retirement who has the
> knolwedge to keep the old kit running. On the other hand, you can try to
> build a new system that replicates the functionality on modern hardware.
> THere is the added complication with the second option that the actual
> software involved will be owned by the OEM supplier, not the end user
> (in this case, Siemens, not DB), and they may decide they have no
> interest in helping you out with this project, in which case it baloons
> into not only emmulating and testing, but potentially also
> reverse-engineering. In this situation, I can see why DB might feel that
> the lower cost route is to find some old guy who still remembers how to
> keep the old stuff running until the trains it is used in are retired.

It's quite possibly harder than that. Siemens made the train, but they
could have bought in the software from some third party company. Their
forte is not writing software for controlling train displays, so they maybe
bought in a system from XYZ Displays GmbH, which was perhaps a small company
whose founder is long dead. Siemens maybe have a floppy disc with the
software for generating new screens, but they don't have the source code for
it, nor any documentation of the protocols.

Often such software is really tied to the hardware of the machine it runs
on. For example we have an industrial CNC machine with a serial interface.
The serial interface doesn't take commands as you might expect, it takes
bytes which directly drive the motor controllers in the machine: 8 bits
control 4 motors (1 bit for direction, 1 bit for step). So to make the
motor spin forwards at a particular speed you send hex chars on the serial
port:
(suppose bit 0 = direction, bit 1 = motor step)

0x3 0x1 0x3 0x1 0x3 0x1

and to make it spin slower you step the motor slower by changing the timing
of the bytes:

0x3 .... 0x1 .... 0x3 .... 0x1 .... 0x3 .... 0x1

The software runs on Windows XP but no newer, because it installs a driver
to have direct control of the serial port rather than going through the
Windows driver stack. There is no chance of running this on a modern
machine, or using a USB serial port, because it's so timing critical. If
the timing goes wrong your smooth CNC tool paths turn out all juddery
because of the serial port latency and you mess up your workpiece.

These are the kind of challenges that someone working with this kind of kit
faces. In the case of the CNC, the solution is to throw away the controller
and its software stack and refit a new controller to control the stepper
motors (for which there is, thankfully, a modern ecosystem of software to do
that). But that option is not available to a more niche system, or one
which is safety critical and subject to certification requirements.

Theo

Interesting to note the longevity of the humble calculator though. I've
got a Casio FX82a sitting on my desk here which still seems to work fine
and does everything I need from a basic scientific calculator. I really
cannot remember when I bought it, but if someone told me that it was 25
or 30 years ago (if not longer) then that's very plausible and would not
surprise me at all. Probably outlasted 6 or more generations of PC.

--
Order alone is boring; complexity alone is chaos

On 2024-02-02 21:14, JohnD wrote:
> Interesting to note the longevity of the humble calculator though. I've
> got a Casio FX82a sitting on my desk here which still seems to work fine
> and does everything I need from a basic scientific calculator. I really
> cannot remember when I bought it, but if someone told me that it was 25
> or 30 years ago (if not longer) then that's very plausible and would not
> surprise me at all. Probably outlasted 6 or more generations of PC.
>
I'm still using a Sharp EL-230 that's about 40 years old, and I have a
working Sinclair Scientific Programmable from 1977, though the legends
on the keys are getting a bit faint!

nib

Tweed <usenet.tweed@gmail.com> wrote:
> Roland Perry <roland@perry.uk> wrote:
>> In message <upg48j$22qt3$1@dont-email.me>, at 12:53:39 on Thu, 1 Feb
>> 2024, Tweed <usenet.tweed@gmail.com> remarked:
>>
>>> I’m wondering if all the electronics inside modern rolling stock is going
>>> to be the ultimate limiter of lifetime. There was a recent article in The
>>> Register about DB advertising for a Windows 3.11 administrator to keep some
>>> in train display stuff going. There’s going to have to be a very big spares
>>> inventory to keep things going over a 30 to 40 year lifetime, plus the need
>>> to maintain the knowledge of how it all works. All the things that look
>>> terribly modern now are going to be very old fashioned in a few decades
>>> time.
>>
>> On the other hand, as you can get Raspberry PIs with onboard emulators
>> of 1980's 8-bit home computers, the idea you need inventory of old
>> hardware is false.
>
> Emulating the OS is not the only problem to be solved. At least the
> following issues still arise: hardware interfaces, interface protocols,
> people who understand the old stuff (new hires aren’t going to want to base
> a career on obsolete technology). Retrofits of new hardware and their
> associated code will need to be tested and certified, especially anything
> safety critical. Everything is solvable with enough money and I suspect the
> aerospace industry works around the problem because there’s lots of old
> aircraft of the same type to make it worthwhile. However relatively short
> production runs of UK bespoke stock is a different matter.

Even emulating the OS may not be trivial. There used to be a YouTube video
of, I think, a lecture by a guy who emulated a BBC Micro. The fact that
the video was ~1h long shows you that it wasn’t an easy job. IIRC there
was stuff in the OS which relied on how long particular tasks took on
particular bits of the hardware.

Sam

--
The entity formerly known as Sam.Wilson@ed.ac.uk
Spit the dummy to reply

On 02.02.2024 22:42, nibble wrote:
> On 2024-02-02 21:14, JohnD wrote:
>> Interesting to note the longevity of the humble calculator though.
>> I've got a Casio FX82a sitting on my desk here which still seems to
>> work fine and does everything I need from a basic scientific
>> calculator. I really cannot remember when I bought it, but if someone
>> told me that it was 25 or 30 years ago (if not longer) then that's
>> very plausible and would not surprise me at all. Probably outlasted 6
>> or more generations of PC.
>>
>  I'm still using a Sharp EL-230 that's about 40 years old, and I have a
> working Sinclair Scientific Programmable from 1977, though the legends
> on the keys are getting a bit faint!

When I started as an undergrad, a fair few years ago now, when I first
visted the engineering department, the university policy was that a
select set of calculators were approved for use in university exams, and
if you had one of your own, you could submit it to be marked with a
university approved sticker, or you could buy an approved model,
pre-marked. I chose the latter, and I still use it from time to time in
my day job of building big industrial gas turbines.

There is a long drawn out story about how I relished in putting a
graduate student in his place over those rules, but perhaps it is too
long winded just now.

Robin