On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on with it...

I believe they eventually got rid of him!

Adrian

SteveW

John Williamson wrote:
tony sayer wrote:
In article , John Williamson johnwilli
scribeth thus
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
I normally end up doing a clean reinstall every year or so. An
evening getting it to the stage of being safe on the internet,
update Windows overnight, which usually takes a couple of
evenings while I watch TV, then re-install the programs one by
one as I use them from the install images on the HD, or dig out
the CDs from the piling system.

Why do you do that John?. Which version of Win are you running as we
have a good mix of WIN 200 Pro and XP machines around and all in
there're running fine, some been on the go 6 odd or more years!...
XP, which I find slows down over time. A fresh install every year
or so gets the performance back where it should be, though that may
be partly psychological. I also find that it gets less stable over
time, and, no, I don't visit dodgy download sites, though I do tend
to push the limits with what I use the computers for, which means I
notice any problems early on.

Well I find that it all seems fine, just runs does what we need and
haven't seen any lock up's for a very long time now..

And that's on all the machines, let alone the WIN 7 engined ones the
only WIN that we don't give house room to is Vista!..

So if it slows what's the reason ?.. Why?...
Dunno, I just know that it does, and slowly gets more and more
unstable with time.


Wonder -why- you are finding that?..

Ask Microsoft and Intel. I don't know. I do know that a re-install gets
it back to the original speed and stability.


browser logs everythng into fragged cache files

The only machine that wll run 7 is the only one that doesn't seem to
suffer, but that's the only one that's got a lot of reserve capacity.

Yes but thats a differing operating system?.. Are you saying thats got
more storage capacity or processing power ?..

XP runs with a degree of success on oldish machines with 256 Meg of
memory, and various processors.

7 won't even install until the machine has a 1GHz processor and a Gig of
RAM, and 15Gig spare on the HD. This EEE (4Gig C: drive, and 2Gig of
RAM) will run XP, the twin core 1.6GHz machine will run 7 almost as fast
as the EEE runs XP. Which is why all my machines run XP, and will do
until M$ withdraw support, at which point, given the way 7 has had stuff
moved round, I may well go for Linux.

I've just found an old Toshiba Satellite 1800, which I'll probably end
up installing ME on, as I need a particular item of hardware supported,
and 98 won't run the drivers. At the moment, it's trying to install
Debian, so I can see if that will do the required job.

it probably will. I have debian so ask if you run into crap

Steve Walker wrote:
On 12/05/2011 12:01, Nitro® wrote:

"Invisible Man" wrote in message
news:iqevb9$brb$2@dont-

Take care with CCleaner


Why? Having passwords and personal information held by
a browser is foolhardy at best.

Personally I don't have any problem with my browser remembering
passwords and ids for forums, newspapers, etc. that I visit. It's far
easier letting it just get on with it and saves me any effort. If it
ends up compromised, there's no great loss.

Passwords for banking, inland revenue, etc. are however, a different
matter and I certainly wouldn't let the browser store those.

Firefox wont store anything from an https: site.

SteveW

Adrian Brentnall wrote:
On 13/05/2011 14:20, The Natural Philosopher wrote:
Adrian Brentnall wrote:
On 13/05/2011 10:58, Adrian C wrote:
On 13/05/2011 10:41, Adrian Brentnall wrote:
HI All

Thanks for the comments so far....

I have looked in the Event Logs and seem to be getting a fair few
warnings (not errors) "error51 on \Harddisk1\D during a paging
operation"


Following may help to decode the data section recorded in the event log
http://support.microsoft.com/kb/244780

I've run Dell's built-in diagnostics and they come up clean (both
quick
and extended) - also downloaded Hitachi's own disk test suite and that
says 'OK'....

Do a check with HDtune and those SMART registers. Some manufacturer HDD
testing tools are designed to keep failing customer drives away from
their RMA department. When the drive starts reallocating sectors, is
the
time that I tend to ditch it - however the top line of a diagnostic
could still say 'PASS' in that senario.

Interesting.... g Thanks



I'm wondering if it could be an obscure memory fault - could replace
all
the system memory for 50 euro or so - might this be a Good Plan ?

Problem with that is 1) memory itself rarely fails unless if badly
overclocked or over volted and 2) What if it turns out to be the
motherboard?

Apply the business case. Sling it.

Hmmm.
I've just specced out the nearest equivalent Dell machine - and it's
best part of ‚¬340 + vat + delivery.
The bigger cost to me is that the box would come with Windows7, and
various bits of kit hanging off this PC would probably not have
drivers available, plus there's the grief of reinstalling software &
so on...

I think my business case is best served if I can get this machine
running again g

Checking up on the memory, it seems that I've got a mix of memory
modules installed - 2 x DDR2 PC2-5300, 2 x DDR2 PC2-6400, DDR2
(non-ECC) - so whether that's significant or not ...??

Also contacted my friendly local freelance Dell fixit-man - who's
going to call by with a bag of bits, an AVO, and a prayer-mat some
time next week - we'll see if he can work some magic for me!


So don't buy Dell.
Buy from a pc maker who will sell you just the hardware, if you have
full install disks available for generic machines.

Don't know about that - I do have the Dell disks, plus some other WinXP
disks that seem to install OK on one or two other 'home-built' machines

What actually do you need out of this machine anyway?

Nothing earth-shattering g

General w/p / openoffice activities,

Open office of the libra fork

graphics stuff (Paint-Shop Pro),

Hmm. Gimp may do on Liniux

web authoring (Netobjects Fusion & WYSIWYG plus anything else that takes
my fancy).

I use a text editor :-)

Email & web. Very infrequently editing music files (Goldwave).
Mozilla comes as standard on most linux..
Audacity is not bad for editing sound files
Runs a set of software to drive a vinyl cutter.
ah that's probly win only...


Output for two screens..... that's about it


Get half a dozen virtual screens..not sure about twin reals on X windows.


Oh - and not falling in a heap would be a bonus! g


That comes as standard with linux.
Id say linux plus xp in a virtual for the vinyl cutter. Can be restored
at the touch of a button if windows goes tits up. Juts keep data on a
LINUX drive.


Regards
Adrian

The Natural Philosopher wrote:
John Williamson wrote:
tony sayer wrote:
In article , John Williamson johnwilli
scribeth thus
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
I normally end up doing a clean reinstall every year or so. An
evening getting it to the stage of being safe on the internet,
update Windows overnight, which usually takes a couple of
evenings while I watch TV, then re-install the programs one by
one as I use them from the install images on the HD, or dig out
the CDs from the piling system.

Why do you do that John?. Which version of Win are you running as we
have a good mix of WIN 200 Pro and XP machines around and all in
there're running fine, some been on the go 6 odd or more years!...
XP, which I find slows down over time. A fresh install every year
or so gets the performance back where it should be, though that
may be partly psychological. I also find that it gets less stable
over time, and, no, I don't visit dodgy download sites, though I
do tend to push the limits with what I use the computers for,
which means I notice any problems early on.

Well I find that it all seems fine, just runs does what we need and
haven't seen any lock up's for a very long time now..

And that's on all the machines, let alone the WIN 7 engined ones the
only WIN that we don't give house room to is Vista!..

So if it slows what's the reason ?.. Why?...
Dunno, I just know that it does, and slowly gets more and more
unstable with time.


Wonder -why- you are finding that?..

Ask Microsoft and Intel. I don't know. I do know that a re-install
gets it back to the original speed and stability.


browser logs everythng into fragged cache files

The only machine that wll run 7 is the only one that doesn't seem to
suffer, but that's the only one that's got a lot of reserve capacity.

Yes but thats a differing operating system?.. Are you saying thats got
more storage capacity or processing power ?..

XP runs with a degree of success on oldish machines with 256 Meg of
memory, and various processors.

7 won't even install until the machine has a 1GHz processor and a Gig
of RAM, and 15Gig spare on the HD. This EEE (4Gig C: drive, and 2Gig
of RAM) will run XP, the twin core 1.6GHz machine will run 7 almost as
fast as the EEE runs XP. Which is why all my machines run XP, and will
do until M$ withdraw support, at which point, given the way 7 has had
stuff moved round, I may well go for Linux.

I've just found an old Toshiba Satellite 1800, which I'll probably end
up installing ME on, as I need a particular item of hardware
supported, and 98 won't run the drivers. At the moment, it's trying to
install Debian, so I can see if that will do the required job.

it probably will. I have debian so ask if you run into crap

It won't use the built-in display at anything better than 800x600.
Apparently, according to t'internet it's a minor change to the xorg.conf
file, setting the video driver to "trident" instead of "vesa", and
setting the refresh to 60Hz. There are even sample files on varius
sites. That I can do, either using mc or Gedit. It would help if the
xorg.conf file was in /etc or /usr/bin/X11 or even findable by root.

Debian 6 using Gnome, clean HD install.

When that's sorted, the fun begins with the sound configuration, as I
want to use it as a multitrack recorder. People claim that the
interfaces I want to use "just work".

--
Tciao for Now!

John.

"Adrian Brentnall" wrote in message
...
On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some Texas
99000 kit that was to be used in a railway points switching application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2' -
and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

I wrote a test for the 8086 while I was at GEC.
It didn't really attempt to test instructions at all.
It moved data about to test the internal and external buses for stuck bits
and shorted bits.
It also tested the address busses in a similar way.
It also tested the functions like increment but not every value just the
boundaries.

It made no real attempt to report errors it just set some bits in a bit of
IO hardware at the start of each test and went into a loop if it failed.

John Williamson wrote:
The Natural Philosopher wrote:
John Williamson wrote:
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
I normally end up doing a clean reinstall every year or so. An
evening getting it to the stage of being safe on the internet,
update Windows overnight, which usually takes a couple of
evenings while I watch TV, then re-install the programs one by
one as I use them from the install images on the HD, or dig out
the CDs from the piling system.

Why do you do that John?. Which version of Win are you running
as we
have a good mix of WIN 200 Pro and XP machines around and all in
there're running fine, some been on the go 6 odd or more years!...
XP, which I find slows down over time. A fresh install every year
or so gets the performance back where it should be, though that
may be partly psychological. I also find that it gets less stable
over time, and, no, I don't visit dodgy download sites, though I
do tend to push the limits with what I use the computers for,
which means I notice any problems early on.

Well I find that it all seems fine, just runs does what we need and
haven't seen any lock up's for a very long time now..

And that's on all the machines, let alone the WIN 7 engined ones the
only WIN that we don't give house room to is Vista!..

So if it slows what's the reason ?.. Why?...
Dunno, I just know that it does, and slowly gets more and more
unstable with time.


Wonder -why- you are finding that?..

Ask Microsoft and Intel. I don't know. I do know that a re-install
gets it back to the original speed and stability.


browser logs everythng into fragged cache files

The only machine that wll run 7 is the only one that doesn't seem
to suffer, but that's the only one that's got a lot of reserve
capacity.

Yes but thats a differing operating system?.. Are you saying thats got
more storage capacity or processing power ?..

XP runs with a degree of success on oldish machines with 256 Meg of
memory, and various processors.

7 won't even install until the machine has a 1GHz processor and a Gig
of RAM, and 15Gig spare on the HD. This EEE (4Gig C: drive, and 2Gig
of RAM) will run XP, the twin core 1.6GHz machine will run 7 almost
as fast as the EEE runs XP. Which is why all my machines run XP, and
will do until M$ withdraw support, at which point, given the way 7
has had stuff moved round, I may well go for Linux.

I've just found an old Toshiba Satellite 1800, which I'll probably
end up installing ME on, as I need a particular item of hardware
supported, and 98 won't run the drivers. At the moment, it's trying
to install Debian, so I can see if that will do the required job.

it probably will. I have debian so ask if you run into crap

It won't use the built-in display at anything better than 800x600.
Apparently, according to t'internet it's a minor change to the xorg.conf
file, setting the video driver to "trident" instead of "vesa", and
setting the refresh to 60Hz. There are even sample files on varius
sites. That I can do, either using mc or Gedit. It would help if the
xorg.conf file was in /etc or /usr/bin/X11 or even findable by root.


There may not be one if its the latest debian
Much is now automatically adapted.

I think you may need to make one in /etc/X11..IIRC.
That's where it is on lenny..you may have squeeze tho.



Debian 6 using Gnome, clean HD install.

When that's sorted, the fun begins with the sound configuration, as I
want to use it as a multitrack recorder. People claim that the
interfaces I want to use "just work".


sound is a bit of a doss up on 5. Mostly sorted on 6.

ALSA is the better way to go Id say

The Natural Philosopher wrote:
John Williamson wrote:
The Natural Philosopher wrote:
John Williamson wrote:
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
I normally end up doing a clean reinstall every year or so. An
evening getting it to the stage of being safe on the internet,
update Windows overnight, which usually takes a couple of
evenings while I watch TV, then re-install the programs one by
one as I use them from the install images on the HD, or dig
out the CDs from the piling system.

Why do you do that John?. Which version of Win are you running
as we
have a good mix of WIN 200 Pro and XP machines around and all in
there're running fine, some been on the go 6 odd or more years!...
XP, which I find slows down over time. A fresh install every
year or so gets the performance back where it should be, though
that may be partly psychological. I also find that it gets less
stable over time, and, no, I don't visit dodgy download sites,
though I do tend to push the limits with what I use the
computers for, which means I notice any problems early on.

Well I find that it all seems fine, just runs does what we need and
haven't seen any lock up's for a very long time now..

And that's on all the machines, let alone the WIN 7 engined ones the
only WIN that we don't give house room to is Vista!..

So if it slows what's the reason ?.. Why?...
Dunno, I just know that it does, and slowly gets more and more
unstable with time.


Wonder -why- you are finding that?..

Ask Microsoft and Intel. I don't know. I do know that a re-install
gets it back to the original speed and stability.


browser logs everythng into fragged cache files

The only machine that wll run 7 is the only one that doesn't seem
to suffer, but that's the only one that's got a lot of reserve
capacity.

Yes but thats a differing operating system?.. Are you saying thats got
more storage capacity or processing power ?..

XP runs with a degree of success on oldish machines with 256 Meg of
memory, and various processors.

7 won't even install until the machine has a 1GHz processor and a
Gig of RAM, and 15Gig spare on the HD. This EEE (4Gig C: drive, and
2Gig of RAM) will run XP, the twin core 1.6GHz machine will run 7
almost as fast as the EEE runs XP. Which is why all my machines run
XP, and will do until M$ withdraw support, at which point, given the
way 7 has had stuff moved round, I may well go for Linux.

I've just found an old Toshiba Satellite 1800, which I'll probably
end up installing ME on, as I need a particular item of hardware
supported, and 98 won't run the drivers. At the moment, it's trying
to install Debian, so I can see if that will do the required job.

it probably will. I have debian so ask if you run into crap

It won't use the built-in display at anything better than 800x600.
Apparently, according to t'internet it's a minor change to the
xorg.conf file, setting the video driver to "trident" instead of
"vesa", and setting the refresh to 60Hz. There are even sample files
on varius sites. That I can do, either using mc or Gedit. It would
help if the xorg.conf file was in /etc or /usr/bin/X11 or even
findable by root.


There may not be one if its the latest debian
Much is now automatically adapted.

I think you may need to make one in /etc/X11..IIRC.
That's where it is on lenny..you may have squeeze tho.

I do. /etc/X11 is where I started looking, but the nly copies I culdfind
were commented out copies in a sample directory. I'll have another go
next time I'mnear the machine.


Debian 6 using Gnome, clean HD install.

When that's sorted, the fun begins with the sound configuration, as I
want to use it as a multitrack recorder. People claim that the
interfaces I want to use "just work".


sound is a bit of a doss up on 5. Mostly sorted on 6.

ALSA is the better way to go Id say

ALSA is apparently installed. I'd prefer Ubuntu Studio, but the DVD
drive's a bit marginal, and has trouble with recorded discs, and the
download version I tried screwed up installing GRUB. Maybe another go at
that, as the problem wasn't with the boot CD.

--
Tciao for Now!

John.

In article , The Natural Philosopher
scribeth thus
John Williamson wrote:
tony sayer wrote:
In article , John Williamson johnwilli
scribeth thus
tony sayer wrote:
In article , John Williamson
johnwilli
scribeth thus
tony sayer wrote:
I normally end up doing a clean reinstall every year or so. An
evening getting it to the stage of being safe on the internet,
update Windows overnight, which usually takes a couple of
evenings while I watch TV, then re-install the programs one by
one as I use them from the install images on the HD, or dig out
the CDs from the piling system.

Why do you do that John?. Which version of Win are you running as we
have a good mix of WIN 200 Pro and XP machines around and all in
there're running fine, some been on the go 6 odd or more years!...
XP, which I find slows down over time. A fresh install every year
or so gets the performance back where it should be, though that may
be partly psychological. I also find that it gets less stable over
time, and, no, I don't visit dodgy download sites, though I do tend
to push the limits with what I use the computers for, which means I
notice any problems early on.

Well I find that it all seems fine, just runs does what we need and
haven't seen any lock up's for a very long time now..

And that's on all the machines, let alone the WIN 7 engined ones the
only WIN that we don't give house room to is Vista!..

So if it slows what's the reason ?.. Why?...
Dunno, I just know that it does, and slowly gets more and more
unstable with time.


Wonder -why- you are finding that?..

Ask Microsoft and Intel. I don't know. I do know that a re-install gets
it back to the original speed and stability.


browser logs everythng into fragged cache files

So is that it?. Surely something that can be put right without
reinstalling?...




The only machine that wll run 7 is the only one that doesn't seem to
suffer, but that's the only one that's got a lot of reserve capacity.

Yes but thats a differing operating system?.. Are you saying thats got
more storage capacity or processing power ?..

XP runs with a degree of success on oldish machines with 256 Meg of
memory, and various processors.

7 won't even install until the machine has a 1GHz processor and a Gig of
RAM, and 15Gig spare on the HD. This EEE (4Gig C: drive, and 2Gig of
RAM) will run XP, the twin core 1.6GHz machine will run 7 almost as fast
as the EEE runs XP. Which is why all my machines run XP, and will do
until M$ withdraw support, at which point, given the way 7 has had stuff
moved round, I may well go for Linux.

I've just found an old Toshiba Satellite 1800, which I'll probably end
up installing ME

ME!, Arrgh!!!, even Vista is better than that pile of cack.

Can't you find a copy of WIN 200 Pro or XP even?..

on, as I need a particular item of hardware supported,
and 98 won't run the drivers. At the moment, it's trying to install
Debian, so I can see if that will do the required job.

it probably will. I have debian so ask if you run into crap

--
Tony Sayer

"Adrian C" wrote in message
...
On 12/05/2011 12:41, Mike Tomlinson wrote:
In , Invisible Man
writes

This is way OT for uk.d-i-y, uk.comp.homebuilt would be a better choice.

Whenever I have reinstalled windows I have had to reinstall everything
else afterwards. If there is an easy way I Would love to hear it.

Do a repair install, preserves all your installed apps and settings.


Yup if done correctly, and the right option is available.

http://www.michaelstevenstech.com/XPrepairinstall.htm

There is so much confusion between different builds and releases of XP
(Professional, Home, SP1,2,3, Retail, OEM, Brand custom) that the unaware
can start the process off and be next looking at either a completely
virgin copy of XP (with all data and applications gone), or something that
is almost there for the data but all the applications have lost their
registry marbles and no longer work.

If the right option isn't there, see my post above in the thread for a way
that does work.
--
Tinkerer

On 14/05/2011 18:30, Huge wrote:


*So* much easier than Linux, eh?

This practice is admitedly deeply technical, a challenge that end users
shouldn't normally find themselves involved in.

In similar hot water in some builds of linux, ye'll be swimming in the
seas of the command prompt, wishing ye had google by your side (ye
don't) and having a fine time learning/crying about partition tables and
grub. And where is/was /home?

Interesting stuff though. I can't live in camps anymore. Skill up on
both!!!

--
Adrian C

tony sayer wrote:
In article , The Natural Philosopher
scribeth thus
John Williamson wrote:
I've just found an old Toshiba Satellite 1800, which I'll probably end
up installing ME

ME!, Arrgh!!!, even Vista is better than that pile of cack.

Can't you find a copy of WIN 200 Pro or XP even?..

XP runs like a snail on Valium on the machine in question, which is why
I'm investigating Linux. For the use intended, ME is known to work
without problems. It should never even be connected to a network more
closely than by sneakernet and a shared USB drive, so security won't be
a problem.

Workflow will be 4 or 6 audio sources - record - edit- burn CD.

While editing, backups will be made, but not over a network.

For 2 track work, I've got a nice, silent laptop that runs 98, and a
very nice little interface to gowith it. Total power use is about 30W.

--
Tciao for Now!

John.

On 14/05/2011 17:23, Tinkerer wrote:

If the right option isn't there, see my post above in the thread for a way
that does work.

Hmmm, apart from a new boot sector I can't see what that fixes. However
if it works for you then fine ;-)



Have a look at this wonder.

http://www.instructables.com/id/Comp-ible/

A dumb solution if ever there was one...

--
Adrian C

On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on with it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.
--
(\__/) M.
(='.'=) Due to the amount of spam posted via googlegroups and
(")_(") their inaction to the problem. I am blocking some articles
posted from there. If you wish your postings to be seen by
everyone you will need use a different method of posting.

On 16/05/2011 09:10, Mark wrote:
On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on with it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).
This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and reviewed
the spec for my POST, and discussed in infinite detail whether it'd be
better to clear a register or load 'zero' into it.... that kind of
committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI assembler.

Then they realised that nobody else on the project spoke TI assembler -
(which is why I was there on contract in the first place!)
so there was no way that anybody could independently review by finished
work.

So they just scribbled some made-up initials in the 'reviewed' box, and
carried on......

Daft! - but it kept me off the streets for a year or so.... g

Adrian

HI John

On 16/05/2011 12:18, John Rumm wrote:
On 16/05/2011 10:33, Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).

That takes me back a bit... the joys of the ACP diagram! A standard that
does not quite know if its a design methodology, a language, or a run
time system ;-) (answer: its most of those, with a strong leaning toward
CORAL 66 IIRC).

Ah yes - the ACP diagram.
I seem to remember there was strong competition among the high-level
designers to see if the ACP diagrams could be made to resemble
a) Cartoon characters
b) various forms of genitalia or assorted sex acts
...seriously!

Complete waste of time.....(IMHO)

Better yet was that the micro was controlling something called a
'Y-relay' - and obscenely large relay with something like 1/2"
separation between the switched contacts, and actuated by a coil
the size of a toilet roll. This was a safety-critical component, and
couldn't be modified.

Trouble was, when the thing de-energised, the resultant back emf was
sufficient to power the uP board for several tens of seconds - during
which time it powered up, went through it's self-test, and started to
communicate up the line - only to be cut off in mid-flow when the
'oomph' finally died away. This generated comms / system errors further
up the chain of command, and it all went bad.... Hey ho!

I never did find out if the system eventually went live....

Adrian

Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:
On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on with
it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).
This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and reviewed
the spec for my POST, and discussed in infinite detail whether it'd be
better to clear a register or load 'zero' into it.... that kind of
committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI assembler.

Then they realised that nobody else on the project spoke TI assembler -
(which is why I was there on contract in the first place!)
so there was no way that anybody could independently review by finished
work.

So they just scribbled some made-up initials in the 'reviewed' box, and
carried on......

Daft! - but it kept me off the streets for a year or so.... g


Bin there done that ....

Actually some of the mil spec stuff made sense, like watchdog timers and
filling spare blocks of ROM with reset instructions..so if your missile
got EMP'ed it would reset itself, or self destruct if that failed.

And there was a bit of a fuffle round the lab when some bright spark
remarked that if an early version of sea wolf would pop its clogs 1/4
second after losing radar beam ride, this would be embarrassing as it
didn't get INTO the beam for half a second..and would self destruct as
it passed the ship's bridge.

Adrian

John Rumm wrote:
On 16/05/2011 10:33, Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).

That takes me back a bit... the joys of the ACP diagram! A standard that
does not quite know if its a design methodology, a language, or a run
time system ;-) (answer: its most of those, with a strong leaning
toward CORAL 66 IIRC).

This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and reviewed
the spec for my POST, and discussed in infinite detail whether it'd be
better to clear a register or load 'zero' into it.... that kind of
committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI
assembler.

Sounds somewhat similar... played much the same game for a network of
milspec workstations running 16MHz 80386 when 80386 was new and sexy[1]
in the late 80's.

[1] They built a working system running on standardised (to the project)
8086 cards, and found firstly this particular app was going to be two
slow on them, and secondly, although each application would fit in the
available EPROM space, the box was supposed to be able to behave in one
of four different roles selected at startup, and there was not room to
get all four in there. The hardware team had come up with a '386 card to
"solve" the problems. Had they have left it at that, they could have
slapped some extra storage in the box and a bit of bank switching logic,
let the thing power up in real mode and stay there, and they would have
been back to a working system in a couple of weeks what what looked like
a 'kin fast 8086 and some switchable ROM space.

Alas someone thought it would be fun to play with the new '386 toy and
set about building a MASCOT kernel for it in protected mode, and then
titting about attempting to create virtual 8086 tasks to run each role,
having cobbled together the required address space for it using the
memory protection tools the processor has. All clever stuff at the time,
but in the days when you asked intel for assistance and their response
was "wow, no one has tried that before, can you let us know how you get
on!" - it took over two years to get back to square one!


I had the reverse. Did some work for Gould - digital scope.

After a week looking aghast at getting on for a meg of bank switched
RAM on a 6809 I sweetly asked 'frankly, why didn't you get a PC
motherboard, plug in all your clever analogue D to A, and use it as a
scope instead with a standard VGA screen?'

"mumble mumble..we didn't know it would get this big when we started.."

Adrian Brentnall wrote:
HI John

On 16/05/2011 12:18, John Rumm wrote:
On 16/05/2011 10:33, Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).

That takes me back a bit... the joys of the ACP diagram! A standard that
does not quite know if its a design methodology, a language, or a run
time system ;-) (answer: its most of those, with a strong leaning toward
CORAL 66 IIRC).

Ah yes - the ACP diagram.
I seem to remember there was strong competition among the high-level
designers to see if the ACP diagrams could be made to resemble
a) Cartoon characters
b) various forms of genitalia or assorted sex acts
..seriously!

Complete waste of time.....(IMHO)

Better yet was that the micro was controlling something called a
'Y-relay' - and obscenely large relay with something like 1/2"
separation between the switched contacts, and actuated by a coil
the size of a toilet roll. This was a safety-critical component, and
couldn't be modified.

Trouble was, when the thing de-energised, the resultant back emf was
sufficient to power the uP board for several tens of seconds - during
which time it powered up, went through it's self-test, and started to
communicate up the line - only to be cut off in mid-flow when the
'oomph' finally died away. This generated comms / system errors further
up the chain of command, and it all went bad.... Hey ho!

I never did find out if the system eventually went live....

Adrian

Cockups we have known...try the deicing on a radar dish that melted and
charred the fibreglass when tested in full view of the customer...
...the idiot who switched off the CNC milling machine at 5pm halfway
through milling a 100,000 quid phased array out of an 8ft slab of solid
alloy, and when restarted, didn't recalibrate it. One half was 1/4" out
with respect to the other.

Wouldn't have looked out of place in a museum of modern at, but was in
fact pure scrap.

Or the two months I spent working on a project that turned out to be
theoretically impossible.

On 16/05/2011 13:01, The Natural Philosopher wrote:
Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:
On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any more
but was '1', then incremented it again to see if it was round-about '2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on
with it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).
This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and
reviewed the spec for my POST, and discussed in infinite detail
whether it'd be better to clear a register or load 'zero' into it....
that kind of committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI
assembler.

Then they realised that nobody else on the project spoke TI assembler -
(which is why I was there on contract in the first place!)
so there was no way that anybody could independently review by
finished work.

So they just scribbled some made-up initials in the 'reviewed' box, and
carried on......

Daft! - but it kept me off the streets for a year or so.... g


Bin there done that ....

Actually some of the mil spec stuff made sense, like watchdog timers and
filling spare blocks of ROM with reset instructions..so if your missile
got EMP'ed it would reset itself, or self destruct if that failed.

Never worked on miltary kit - but similar requirements on warehouse
automation / automatically guided vehicles etc - though we tended
towards the 'I'm confused - I'll sit here, flash my lights and look
pathetic' rather than actually self-destructing g

And there was a bit of a fuffle round the lab when some bright spark
remarked that if an early version of sea wolf would pop its clogs 1/4
second after losing radar beam ride, this would be embarrassing as it
didn't get INTO the beam for half a second..and would self destruct as
it passed the ship's bridge.


Mmmm!
There was the story (possibly true) of the torpedo guidance system that
had a fail-safe mechanism built into the code by a helpful programmer.

Logic ran - 'if I suddenly find that I'm on a reciprocal compass bearing
to the one I was on when I left the sub, then self-destruct'

Allegedly, one day on weapons trials, the armed fish got stuck in the
launch tube, at which point the Captain said - "B*gger this, lets go
home" and executed a swift 180-degree turn... and blew the sharp end of
the sub to bits!

Probably not true g
Adrian

In article , John Williamson johnwilli
writes

XP runs like a snail on Valium on the machine in question, which is why
I'm investigating Linux.

Try Puppy Linux. Runs well on very modest hardware.

--
Mike Tomlinson

In article , John
Rumm writes

Nope, dumb error messages are at large in the wild!

"lp1 on fire" (*nix)

--
Mike Tomlinson

Mike Tomlinson wrote:
In article , John Williamson johnwilli
writes

XP runs like a snail on Valium on the machine in question, which is why
I'm investigating Linux.

Try Puppy Linux. Runs well on very modest hardware.

As long as what I want to do is supported. The machine I'm working on is
to be used for recording 4 tracks of audio to 24 bit, 44.1KHz files at a
time, through an interface which Linux supporters claim "just works".

Ubuntu studio with Ardour or XP with Cubase would be the ideal as far as
software goes. I'll have a look tonight, maybe, if I'm not starting work
at sparrowfart tomorrow.

--
Tciao for Now!

John.

John Rumm wrote:
On 16/05/2011 14:09, Adrian Brentnall wrote:
On 16/05/2011 13:01, The Natural Philosopher wrote:
Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:
On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days,
but I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first
job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any
more
but was '1', then incremented it again to see if it was round-about
'2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on
with it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).
This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and
reviewed the spec for my POST, and discussed in infinite detail
whether it'd be better to clear a register or load 'zero' into it....
that kind of committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI
assembler.

Then they realised that nobody else on the project spoke TI assembler -
(which is why I was there on contract in the first place!)
so there was no way that anybody could independently review by
finished work.

So they just scribbled some made-up initials in the 'reviewed' box, and
carried on......

Daft! - but it kept me off the streets for a year or so.... g


Bin there done that ....

Actually some of the mil spec stuff made sense, like watchdog timers and
filling spare blocks of ROM with reset instructions..so if your missile
got EMP'ed it would reset itself, or self destruct if that failed.

Never worked on miltary kit - but similar requirements on warehouse
automation / automatically guided vehicles etc - though we tended
towards the 'I'm confused - I'll sit here, flash my lights and look
pathetic' rather than actually self-destructing g

Military kit tends to avoid safety critical software if at all possible
(using easily provable hardware interlocks etc). However the MoD
definition I always found slightly amusing, as begin "software that has
the capability of unintentionally causing death". I.e. accepting that
its quite possible to have software designed to kill you that is not
safety critical.


And there was a bit of a fuffle round the lab when some bright spark
remarked that if an early version of sea wolf would pop its clogs 1/4
second after losing radar beam ride, this would be embarrassing as it
didn't get INTO the beam for half a second..and would self destruct as
it passed the ship's bridge.


Mmmm!
There was the story (possibly true) of the torpedo guidance system that
had a fail-safe mechanism built into the code by a helpful programmer.

Logic ran - 'if I suddenly find that I'm on a reciprocal compass bearing
to the one I was on when I left the sub, then self-destruct'

Allegedly, one day on weapons trials, the armed fish got stuck in the
launch tube, at which point the Captain said - "B*gger this, lets go
home" and executed a swift 180-degree turn... and blew the sharp end of
the sub to bits!

Probably not true g

Well not having read your post I just posted the version of this I
heard... so there may be some truth.

There are a few other possibly apocryphal tales of a similar nature that
do the rounds.

The one I herd was a fly by wire fighter on a test flight and he pulled
the stick right back.

At some point this caused a digital multiplier to overflow offering a
large negative number and it went into a steep dive instead.

I alwys liked the 'Slide Rule' book where Neville Shute works on the
private R100. Their competition is the R101 which is to be given diesel
engines, by government decree. It is heavier, and therefore they worry
about the ability of the pilot to turn the wheel that works the rudder.
Power steering is employed, and of course, this has to be limited to
prevent the pilot from tearing it off.

This worries Shute and his colleagues, as they are simply using
cables..they redo the calculation and indeed confirm that with the
gearing they have, the pilot simply wont be strong enough to tear the
rudder off against the back pressure....

"John Rumm" wrote in message
o.uk...
On 16/05/2011 13:01, The Natural Philosopher wrote:
Adrian Brentnall wrote:
On 16/05/2011 09:10, Mark wrote:
On Fri, 13 May 2011 21:27:57 +0100, Adrian Brentnall
wrote:

On 13/05/2011 21:13, Steve Walker wrote:

snip

Nope, dumb error messages are at large in the wild!

I'm sure I remember a great one from my ZX Spectrum and QL days, but
I
don't seem to be able to find any reference to it at the moment -
"Unprintable Error".

Does anyone else have any recollection of it?

No - but I remember being subcontracted to write a uP test for some
Texas 99000 kit that was to be used in a railway points switching
application.

At power-up the system was supposed to run a POST.
The Project Manager was a Postgrad Computer 'Expert' on his first job.
People skills of a water-cooler, and technical grasp even worse.

Had me writing Texas assembler code that solemnly loaded zero into an
internal register, incremented it, checked that it wasn't zero any
more
but was '1', then incremented it again to see if it was round-about
'2'
- and so on.
Then we tested the 'decrement' instruction ......

and so on.....
and so on .....!

He also wanted the uP to report back up the line if it found errors -
using a highly complex comms routine. Tried to point out that if the
micro couldn't handle simple maths then it wasn't likely to be very
successful at telling the outside world - but was told to get on
with it...

I believe they eventually got rid of him!

That kind of obsessive BIT was commonplace in the military and
aerospace environments years ago. It may still be but I got out of
the industry. IIRC I pointed out the same thing myself but got the
standard answer "It's in the spec" so we had to implement it.

Yes....
It was actually 'better' than that....
The entire project was designed in some kind of supa-duper high-level
design methodology (MASCOT - if my memory serves me right).
This was all about high-level graphical designs (using pencil & paper!)
and multiple reviews at every stage. They solemnly designed and
reviewed the spec for my POST, and discussed in infinite detail
whether it'd be better to clear a register or load 'zero' into it....
that kind of committee-based nonsense g
Eventually, a detailed spec arrive with me, which I coded in TI
assembler.

Then they realised that nobody else on the project spoke TI assembler -
(which is why I was there on contract in the first place!)
so there was no way that anybody could independently review by
finished work.

So they just scribbled some made-up initials in the 'reviewed' box, and
carried on......

Daft! - but it kept me off the streets for a year or so.... g


Bin there done that ....

Actually some of the mil spec stuff made sense, like watchdog timers and
filling spare blocks of ROM with reset instructions..so if your missile
got EMP'ed it would reset itself, or self destruct if that failed.

And there was a bit of a fuffle round the lab when some bright spark
remarked that if an early version of sea wolf would pop its clogs 1/4
second after losing radar beam ride, this would be embarrassing as it
didn't get INTO the beam for half a second..and would self destruct as
it passed the ship's bridge.

I seem to recall one tale of a homing torpedo that had a nice bit of logic
in the control software to self destruct it should it find itself on the
reverse of its launch bearing. Was a nice idea until someone tried a test
launch that failed totally - and it was left sat on the deck not having
moved. Then they turned the ship round to come home...

(it was not fitted with a main charge, but it did have a detonator!)

Torpedoes don't arm until they have travelled a few hundred yards.

tony sayer wrote:
In article , Adrian Brentnall
scribeth thus
HI All
Sorta DIY - I'm trying to fix it myself g

WinXP Pro SP3 PC, Dell fairly basic desktop (Vostro 200), about 4 years
old, fully updated, running Avast antivirus & Spybot.
Every so often, the PC just locks up solid, not responsive to mouse,
Alt-Ctl-Del or anything else...
Only way out is to pull the mains plug (I know it's not a good idea -
but what can you do ?)

Happens most often when running a particular web authoring package
(Netobjects Fusion) - but has also happened (once) with another web
authoring package (WYSIWYG).

Tried blowing the muck out of the PC fans etc. - but doesn't seem
to have made any difference. Lifted & reseated the memory.

I'm a bit at a loss for where to look next?
Any ideas, please ?

Thanks
Adrian


With that sort of symptom I'd try a substitute power supply first at
that sort of age. Most all the instances of this happening have been
hardware related. XP is a stable system otherwise..

Don't know if that make was subject to the dodgy capacitors that did
plague many motherboards some time ago, but worth a look to see if any
of them are leaking or bulging etc...

I had two caps go on a 14 month old PC. Got a new MB and it's fine, but
well worth a quick check for bulging caps.