Re: No Y2k Bug

[/dev/null@.com (Justa Lurker)]

It was Fri, 07 Jan 2000 13:45:03 -0500, and Joshua P. Hill
<XXjoshhill@mindspring.com> wrote in misc.transport.urban-transit:
| >Yes.  One uses what is available, usually until something better is
| >avaiable.  My first computer was a 4k RAM whoknows clock speed box
| >that WORKED.  The next was a glorious 286-10Mhz with a 30meg HD - wow
| >that WAS fast.  Then the next one came.
| >Each machine was state of the art when I used it.  And now I am on a
| >different state of the art machine.  Mission critical - REAL TIME.
| 
| None of the machiens you refer to, TRS-80 or otherwise, was ever
| state-of-the-art. They were all personal computers.

State of the art FOR THEIR MISSION.

| Look, we aren't talking pimply faced kids here. In the days of the
| TRS-80, small scale control applications were either built around
| dedicated boards with their own processors or microcontrollers, or, at
| the next higher level, minis like the PDP-11. No one beyond
| adolescence would have used a TRS-80 in an engineering project.

Just like today.  The fail safe portion of the equasion is not based
on a machine that could forget how to tie its own bits together.  It
is put into equipment that senses failure and falls to a protective
state.  (Unless the whole applied system is poorly designed, and the
fail safe is missing or broken.)

| NT is far from the only operating system avaikable, and among the many
| that are, it is not the appropriate one for mission critical
| applications. At best its best, it works in apps that require
| semi-reliability such as servers.

I didn't say it was the only one.  Just that it will do the job.  You
seem to be unable to accept the fact that people DO use NT in mission
critical roles.  Just because you are biased against that OS does not
mean that it is not successfully used in other people's systems.

| >As I said before, it depends on the mission.  A TRS-80 is fine IF you
| >don't mind its failings AND you build what the box controls with
| >appropriate failsafes so that if/when the box does die, it fails SAFE.
| 
| That's ridiculous. Why use something unreliable and put a kludge
| around it when there are more reliable OS's that are designed for this
| sort of application? There are no issues of economy or functionality
| here; the architecture of NT just isn't good enough for sufficient
| stablity or consistency.

If you are not putting fail safes around whatever OS you promote, then
you are operating in a dangerous manner.  Please give a list of all
the designs that you have worked on so that we can avoid the danger.

| >| And NT was *not* designed for real time control. Period.
| >
| >That is your opinion.  You are welcome to it.
| 
| Rubbish; it's a basic engineering fact. I've read several articles
| that mention the architectural issues involved, and the possibility
| that Microsoft might fix them some day.

This coming from the same company that believes that not being able to
type 'DATE 1-1-00' at the command prompt is a "bug".  You are
listening to lawyer talk, trying to keep MS off the hook.  MS also has
the handicap of being a software company - if the hardware processor
fails their OS is useless.

| >I'm sure you can name a few OS's that you believe are better for
| >mission critical or real time applications.  Name ONE that has never
| >failed.  Note the word NEVER.
| 
| You: "I send my clients across the Atlantic in rowboats."
| 
| Me: "That's ridiculous; a rowboat isn't the right kind of boat for
| that."
| 
| You: "Name ONE kind of ship that has never sunk. Note the word NEVER."

No comparison to reality unless you consider NT to be a rowboat,
and yet you still fail to name an OS that doesn't fail.

Unless your OS NEVER fails you still have to build somthing around it
to protect the mission you have assigned it to.  Nothing worse than an
OS that says everything is fine as it systematically destroys
everyhing in its path.  The most basic failsafe is knowing when the OS
has ceased to function.

JL