JAMECO sells Pentium I and II processors. Thoughts?

[Doctor]

As I was browsing JAMECO'S catalog, and saw that they have Pentium I and II processors for sale. Then I started to wonder... Why? Do some people still use Pentium processors in computers? Is there a way to use them in a similar fashion to having an Atmega on a board?

It's just something odd that I was wondering if anyone else noticed. I brought it up at the hive, but there were only 5 of us to discuss it... Maybe someone else has an idea what to do with a Pentium. Any thoughts?

[Louis Gerbarg]

Legacy systems. If you have some old rack mount industrial system
running a custom piece of DOS (or OS/2 or Xenix) software to control
industrial hardware you don't want or need to upgrade. If something
breaks you want to replace it with an exact match instead of having to
migrate to new hardware including potentially risky ports of software
you might not even have the source to any more.

> --
> To post to this group, send email to hive76-d...@googlegroups.com
> To unsubscribe send email to hive76-discuss...@googlegroups.com
> For more awesome goto http://groups.google.com/group/hive76-discussion?hl=en

[Doctor]

True. I just think it might be a bad idea to stick with obsolete technology for too long, because an upgrade may be impossible down the line.

[Louis Gerbarg]

Consumer vs Industrial users have very different requirements. A lot
of embedded device are never upgraded. My father runs a manufacturing
plant that has "computer controlled" embroidery machines. They are
8086 machines and take 3.5 inch disks that have their patterns on
them. The file format of the embroidery designs is directly derived
from the magnetic tape readers the original model of the machine had.
Yes, if one breaks down it might be a real pita to deal with it, but
the incremental cost of upgrading the machines for the last three
decades is much higher than that. When something goes wrong they
scavenge the boneyard of other machines for parts rotate things
around, then go looking for a spare parts to replace whatever they
used. If they actually lose the ability to use a machine due to
irreplaceable parts they will likely just move to newer more capable
machines (they have a few already).

For a lot of users stuff working is much more important than anything
else, including upgradability. Spending weeks tinkering with a machine
to redesign interface boards created 30 years ago is simply not worth
the effort of finding some parts on eBay and moving on. Unlike us,
those users really don't want to tinker, time spent tinkering is just
lost money.

Louis

[Doctor]

Yeah... That all makes sense. I guess it depends on who you are, what you have, and what you want and need to do.

[pezman]

Interesting to read ...

I work in industrial control, specializing in migration.  It is often remarkably difficult to replace legacy systems with newer "better" systems -- in many cases, the older the system, the tougher it is, since the customer's entire way of doing things ultimately molds itself to the legacy system over the years.

You would be shocked at the prices that old hardware commands.

[Doctor]

I would guess legacy hardware is only cheap if there were literally millions of extras, or someone is re-manufacturing them for cheap.

[Jonathan Simpson]

One of my favorite stories is about an old unix box that I got called out to take a look at.

This box ran a program called orden, and was their order entry system.  It died a horrible death, but orden had been custom written for them. The customer needed orden back, at any cost.

It turns our orden was written in an archaic, interpreted basic language.  All the source files for orden were right there on the machine, and I had already successfully extracted the contents of the disk, so I had them.  After 3 weeks of wading through orden's collection of gotos and spaghetti, they had a new orden.  This new orden was built in quickbasic 4.5, which itself was nearly 20 years old at the time.   The check I got for "updating" this is still the largest deposit I've ever made in my life, but it felt awful dirty.

I still have the orden source code in my storage system.  You never know when someone might need a DOS based, goto laden order entry system designed in the early 80s and "updated" to the late 80s.  I pray they never call me again.

Jonathan Simpson

[Doctor]

Sometimes it does come to that... But my question is, if you HAVE to update, why not update to a new system? Modern, even if basic?

[andrew sooy]

For the sewing machine couldn't there be something made like the tape that plugs in to an ipod to play mp3 on a tape player but to have it hook to a new computer.

[Sean McBeth]

Because you should almost never rewrite software:  http://onstartups.com/tabid/3339/bid/2596/Why-You-Should-Almost-Never-Rewrite-Your-Software.aspx 

Rewriting software, rather than upgrading it, is tantamount to wiping out all of your memories for the last 10 years just because you had one bad date.

Only amateur programmers immediately jump to the "rewrite" solution. You can extend with newer technology, you can replace bits and pieces over time, but starting from scratch is dumb.

[Sean McBeth]

Also, this:  http://www.joelonsoftware.com/articles/fog0000000069.html 

[Louis Gerbarg]

I didn't say audio tape, I said magnetic tape (of which audio is a
specific type). The key point being that machines don't have anything
resembling an audio hookup. Introducing audio into the mix is just a
whole bunch of extra hardware pieces that can break and/or introduce
errors. It might have an appeal as a Rube Goldberg style hack, but it
is generally not something you want to try deploy and support.

I had a coworker once tell me that when given a certain class of
technical challenges there are 3 types of engineers

1) Those who think it can't be done
2) A few talented ones who figure out how to do it
3) An even fewer number of talented ones with enough experience who
know both how to do it and why despite the fact it is possible it will
be more grief than it is worth

Louis

[Kyle Yankanich]

Because older architectures are often more robust. We have several mainframes at work, and each only does a few functions. There's a whole multi-million dollar mainframe that literally billions of orders go through - that can only be accessed via terminal. Why? The software was wrote that was in 1986, and it's worked great. We scale the mainframe's hardware requirements up as needed, and patch the latest security issues and call it a day.

Perfect example - there's a program in the mainframe community called CICS. It's a customer transaction systems. Takes stock out of inventory, manages inventory, helps with purchase orders, tracking stock, etc. It was written in 1968. It receives regular updates, but the core of the program is the same - and over 90% of Fortune 500 companies use it. It just works, and it works well.

I did an internship at Comcast's West Chester Data Center - they ran one mainframe, that I regret not asking questions about at the time. They literally had boxes of CPU's for when one burned out - they'd just swap in a replacement. It was worth it for them, instead of buying all new hardware, finding software to run on top of that, rewriting any extensions and modifications they had, modifying all client software - to just throw some new outdated CPU's in.

[andrew11235]

i Didn't mean over audio i ment like the magnetic tape jawns for mp3 players but converted to a floppy that can go in it and use USB or somthing.
Then have it come up on the computer and have software to just switch/mount files when needed.

[Louis Gerbarg]

I see what you are saying, but it makes no sense except as an exercise
to prove it can be done.

1) Using floppy disks works, what does adding some complicated rube
goldberg floppy disk emulation system to avoid the floppy disk
achieve?
2) The device you are describing would have to substantially more
complicated than the MP3 tape adapters you describe. Magnetic tapes
are linear access, floppy disks are random access. In order to do what
you want you would either need a system that emulates the entire
magnetic profile of the disk surface (IOW, is a floppy disk), or can
dynamically intuit where the drive is (or emulate an entire track
sector and swap it on a loop synchronized to the drives speed, which
is tricky since some floppy formats use soft sectoring and variable
speed drives).

If you really want to get off of floppy disks what you would do is
build an embedded controller that electrically emulates a floppy
drive, not magnetically emulates a floppy disk. And that is a common
enough retrofit for legacy systems you can buy them off the shelf:
http://en.wikipedia.org/wiki/Floppy_disk_hardware_emulator .

Louis