Rod Speed wrote:
clare at snyder.on.ca wrote
Rod Speed wrote
Too_Many_Tools wrote

Most companies data isn't worth anything after only a handful of years.

Engineering data is the heart of a business.

Not data thats a handful of years old.

Management often forgets that.

Then a competitor eats them alive.

Bet you cant list any examples of that with data thats older than a handful of years old.

I sure can.

Nope, you couldnt.

I'm sure he could, and I can add a few more, both of own and from other
references as well--

From own experience, it's a regulatory requirement of NRC to keep _all_
safety-related design documentation and calculations for 40 years of
"life of the plant". That's simply one instance of on need for
longterm records-keeping.

There's a whole industry dedicated to preserving data for companies
from finanical to manufacturing and everything in between. It's a
major use of the excavated areas of the salt mines in central KS as
they're fantastically dry, constant temperature, fire and vermin-free
and of humongous size.

For a couple of stories you might check out Jack Ganssle's columns that
he writes for Embedded System magazine -- a mostly unheard of by very
important niche of the microprocessor world. In fact, there are far
more processors used in such applications than in PCs though they don't
have the glamour of the "lastest and fastest" whatever of the day...

http://www.embedded.com/columns/bp/s...cleID=22103292

Jack also distributes a monthly newsletter that has had as one of its
subjects recently reconstructing "legacy" systems. I myself have had
requests for modifications of some systems I had previously worked on
that I would have thought long since "dead and buried" having moved on
to other projects and even other companies, but was tracked down as the
only individual they could find that had any recollection of the actual
system.

Another reader of Jack's newsletter sent an interesting tale of his
experience --

....
"I was brought in as a consultant for one of the downstream users of
an early video-on-demand companies, who supplied complete systems and
programming to hotels and hospitals, even providing a broadband network
infrastructure for free to sell their services.

"There was a need to add new educational programming services for a
client market, or be displaced by a competitor.

"The company had not built their code from scratch in more than 10
years. In fact, they had decided to move to cross compilation rather
than self hosting for a while, had bought a commuter and new tools,
never tried the tools, and had subsequently sold the cross host machine
for scrap.

"Our first task was to put together a development environment hosted
on a "dead" OS, including compilers, linkers, and build control files,
gather known source, and attempt to rebuild the shipping object from
known source.

"This took several months, and was a real adventure. A year and a
half down the road, job complete, ..."

He goes on to describe the system and other technical details probably
of little if any interest here, but needless to say, that little
misadventure of not preserving nor updating their ability to rebuild
their product's software undoubtedly cost that company a pretty penny
and without that effort likely could indeed have put them out of at
least that particular business.

Undoubtedly, these few instances given here are far from the only
occurrences of such in industry. And, for every one that did manage to
recover, how many were there who were unable to?