Tales from the flying disk doctor - Crime and Punishment

Copyright Dr Alan Solomon, 1986-1995

I got a phone call from Mark at XYZ PLC.  "We have a team of
programmers developing a dBase application.  They're using a Novell
network, with an AT as a fileserver.  They moved the AT, and ran Park
before they moved it.  But it was the wrong version of Park, and now
they can't read their hard disk."

I didn't bother asking if there was a backup - they wouldn't be
consulting me if there was.  I've seen this problem several times
before;  the wrong version of Park (HDPark, Shutdown, whatever it
mught be called on various machines) can be fatal to hard disks.  If
you're lucky, it will park the heads on a data track, and if you're
unlucky, it will park them half an inch outside the drive casing;
goodbye drive.  I know how to deal with it, too;  it isn't as
impossible to fix as IBM thinks.  "Yes, I can fix that for you", and I
named a very modest three figure sum.  Mark was over the moon, so I
explained that I'd need the machine, the Novell software and manuals,
and a replacement hard disk.  The replacement hard disk is because I
advise people not to reuse a disk that has had a hardware failure
unless the application is unimportant, and I think that this counts as
a hardware problem, even though it is caused by software.

Mark said he was happy with the price, but that it was very urgent,
and could I do it by tomorrow noon - money no object.  This would mean
that I'd have to drop everything and work full-time on his disk, with
a risk that there was more wrong with it than usual, and that if it
was bad and would take longer than I was given, I'd have to work all
night (and I'm a bit long in the tooth for wraparounds), or even that
the disk could be undoable, in which case I wouldn't get paid (I only
take money if successful, which is 90-95% of the time).  So I added a
bit to the price to cover the risk of having to work all through the
night, and Mark was still over the moon.  I'm not sure why time was so
much of the essence, but I make a practice of not arguing with people
who use the magic words "money is no object".  I think they were also
using the network for some kind of financial business, and down time
would cost them quite a lot.  Mark said he'd make the arrangements,
and I started clearing out my disk quacking room, and phoning people
up to cancel things.

An hour later Mark phoned back, and he sounded awful.  "We can't do
it", he said.  "Why not?", I asked.  This does happen sometimes;
either it turns out that they haven't lost as much as they thought, or
they have a security problem and don't trust the integrity of a
professional, or they find they did have a backup after all, only a
few days out of date.  But there is another reason, and I've heard it
before, and it makes me feel ill.  "The DP department says it serves
us right, we should have been doing backups, and we shouldn't have
moved it without permission, and they aren't going to authorise the
expenditure.  And they've come and started doing a CompSurf, and
that's the end of it."

In other words, they were being punished by the computer department
for their crime.  Can you believe that?  And can you believe that it
isn't the first time I've had this?

Now it doesn't affect me, except emotionally - I just phoned up the
guy I'd cancelled and told him I could do his disk after all.  But I
keep thinking about this bunch of professional programmers developing
their dBase thing, and they're going to have to re-create a month's
work.  I had to recreate an hour's work once, because I only backup a
couple of times per day, and that was a horrible thing to have to do,
and I don't think I did it nearly so well the second time, because I
couldn't remember what I'd done and what I hadn't done.  Can you
imagine, a month's work, and all the time knowing that it was
unnecessary?

But you should also spare a bit of sympathy for the poor guy who
refused to authorise the data recovery.  A month's work could perhaps
be re-created in a week if you're lucky, and three people working for
a week will cost the company perhaps five grand.  So he's in a
slightly delicate situation, politically, which wouldn't matter if you
didn't have any enemies, only that guy has three very dedicated
enemies who will do anything to get even - you should have heard the
way Mark said his name.  If he is still in that company next year,
he'll be lucky.

Anyway, I didn't bother to explain to Mark that I'd have a go at a
CompSurfed disk, as it would only have made him feel worse, but I did
remind him that he would be using a damaged disk from now on, and he
should take extra care.  He was very apologetic about the trouble he'd
put me to, and I said it was all right.

All over the country there are people like Mark, committing the crime
of not doing a backup, and being punished for it by their DP
department.  Most of them I never hear about, because they don't even
get as far as talking to me, and I'm sorry guys, but I can't see how I
can help you.  But it really is very depressing to see someone fall
over a cliff, manage to grab a branch as they go over, and then have
that branch sawed through by someone who is supposed to support them.

Rather than leave you with such a depressing thought, let me tell you
a more cheerful story.  I got sent a Reflex database on a diskette,
and one of the sectors of the diskette had gone bad.  If it were
dBase, or PC-File, it would have been a piece of cake, as they store
their data in nice simple, rectangular ASCII files, and quacking that
is easy.  But when I looked at the Reflex file format, I could see
that it was a complex, structured thing, very reminiscent of a Lotus
spreadsheet.  It took me six weeks to reverse engineer the Lotus file
format, and it was worth every minute.  But Reflex isn't so widely
used, and I didn't fancy doing a reverse engineering job just for the
sake of one diskette.

So I rang Borland, and spoke to one of the tech support people there,
and asked him for the file format.  "No problem", he said, cheerfully,
"I'll pop it in the post." When I tried the same thing with
Ashton-Tate on Framework, the answer was "No dice, and furthermore, we
won't tell you the file format for Multimate or dBase".  Which is a
good trick, since they've already published the dBase format.  "Why
not?", I asked.  "We don't want you getting inside our product," they
said.  So, if you've for a trashed disk full of Framework, forget it.
But if you use Reflex, I can sort you out.  Who do you think is being
clever, Borland or Ashton-Tate?