Is it a hoax? 
Author Message
 Is it a hoax?


> > Does anybody feel the problem is being over hyped?

> I don't.

> > Is it really as bad as the media suggests?

> It's much worse.  Mission critical systems fall over now and then.  When it
> happens, those involved remember the terror for their entire careers.  In Y2K
> every system has a high probability of doing something wrong, something
> strange and wonderful.  The problem occurs when every production system goes
> berserk, starts flinging bytes around like a crazy.

Cory, here's a real world disaster.

I worked for a company that serviced a utility.  Their maintenance was
CICS based.  Pending Maintenance was entered, and for the pending record
the ENTIRE mapset was stored as the record, appended to some key
fields.  When the pending maint was recalled, the entire record was fed
back into the map.

Now, this needed modifying.  More often recently.  Now, throw in the
fact that it is actually 2 screens of data, concatinated and stored in a
single record.  Now insert a field in the first of the two.  For data
conversion of existing records, you can't just move the fields down
uniformly.  You have to move the first set down into the slack bytes at
the end.  

Well the time came when it was time to expand the fields (insert on
screen 1) beyond the limit of the remaining slack bytes.  So it was
determined to do what should have been done the first time, and move the
fields individually to the record.

Well well well.  The conversion looked good.  The entry of maintenance
looked good.  Recall of newly entered maint looked good.  Some minor
thing, about some fields not showing up - ones that were not changed,
but used for inquiry, but hey, we're ready.  (I wasn't involved up to
this point - can you tell?)

Anyway, put the sucker in production.  Somehow recall of existing
maintenance had a wierd affect.  Fields were changed that were not meant
to be, new maintenance was messed up.  Wierd and wonderful things were
happening.  The untility could not control who had service and who did
not.  Readings failed.  Whattamess.

This is where I entered the picture.  What I did first was to learn HOW
the original application worked.  Here's the kicker.  What it did was
retrieve the record.  Displaying all iformation and placing it in the
map.  Key information with Modified Data Tags on.  Then of course, the
changed info would have the MDT's on.  Then before the write, the other
display information was re-retrieved from the record and saved in the
maps data fields, but the MTDs were not on.  Then that record was
written.  Recall of maint would read the record directly into the map.
Walla- MDT's are on on the fields you last modified! <G>.

So, the fix.  First - and this is the Y2K relationship - the DAMAGE WAS
DONE. The converted file LOST all of the MDTs.  NEW data had been
entered.  Recovery was important.

I took the backup of the prior records, before conversion and compared
them with was was on file.  I printed the NEWLY added maint, and deleted
it.  They had to re-enter it.  I added the attribute bytes to the end of
the record.  Once the slack bytes were gone, there was plenty of room.
I then used the original data to recover the attribute bytes.  Then I
had the guy who made the changes (This was a team effort you know)
change the program to store the attributes and the data, and to restore
them on a read.  The convoluted, VERY efficient program logic was
maintained and disaster was averted.

This is Y2K in miniature.  The damage that will be done by unremediated
systems will be HARD to fix.  The newbie programmers won't have the
skill or experience to UNDERSTAND the original logic (just as the person
who started the above mentioned project didn't understand how it worked
before he changed it), and they won't be able to fix it, or they will
cause things to break in wierd and wonderful ways.

Sat, 23 Sep 2000 03:00:00 GMT  
 [ 1 post ] 

 Relevant Pages 

1. I am not deaf, but am I mute?


3. Hoax AOL Virus

4. Virus hoaxes

5. OCX in C4 - a hoax?


7. Creators Admit UNIX, C Hoax

8. Fw: Major Virus Warning -- Is a Hoax

9. forth the big hoax

10. FORTH, the big hoax.

11. The C Hoax

12. Software Engineering is not a hoax...


Powered by phpBB® Forum Software