DC Y2K Weather Report # 7 
Author Message
 DC Y2K Weather Report # 7

Cory Hamasaki's DC Y2K Weather Report # 7
    (c) 1997 Cory Hamasaki /* send copies of this around, just leave the
(c) in and the entire text.  */

Over the last few months the Washington Post has been slamming various
federal agencies,  not for Y2K failures, no, no,  that's too far in the
future and too hard to understand.  The Post has been hitting them for
barfing on projects in the range of a half billion dollars
and up.  Remember, this is Washington DC,  a mere hundred million
dollars is not worth thinking about.  Check out the
Washingon Post's archives, http://www.*-*-*.com/ available 24
hours a day.

You know the big duds, I won't embarrass the agencies any further, after
all, some of them are very well armed.  "You, you in there, throw out
your keyboard and come out with your hands up."

While it's easy to blame the feds, they shoulder only half the
responsiblity, the other half is the shadow government. A collection of
consultants, contractors, and rent-a-programmer shops that encircle
Washington DC.  Yes, I'm one of them,  but I struggle very hard to spend
my time cranking code and bring code knights back from the dark side of
the force (feasibility studies, wacked out looney-tunes research
projects to build software death stars).

If you drink Mai-Tai's all day with the bad-girls at Club ChoCho on
Keamoku Street, paddle out to 4 foot surf off Kaiser's, and have a real
weird dream as the hydraulic tumbles you off your board, you have the
makings of a software research proposal to a DC area think tank.  The
nuttiest, looniest, most far out thinking can get funding.  It's because
we've downsized, rightsized, and dumbed down for so long that they can't
tell fantasy from reality.

Remember the old days when 1) Everyone used ANSI COBOL because they were
afraid to become locked into a proprietary language and lose
support when the vendor went belly up, 2) vendors had to escrow the
source code to their products and had to prove that the product was
written in a standard, maintainable language, 3) A central control
authority reviewed and logged all source, source updates, job control
language, documentation, 4) Another central control authority controlled
data and data definitions?

That kind of efficient thinking vanished with the dumbing down.  For the
last 15 years, the dumbing down and rightsizing has cut all the
sensible support infrastructure and left us with a bunch of
scammers, wanna-be's, and computer professional pretenders.  You've met
them; even if your organization escaped the madness, you've read their
memos, heard them talk at interagency meetings, had to work with their
cr*p.  Perhaps you've had to fire one of them or (hopefully not) been
fired by one.

Almost as bad is being interviewed by them.  They can pronounce some of
the words but they don't know what they mean.  You're frustrated because
you have EXCPVR experience with ASYNC and BISYNC communications devices
and they think you can't handle a CICS job.  It's even worse when they
have some small understanding but still can't appreciate your depth.
It's their loss.

Here's my assessment of this Sunday's Post (March 22, 1997). 16+ pages
of "Computer Geek Needed, must have pocket protector" ads.  Still mostly
PC stuff, C/C++, Oracle-SQeaLer, Lan-grunt but the mainframe, MVS, COBOL
ads are increasing.  The turning point seems to have been Jan 1997.

Remember Y2K is a "come as you are crisis."  If your company has 30 code
crankers, that is all you will have through the crisis.  As fast as you
hire, someone will toss a net over your crankers and steal them away.
I've had first-hand reports of 30-40% salary increases over the last
year; and the raiding has yet to begin.

What's in the Post?  The ads are starting to list salaries!  Glancing
over the pages, $75K seems to be the common salary (Yes, I know they
list as $55K-$75K, but come-on, if they can go $75K, you're not going to
talk them down.)  Several had nice big numbers.  There were enough $100K
ads that we have reached the magic $100K/year number for experienced
crankers with specialties.   Common salaries are somewhere in the
$55K-$75K/year range.

If you're Joe COBOL, 5 years, understand copy books, did a little VSAM,
some tape processing, maybe extracted data from DB2, did maintenance on
a CICS transaction, you can get $75,000/year around Washington DC.

The jump to $100,000/year is a big one but someone with project
management experience,  a few more years,  more understanding of DB2,
CICS, and maybe some S/370 assembler should be able to command that
salary.   These guidelines change if you move 50 miles away.  The rates
will also be different next month.

-------------------------------

The Guild (a recruiting firm and job shop) - Project Manager (60K-180K),

DynCorp has a nice big ad, still begging for MVS mechanics, OS/390
GenMasters, VTAM'ers, HSM-SMS, these are separate listings.  Odd, they
list IEBCOPY skills as a REQ for the senior MVS communications
programmer.  <I'd do the work for $75K but if they want IEBCOPY,
welllll-lll, now we're talking $175K.>  Find out if it's a subtle joke

with the U.S. Department of Transportation.

Uh-oh, Fairfax Country has three separate ads, a very flashy one on page
K18 looks like an explosion of MVS acronyms.  In Y2K, government
agencies, especially state and local, are at a grave disadvantage.  It's
strictly salary.  They can't offer $100K/year salaries to keep their
talent.  Check them out at http://www.*-*-*.com/

The 'Polish Rocket Company', aka Litton PRC, wants MVS-heads, PL/I,
probably for contracts downtown.  The Rocket Company used to have some
problems but I've heard through the DC area Wizards grapevine that
they've cleaned up their act.  One of their managers, a Mr. Manino,
spent a lot of time (and lost most of his hair as a consequence)
cleaning up their projects, trouble shooting, and fixing things.  PRC is
at http:www.prc.com

Management Systems Designers is still trying to staff the IRS (It Really
Steals, name offered by Frank) with Assembly Language crankers. They
have one of the top S/370 assembly language programmers in the world.
MSD is at http://www.*-*-*.com/

Ooooh, here's a blind ad, Senior Management with COBOL, VSAM, and
JCL.  Sayyyy-whattt? Senior Management and JCL, DB2 a plus?  What's
going on?  If you find out, let us know.  Washington Post Box M1827,
Washington DC 20071-6250.  The Post has a remailing service for the shy,
mostly it's used by 'ISO Wealthy Sugar Daddy' or 'ISO nubile pneumatic
females' type ads.  I checked the box number twice to make sure that I
didn't send you to an ISO box by mistake.

Ooooh again, here's another blind ad, COBOL, MVS in Fredericksburg.
Washington Post blind box M1942.  Wash DC 20071-6250  It's getting
{*filter*}, next we'll be meeting recruiters in dark alleys.

Marriot's still looking for COBOL, email: IS&T.staffing.marriott.com

The Network Automation Technology Group, NATI wants Y2K'er with COBOL,
MVS, etc. fax: (703) 714-6994  <I don't know anything about them.>

Raytheon E-Systems wants various mainframers, they say they won a Dept
of Education contract to service Student Loans.  Excuse-meeee, I thought
SALLIE MAE had a lock on that.  Oh right, someone had a brilliant idea
to re-federalized the student loan program.  Looks like E-Systems is the
stuck-ee.  They want IDMS programmers?  How long before they're in
default?  The good news is that E-Systems has a local reputation for
working on cost is no object projects.  Help them out, email:

SALLIE MAE was back again with a nice big ad, they also want MVS, COBOL,
JCL, they're looking for SMP-E too <Apply at both E-Systems and SALLIE
MAE and let the bidding begin.  Tell us who has the bucks. You know
that E-Systems is already raiding SALLIE.> http://www.*-*-*.com/

Odd, a few nice sized OS/2 ads, CTG wants a DB-tutu DBA in Fairfax,

Chevy Chase Bank wants MVS mainframers and OS/2'ers, email:

Telemet America wants OS/2 C programmers for a real-time stock
information system, fax: (703) 684-7374

<Editorial musings, a lot of OS/2 skills scale up to MVS nicely; Y2K
mainframe remediation may start tapping the OS/2 expert market, DB2/2,
Rexx, Visual Age and CSet, PL/I for OS/2, if you like OS/2, you'll
tolerate MVS with only a few grunts and groans.>  

This isn't Y2K but worth mentioning, DENRO, they make some kind of voice
switch used for air traffic control, wants C++, offers a 10% signon
bonus and a day off every other week during the summer so you can work
on your tan.  <Nice creative ad and I like the 10% pay yourself a
finders fee.  They've raised the bar.>   http://www.*-*-*.com/

WMATA, the local bus and subway company, wants COBOL, VSAM, etc. Datacom
too,  They can't afford to pay the kind of money you deserve.  If you
like mass transportation, check them out,  voice: (202) 962-6262

Here's a joker, QCS wants Adabase (sic), COBOL, VSAM, and says "Part
time, Full time, Consultants Not Y2k"  Huh?  It's all Y2K now.  fax:
(703) 847-5662

Another blind ad, TSO, CMS, COBOL (and a bunch of PC stuff) $85K fax:
(202) 887-8445  <What are you companies ashamed of?  Blind ads are the
realm of overweight middle aged women looking for love and overweight
middle aged (married) men looking for a fling.  Just say who you are,
what skills you want, and the salary range.  Is that so hard? >

And the final touch, a recruiter's blind ad,  "to $275K!", I won't
bother with the contact information on this one.  It's too sleazy for
words.

------------------------

Some closing shots.  If you think Y2K isn't a problem, please keep your
mouth <keyboard?> shut.  Y2K is a huge problem and if you think
otherwise, you are sadly mistaken and embarassing yourself far more than
I ever could.

I've done my research on Y2K; ...

read more »



Fri, 10 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7


Quote:
> Cory Hamasaki's DC Y2K Weather Report # 7
>     (c) 1997 Cory Hamasaki /* send copies of this around, just leave the
> (c) in and the entire text.  */

> Over the last few months the Washington Post has been slamming various
> federal agencies,  not for Y2K failures, no, no,  that's too far in the
> future and too hard to understand.  The Post has been hitting them for
> barfing on projects in the range of a half billion dollars
> and up.  Remember, this is Washington DC,  a mere hundred million
> dollars is not worth thinking about.  Check out the
> Washingon Post's archives, http://www.washingtonpost.com available 24
> hours a day.

big snip

As for "dumbing down," how about this one?

I did some work for a federal agency, which shall remain nameless to
protect the guilty.  A few years later, a friend who was doing an
evaluation of their procedures called me.  When he asked why they didn't
have range checking and validation of data in their programs, their
response was "Pallat had a lot of that in his programs.  It didn't seem to
do anything, so we took it all out."

Sure.  Why bother with fuses and circuit breakers.  They never seem to do
anything either.

Remove the  '.'   from orion.data for sending email to me.


Orion Data Systems

Solicitations to me must be pre-approved in writing
by me after soliciitor pays $1,000 US per incident.
Solicitations sent to me are proof you accept this
notice and will send a certified check forthwith.



Fri, 10 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7

Quote:
> happy, I'm so happy to be able to read S/370 machine language.  SVC-11
> hits more than the TIME macro, it rips the guts out of the PL/I DATE()
> function.  There are implications that are an absolute scream.  Rex had

Does the PL/I DATETIME() function survive? I'd also be quite surprised if
LE370 date functions take a hit.

BTW, there is a mod out now that allows COBOL or PL/I programs to run with
NOSTAE,NOSPIE to run without the overhead of LE320 runtime telling you
that it has overridden your PLIXOPT or COBOL IGZEZOPT. It behaves as it did
before LE370 runtime. We haven't had it in production yet, so we can't say
whether this resolves the performance hit of IMS routines burning CPU in
LE runtime routines. IBM advised running IMS NOSTAE,NOSPIE prior to LE370, so
many online programs now burn more CPU to tell you that your COBOL or PL/I
runtime options have been overridden by the TRAP(ON) mandatory(used to be)
option. We may have had a double dose of this, since some of our load modules
combine PL/I, COBOL, and assembler.

Apart from the IBM IMS recommendation we had seen some obscure DB2 abends
before I assembled and linked an IGZEZOPT runtime parms module to suppress
COBOL II error handling. It seemed to be a case of too many error handlers
trying to own control of a contention/timeout/rollback/requeue situation.
--
notice: by sending advertising/solicitations to this account you will be
indicating your consent to paying me $70/hour for a minimum of 2 hours for
my time spent dealing with it



Fri, 10 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7


 <snip>

Quote:
>The SVC-11 problem is bigger than I thought.  Rex explained the problem,
>it's huge, a monster,  I'm terrified, then we talked about the
>impact on consulting rates, It's huge, a monster, I'm so happy, I'm
>happy, I'm so happy to be able to read S/370 machine language.  SVC-11
>hits more than the TIME macro, it rips the guts out of the PL/I DATE()
>function.  There are implications that are an absolute scream.  Rex had
>me rolling in laughter by the end of the phone call, a-hahaha-ha, CMS
>mini-disks, DATE(), hohoho, I won't give away the story. You'll have to
>contract with Rex or me.  IBM said what?  Oh-Rex, tears are rolling down
>my cheeks,  how much DASD?  Better stock up now before the DASD shortage
>hits.

Cory!

Surely such a kind and sharing individual such as yourself might have
trouble denying your c.s.y2k friends the same rolling-in-laughter
experience.

How bad is it?

Rick Plotke



Sat, 11 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7

The apars on PL/I affect DATE(), I do not beleive that they hit DATETIME();  The hit is on the linked in library modules.  Once the fixes are on, you dont need to change source (to get correct truncated 2 digit years), but you have to relink everything that uses DATE() with anything before LE/370 run-time.

These particular cases are scheduled for a "predictability" test with various combinations of compiled code and run-time modules within the next week or 10 days.  I need to either spend 24 hours backing up and restoring 20 or 30 GB of data, or get some more disks so I can have a complete copy of OS/390, the "Rex's compiler and run-time collection" and the test cases dedicated to a system image with the date set to January, 2000.

WARNING, WARNING Will Robinson.....  BLATANT COMMERCIAL AHEAD!!!

But no big deal, the Portfolio analyzer will find all of the affected load modules with the PL/I DATE() gotcha and generate the needed relink jobs to fix it in about an afternoon for the average 50,000 load module shop.  

Or you can just use the output of the scan to find all of the PL/I generated csects that reference DATE() and go in and fix them correctly.
While the scan it going, it will also tell you all of the assembler modules that have SVC 11's in them waiting to foul up your new century as well.

HTTP://WWW.EDGE-INFORMATION.COM has some details (now if I could just get Michael to load the SVC 11 examples to the web page).

Rex Widmer
Builder of software archeology tools and other strange programs to help survive in a legacy based world.



Sat, 11 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7


Quote:

> <snip>
>>The SVC-11 problem is bigger than I thought.  Rex explained the problem,
>>it's huge, a monster,  I'm terrified, then we talked about the
>>impact on consulting rates, It's huge, a monster, I'm so happy, I'm
>>happy, I'm so happy to be able to read S/370 machine language.  SVC-11
>>hits more than the TIME macro, it rips the guts out of the PL/I DATE()
>>function.  There are implications that are an absolute scream.  Rex had
>>me rolling in laughter by the end of the phone call, a-hahaha-ha, CMS
>>mini-disks, DATE(), hohoho, I won't give away the story. You'll have to
>>contract with Rex or me.  IBM said what?  Oh-Rex, tears are rolling down
>>my cheeks,  how much DASD?  Better stock up now before the DASD shortage
>>hits.

>Cory!

>Surely such a kind and sharing individual such as yourself might have
>trouble denying your c.s.y2k friends the same rolling-in-laughter
>experience.

>How bad is it?

>Rick Plotke


OK, here's the story as I got it from the master himself, Rex.  When the
year rolls over to 2000, 'DATE();' returns unpredictable garbage.  The
good news is there's a PTF.  The bad news is the PTF is against code
linked into the load module.   The good news is Rex's EDGE Tool will
find the useage of 'DATE();' in the load modules.  The bad news is, you
have to find all the load modules.  Not too hard a job in an organized
small MVS shop.  A little harder in a disorganized large MVS shop with
production code hidden all over the place, staged on backup mass store
devices.  Very hard in a VM/CMS shop with secret pockets of PL/I load
modules tucked away on user's minidisks.

Where I consult, we have minidisks that belong to users that haven't
been around for years.  We autolog their virtuals and run
disconnected production jobs as if they were still there.

Like all these problems, it's simple until you try to do it.  To dam a
river, build a dam.  How simple.  To fix 'DATE();', apply the PTF.
Well, no, since we have to hit more than just the PL/I transient
library, since we have to find every occurance of 'DATE();', it becomes
a big job.  Just how big depends upon the size and nature of your shop.
I know places where the technical staff has been rightsized and I don't
think they can find all occurances.

My assessment, this is real bad. Or, as a consultant who will be happy
to run Rex's tool for hire and can relink the PL/I, it sounds fine.  I
hope people start soon though.

Cory Hamasaki      http://www.kiyoinc.com
HHResearch Co.     OS/2 Webstore & Newsletter
REDWOOD    



Sat, 11 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7


Quote:

>> happy, I'm so happy to be able to read S/370 machine language.  SVC-11
>> hits more than the TIME macro, it rips the guts out of the PL/I DATE()
>> function.  There are implications that are an absolute scream.  Rex had

>Does the PL/I DATETIME() function survive? I'd also be quite surprised if
>LE370 date functions take a hit.

-----snip-------

Hmm, I thought I found an SVC 11 not only in IBMBJDT1 (DATE())
but also in IBMBJDD1 (DATETIME()).  This does not mean that
Y2K is a "problem" with DATETIME, just that it uses the SVC interface.

Cheers,
Greg



Sun, 12 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7

[text snipped]

Quote:
> When he asked why they didn't
> have range checking and validation of data in their programs, their
> response was "Pallat had a lot of that in his programs.  It didn't seem to
> do anything, so we took it all out."

> Sure.  Why bother with fuses and circuit breakers.  They never seem to do
> anything either.

Cannot believe it. Cannot believe it.  Absolutely dumning down.  Reminds me
of when I was talking with a programmer who related a story of someone who
modified a few lines of code, and, because it was only a few, implemented it
without testing it.  Arggghhhhhhhh!!!!!


Fri, 17 Sep 1999 03:00:00 GMT  
 DC Y2K Weather Report # 7



Quote:

> [text snipped]

> > When he asked why they didn't
> > have range checking and validation of data in their programs, their
> > response was "Pallat had a lot of that in his programs.  It didn't seem
to
> > do anything, so we took it all out."

> > Sure.  Why bother with fuses and circuit breakers.  They never seem to
do
> > anything either.

> Cannot believe it. Cannot believe it.  Absolutely dumning down.  Reminds
me
> of when I was talking with a programmer who related a story of someone
who
> modified a few lines of code, and, because it was only a few, implemented
it
> without testing it.  Arggghhhhhhhh!!!!!

Got another one from that same group.  When the people who were doing the
evaluations couldn't get some code to run without bombing out, they asked
the group what kind of tests were run with the changed code.  Their answer?
 "It compiled without any errors, so we assumed the code was OK.  I'm not
kidding.

Remove the  '.'   from orion.data for sending email to me.


Orion Data Systems

Solicitations to me must be pre-approved in writing
by me after soliciitor pays $1,000 US per incident.
Solicitations sent to me are proof you accept this
notice and will send a certified check forthwith.



Sat, 18 Sep 1999 03:00:00 GMT  
 
 [ 9 post ] 

 Relevant Pages 

1. DC Y2K Weather Report (ALC 9, power, Time Machines, How Bad)

2. Available: DC Y2K Weather Report #17

3. DC Y2K Weather Report #3

4. DC Y2K Weather Report # 4

5. gatherwrite Re: DC Y2K Weather Report # 4

6. DC Y2K Weather Report # 7

7. Available: DC Y2K Weather Report #17

8. DC Y2K Weather Report # 7

9. DC Y2K Weather Report #6

10. DC Y2K Weather Report # 4

11. DC Y2K Weather Report (Real GPS, ICBMs, ALC, 250K grant)

12. DC Y2K Weather Report # 11

 

 
Powered by phpBB® Forum Software