[Info-vax] The Gender Fluid IT Crisis

[Bill Gunshannon]

On 7/29/2017 8:20 PM, Arne Vajhøj wrote:
> On 7/29/2017 7:33 PM, seasoned_geek wrote:
>> On Saturday, July 29, 2017 at 5:26:56 PM UTC-5, Arne Vajhøj wrote:
>>> <quote>
>>> IT workers put in millions and millions of hours fixing things
>>> incompetent management was too greedy to fix.
>>> </quote>
>>>
>>> Guess what. In most companies the IT workers do what management tell
>>> them and Y2K issues got fixed because management told iT workers
>>> to fix them.
>>
>> Spoken like an academic without even one day of production maintenance.
>>
>> They got fixed because the ratings agencies threatened to downgrade
>> credit ratings and issue sell ratings on the company's instruments.
>> Then and only then did the criminal fraud of right-sizing reverse
>> costing both the public and shareholders millions in value. Most
>> large companies were staffed well enough before the Gartner lauded
>> criminal fraud to have handled Y2K at a slow and steady pace and
>> management was told this before they started on the slash and burn
>> for this quarter's bonus check.
> That may or may not explain why management ordered the IT workers
> to fix Y2K.
> 
> But it does not change the fact that you claim was simply wrong.
> Management did fix it.
> 
> #<quote>
> #For COBOL programmers with extensive copylibs this won’t be much of a
> #problem. COBOL never had the concept of a boolean, or at least it didn’t
> #in the early standards versions. All you had was PIC X or 9 and a bunch
> #of 88 levels.
> #...
> #You poor bastards using a database though. How many of you have that
> #stored in a boolean column? How many stored procedures are you going to
> #have to try and find both in production and development? Not just in the
> #database, but where they are actually called, tracking back to fix every
> #data type along the way?
> #
> #You poor bastards using BASIC, FORTRAN or some other language where some
> #fool decided bit-flags should be used to store boolean values. That’s
> #all over your system. You cannot even hope to search by name through the
> #code because not everyone will have used the named mask.
> #</quote>
> 
>>> Crap.
>>>
>>> Practically no one use BOOLEAN for storing sex even if that
>>> data type is available. BOOLEAN is for storing TRUE or FALSE
>>> not anything with two values.
>>>
>>> ENUM, CHAR, INT is what is used for what is truly
>>> an enumeration. None of these have problems with more
>>> than two values.
> 
>> Once again spoken like an academic without even one day of supporting
>> production code. Many of those very same systems with Y2K problems
>> using those 3GL languages stored Sex and a rash of other things in
>> bit-flag fields.
> 
> So you admit that the your original claim of BOOLEAN being
> used for sex was crap and you are now trying to make the claim
> that BIT was used?
> 
>>                  They did it for the _EXACT_ same reason, storage
>> wasn't cheap. One single longword with each bit potentially masked
>> off to mean something. I say potentially because bit-flag fields tend
>> to get added to records in groups when you run out of existing
>> bit-flags. That was done so there was only one conversion of existing
>> records. Usually added 3-4 at a time. Sometimes the unused longword
>> fields meant for bit-flags would get consumed by some other data
>> need.
> 
> Many things are possible.
> 
> But the combination of storing year in 2 digit text form and sex
> in 1 bit does not seem very likely.
> 
> Arne
> 
> 
> 
> 

In all the years I did COBOL SEX was a Character, PIC A (or
usually just X).  If A that gives you 26 possibilities with
X even more.

bill