Gary North site reported today that 81% of vendor supplied software is not compliant. This is outrageous. ALGORE could have fixed this easily back in 1994 after he invented the Internet by getting Congress togreenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
pass a bill requiring that all software shipped after 1-1-1995 would be fully compliant or the customer had the right to withhold payment until the bugs were fixed. Why didn't the government impose some minimum date standards such as month date year even if only 2 digits? It is ridiculous that 5 patches are required to fix date problems and it still is not fixed. We need more emphasis on quality and less on profits by Mr. Gates and other similar officials. The poor programer is told to meet unrealistic dates so the product can ship. He is under pressure to finish and was not even well paid for his efforts.
-- Tom (Notstupid@good.idea), April 16, 1999
" Why didn't the government impose some minimum date standards....."
Hopefully you're kidding. Who wiped behind your ears this time last year?
-- less,not more (firstname.lastname@example.org), April 16, 1999.
I've read this a few times, and have some questions about it:
1) How serious is the noncompliance? Is it cosmetic, or nonfunctional?
2) Is this 81% of different packages, or 81% of all shipped vendor software? If the latter, then the 'noncompliant' Windows 9x versions account for most of it, and those are cosmetic bugs.
When I got a master's degree in public affiars (bureaucracy), I noticed that all my classmates always agreed that the solution to every problem was to pass a law and tax the public to set up an agency to administer that law. That's why they were there in the first place (I was there cuz I liked a girl in the class. When we graduated, she married someone else). Anyway, that's the wrong way to go. Besides, your 'cure' would have the effect of putting all vendors out of business overnight. How is this an improvement?
I can't guarantee that my code is bug-free. I can guarantee that it passed every test I could dream up, or that it met the customer's specified acceptance criteria (some of which are pretty damn stringent). But bugs happen later anyway.
-- Flint (email@example.com), April 16, 1999.
--really wish I had the link, maybe someone can help me here--i read awhile back that at some past date, our benevolent gov actually REQUIRED a 2 digit year in any software it was buying, well beyond that "it's too expensive cuz of storage" time frame. Anyone remember, verify or discredit that? Only throwing it out, I have no verification, just one of those little y2k things ya read..........................
-- zog (firstname.lastname@example.org), April 16, 1999.
wasn't it Greenspan that argued with the .gov over the 2 digit date? Didn't he want it to be 4. I think that is where I read it zog, in a statement or report of him. And if vendors wanted to do business with the .gov they had to conform to the 2 digit date. Most vendors do not like to have multiple software packages for the general public and then the .gov. Public sector software is more like the .gov than the general public software.... at least for accounting and the like... ask me how I know :(
-- (email@example.com), April 16, 1999.
zog - the orignal standardization on a two digit date was done in the mid to late '60's (hah!) by the defense department. Prior to that time there was no standard - in fact some early DOD software actually used One digit for the year. Because DOD standardized on two digits, and then forced all of their vendors to comply with the standard, over time it became the defacto standard across a large part of American industry (ever hear of the military -industrial complex). Also, because DOD managed to develop standards before most of the rest of the government, they had a tremendous influence on other parts of the government which developed their own standards later
-- Arlin H. Adams (firstname.lastname@example.org), April 16, 1999.
You're correct Arlin. I learned this only recently. And here all along, I thought it was some stupid IBM mainframe engineer. It turns out that he was only following orders... <:)=
-- Sysman (email@example.com), April 16, 1999.
Arlin is right zog,
the full story is in the Vanity Fair article a while back - there were a couple of threads on this - worth a read to see how the pentagon are responsible for this mess worldwide - makes one wonder... hmmmm.... why DID they INSIST on that date standard.... this one has been bugging me...
-- Andy (2000EOD@prodigy.net), April 16, 1999.
---yes, that vanity fair article rings a bell, must have been initially where I saw that little winner factoid. And being such an incredibly stoopid decision, I'm left with it was done on purpose. And yes, being an ole timey conspiracy freak, I guess I'll assume a long range plan there, too by the dark forces goons, whoever they are. I'll check next tri-illuminated builderburglars luncheon. Real funny when you get goons like princely chuck et al of the UN/NWO crowd bandying around phrases like "useless eater" etc. Anyone catch the latest Dr. Len Horowitz interview on UnReal or the radio?
-- zog (firstname.lastname@example.org), April 17, 1999.
caught snippets of horowitz - plan in catching up on all this om my days off - crazy as it may seem i think it WAS done on purpose - the Pentagon overruled a committee that had been trying to force through a 6 digit date standard for 3 YEARS!!! they are renowned for long range planning, even inter-generationally...
-- Andy (2000EOD@prodigy.net), April 17, 1999.
flint, and others. the link to the new computerworld article is easy to find. gary north posted it april 16, and even when it scrolls off the new links, you can easily find it under the heading "compliance"
-- jocelyne slough (email@example.com), April 17, 1999.
I read the article. My two questions weren't answered.
-- Flint (firstname.lastname@example.org), April 17, 1999.
Gary North had an article last year describing Bob Bemer's arguments with DOD in the late 60's or early 70"s about this. He fought for a 4 digit year and DOD turned him and others down insisting that the two digit year was what they were going to go with.
-- Charlie (email@example.com), April 17, 1999.
Zog and others,
DOD was pushing two digits in the sixties because they were still smarting in the derierre over the one digit year thing with systems they had bought in the late fifties and early sixties. By the mid- eighties, activities purchasing large DOD computer systems were "being encouraged" to go to four digits.
An "etched in stone" rule requiring four digits could only come from the National Institute if Standards and Technology and they weren't going to make any changes until the early nineties. I still don't know if they've made four digits a US National Standards item.
For what it's worth, I was involved in one acquisition beginning in early 1985 and four digits were part of our design requirements. What seemed like a simple issue turned into a can of worms. Even though the applications software was all written with four digits and the operating systems was modified also, the system failed our rollover test.
Seems that no-one realized that the Real Time Clock chipsets used in the mainframes we were buying couldn't handle rollover to all zeros, no matter what the software used. October 10, 1987 was my big intro to the embedded systems portion of Y2K. And it's gone downhill ever since, cause everyone seems to want to go learn the lesson themselves the hard way and not accept what those of us who've seen it have to say.
-- Wildweasel (firstname.lastname@example.org), April 17, 1999.
Wild weasel Your mention of a problem with the system clock in 1987 when attempting to go with a 4 digit year confirms what Tom suggested at the top of this thread. The recent thead on secondary clocks by Bruce Beach also discussed this issue. Gary North said that his post on the Bruce Beach secondary clock issue was the most important thread he had posted. Apparently there were no standards or inadequate ones. At least the machine tool industry established 2 main standards for bolt threads ASE and metric. It sounds like the computer industry keeps reinventing the wheel and the technical issues like this are not taught to other programmers and designers and errors based on this lack of knowledge are repeated over and over. If the clock issue caused problems in 1987, the problems will be many times worse in 1999. It is unfortunate that the industry did not listen to you and others who tried to issue the warning.
-- Steve (email@example.com), April 18, 1999.
Here is a link that will answer some questions:
-- Ray (firstname.lastname@example.org), April 18, 1999.
Here is the link to the Gary North article: