> In particular, I am looking at the historical
>development of the Y2K problem. Were there other
>warnings given back in the early days? How can I find
>that out? I would be most appreciative if you might
>be willing to share views and insights on these
>subjects. Thank you.
I would be happy to build up a historical presentation based on what I know and the problems I had.
My first experienced Y2k event was in 1972 (9999 days to 2000-01-01) when the tape catalog on the IBM OS systems I was working on became corrupted. i.e tape expiry was reset to zero.
The next major Y2k event was in the early 80's when IBM brought out the 4300 series and rewrote the operating systems. Year 2000 compatibility had been designed by the labs for inclusion in those releases, but the upgrades were canned by Marketing.
As a Legalist, you will understand that knowledge of events is one thing, proof is another. After the consent decree, every piece of paper that IBM produced had to be archived. I seem to remember that one of the reasons for success against the later Justice antitrust suits was that IBM had a database of all this stuff and were able to drown investigators in sheer volume of paper.
The "proof" is sitting in archives in the European software plants somewhere. Finding it or even getting to see it is another story.
I was a very junior IBM employee in those days with no access to Management decisions or documentation. My only contact with the design teams was informally, through Product Managers that I happened to meet in my travels, usually with a jug of beer on the table.
The people involved are long gone. Pensioned off, given the Golden Handshake, or Dead. The chances of taking on IBM over this are slim. Even assuming an overwhelming factual case, (a false assumption), who do the courts and/or public generally believe? A large, wealthy organisation or some indigent idiot from the remote boondocks of Africa? I already know the answer. My career and more important to me, my credibility and reputation, were ruined with the greatest of ease.
In 1986 when I wrote the "Timebomb" letter, I was fobbed off as some sort of crank. "Disgruntled employee" was the term used. The threat of the Advertising Standards Authority (it was never carried through) was water off a ducks back to me personally, but my Publisher threw a fit and banned me and my copy from that day on. That ban still applies to this day.
In a nutshell, I would be happy to help reconstruct a historical trail, but doubt the practical value. The info is "interesting" to the historian perhaps, but no one really gives a damn. And it does not fix the problems that we face now.
Cheers.
At 15:21 1998-04-30 -0700:
>I am particularly interested in your Timebomb letter
>and IBM's response. May I inquire as to how you came
>to write it? What happened in 1986 (or earlier) that
>caused you to give such a farsighted early warning?
Why did I write it? I sometimes wonder. It has caused me nothing but grief.
I think I mentioned that the first Y2k event happened in 1972 (9999 days to year 2000). I was a junior programmer at the time working on internal IBM systems, and that event caused me to lose some important data (critical tapes overwritten) and I had to do a lot of work to recover. It proved to me that the problem was real. I found the IBM solution, ie. use 999 days instead of 9999 days retention to be shortsighted, to say the least. However, that event prompted IBM to change retention date processing and prompted the introduction of 99/365 as the "keep forever" date in JCL.
I later moved from the internal IS department to work in the field. My customers were small Manufacturing companies. At that time there were no tape cataloging systems available for DOS/VS or VM/370. So I decided to write one as a private project. I then discovered major problems with the internal control blocks within the Operating System. These could only be fixed by the Developers. Control blocks are at the heart of the system.
I made contact with one of the product developers who was visiting South Africa from Germany (IBM System Software was built in Germany in those days - probably still is). He advised me of changes that were in plan for the new releases being built to support the new 4300 series.
When DOS/VSE Release 1 and VM/SP Release 1 came out, I was horrified to find that the promised date support had been removed. There was one residual (useless) macro left over from the code that had been ripped out, SYSVDATE or something like that, but the control block changes were taken out. I had the opportunity, in the course of training as a 4300 specialist, to visit the World Trade Support Centre in Poughkeepsie, and various plants in Italy, the UK and Germany. My contacts told me (over mugs of beer) that Marketing had canned the proposed date changes. No reason given.
As it was a career threatening activity to go up against Marketing in those days (probably still is) we all left it that.
I did some lobbying internally to try and stimulate interest in having changes made, but with no effect.
I left IBM in 1981 and became an independent consultant, continuing to work primarily with Manufacturing clients.
I eventually became involved as Technical support for Forman, an MRP product written by Formation Inc of New Jersey. Now Forman internally supports a (windowed) date range of 1970 to 2069. But I discovered major problems in acquiring the date from the Operating System (which of course is the Classic Y2k Problem). The application would run, but the Operating System would fail.
I engaged in some discussions with IBM and got some vague indications that they would look into it. This was done at fairly high level with Tony Dry, later MD, who had been my Marketing manager and to whom I had passed various sales leads in the course of my consulting work. He assured me that the Country SE manager Selwyn van der Veen would take it up with Paris. There may have been some conflict of personalities involved at this stage as Van der Veen and I had crossed swords some five years previously on my departure from IBM. As he said to me at that time, "You really go for the jugular don't you". B>)
So I was getting nowhere fast and finally felt that a major "stirring of the pot" was required.
So the infamous "Open Letter" and the concept of the Deadline 2000 User group were born. In those days we did not have the Net. We had bulletin boards such as Fido (which is still going), but they were inhabited by Techies. I discussed the issue there briefly and had my first encounter with denial. Nobody could care less. So I decided that I had to "go public".
The original text of the letter was designed to be published in Datamation. Until I found out the cost of a half page ad. So I published in a local industry rag "Computing SA" instead. It cost R2000, a considerable sum for me in those days.
If you read the letter today, you would think "Ho hum, boring", but in 1986, IBM were at the peak of their Monolithic power, and it was purest heresy.
I had no idea that such a vitriolic response would be forthcoming. If I had realised that I would be personally attacked rather than involved in a rational debate of the issues, I would have dropped the project there and then.
The IBM reply contains a mixture of truths, half-truths, prevarications and outright lies. I have never seen such a strong response in any IBM public communication. Obviously a nerve was touched.
As my publisher dropped me like a hot potato at mention of the Advertising Standards Authority I was never able to reply to IBM. My "Letter to the Editor" in rebuttal of the IBM reply was never published.
IBM implied that I never discussed the issue with them before publication. Maybe not in writing, but I had personal and telephonic interviews at the highest level. Much was made of my use of the word "congressperson" as opposed to "member of parliament". Based on the fact the article was originally destined for US publication I cannot understand why this was such an problem.
>How did you know what was going to happen?
Because I read the system source code. I had a need to use 4 digit years and went to see where the system would not allow me to do this. The "proof" is the source code. It just does not do the job. This is exactly what IBM had to do in 1995 when they finally woke up to the fact that the Congressional hearings were serious. I just happened to do it earlier. Because of my communications with IBM and their lack of response I could predict with fair certainty what their reaction would be. ie. nothing much.
IBM did in fact make several unpublicised changes to date internals in 1987. Not enough to fix Y2k, but enough to get us past the 999 days to 2000 hurdle. I like to think that my work contributed to that decision.
>Did others in the industry know?
In retrospect it seems there was some activity going on. The ANSI x30 standard on 4 digit years came out around then. The ISO8601 standard came out around 1987. The FIPS standard (4 digit years) is from that era. Obviously many people were working independently on date related issues. I heard that somebody had taken up the issue with the COBOL standards group and had been rudely rebuffed. Apparently there were articles and even an IBM journal discussing some of the implications of dates. But I have not seen these, so cannot comment.
>Was the Y2K problem common knowledge?
Within limited highly technical circles, I would say yes. But it was not a hot issue. In internal discussions with overseas developers on developments on the internals of forthcoming OS releases, (within a group of about 30 systems engineers) I was the only one interested in the topic. I gave several talks to groups such as GUIDE, so there was an awareness within the User community as well. Nobody was interested. The assumption was that "somebody else will fix it in the next 14 years".
> Was anyone else speaking out about the problem?
Not publicly. Well, at any rate, I have no knowledge of anyone doing it. The first public acknowledgement of the issue was the result of Peter de Jagers article in Datamation (1994?)
> IBM's written response to Timebomb smacks of
>intimidation. Was there a cover up going on? Why
>was the industry not addressing the problem in the
>manner suggested in Timebomb? Did anyone else
>other than you?
IBM says (in a letter to Computing SA dated 21 July 1986) "IBM and other manufacturers have known about this for years". I don't know why there was such a "bad" reaction. I (foolishly) thought that Action would be taken to fix the problem once it was made public. I didn't understand then about Denial. I didn't realise that there might be vested interests with reason to suppress this information. Presumably the issue was generally considered as very low priority. That attitutude persists in some circles to this day.
Why it was all dismissed so lightly eludes me. In the death, IBM will lose (and has already lost) market share purely and simply because of Y2k. Millions have been and will be wasted on something which if approached rationally and reasonably in 1986 would never have become a problem. Whether the decision was deliberate or just plain stupid is unknown to me. I suppose you could hypothesise that information was suppressed so as not to "rock the boat" or give customers any sense of insecurity. I still do not understand Why.
I am not so sure that "millions" are going to be involved in litigation. There are already US states who have legislated that Y2k cannot be used as a basis for litigation. I suspect this trend will continue.
As a matter of morbid interest Roman Dutch Law in South Africa does not award punitive damages, documentation is preferred over evidence and in general the maxim "let the loss lie" prevails. Prescription probably applies, and any remedy would have to be based on contract law or delict. So we are not expecting a Litigious frenzy.
I trust that this helps you B>)