It seems the airline reservation system is at least a partial success story. Does anyone know (or have any links) if they remediated the current system, or replaced the system completely? It seems to me that (if they started early enough) that a system replacement has a much better chance of success then remadiationg old code. They real key is *if* they started early enough. Think I read somewhere that the FAA will have a new system installed in November, and THAT could be "just a bit of trouble"...lol

-- Online2Much (justcurious@themoment.com), February 06, 1999

Answers

I really think that you will see lots of honest (but localized) "success stories". Quite a bit of software either *is* fixed or *will* be - I'm not being sarcastic here, either.

The problem to watch out for, if this is your focus are:

* Lot of little software problems appearing at once.

* Fractional Reserve Banking - If only a small (4%?) of a bank's customers withdraw funds, the bank goes bankrupt.

* Countries that cannot afford to remediate, but still are critical suppliers to bigger countries.

* EMBEDDED CONTROLLERS!

I'm still banking on a 4-6 in the 0-10 scale, but who knows? This is one time I truly hope I am a pessimist, and everything will be "just fine".

-- Anonymous99 (Anonymous99@Anonymous.com), February 06, 1999.

I used to work for one of the big three CRSes (Computer Reservation Systems). I left about 1 1/2 years ago.

When I worked for them, I did some Y2K stuff in early 1997. Mainly reviewing code I was responsible for, and putting Y2K dependencies (my code had no Y2K dependencies, it merely counted seconds for timer expiration (not using a date function, either), but the computers using my software had non Y2K BIOSes).

The "main system" is fixed. This is what the big CRSes are pushing out to the press. And it probably is for several reasons.

1. Many years ago, the company become VERY ANAL about source code. In this case ANAL is a good thing. This means that changes are controlled in the system and that all programmers must "check in" their source code into a "source library". The company spends a TON of money and about 20 full time staff personnel to keep this source library reliable, secure, safe, and backed up.

2. Most of the code in CRSes is custom code. We wrote it, we have the source code. So we can fix problems by running source code scanners and such to find potential problems, rather than running in an endless test/fix cycle, like you have to do with other people's code (where you don't have source code).

The reality for this company (and I believe the reality for most companies) is that there are literally millions of computers that have to be remediated, though.

Mainframes are the easiest because "formal" source code libraries exist (in most cases), and because the computers cost so much, formal change procedures exist which manage the change through the computer system.

Client/server and PCs are another matter entirely. Most of the code comes from software companies. The ethics of the companies are less than desirable. Source code for "custom" products is often lost or destroyed. And you have the situation I'll describe in the next paragraph.

Contrast this to the organization I just left. They buy most of their code. The vendor has a Y2K version, but that Y2K version is dependent on a Y2K release of the database and operating system. So, for my prior org's system to be Y2K, they have to upgrade their operating system, their database (now 32 gig), and their applications software, all before the end of the year.

In addition, they have tons of embedded system which will have to be certified Y2K before the end of the year. Most of which, I believe, won't get done and will be "fix on failure".

The key point here... Client/server and PCs did the Y2K problem no favors. Code is spread everywhere, and will be a nightmare to find and change.

My only hope is that when Y2K finally washes out, that we'll go back to centralized source code bases, careful change management and reinstate some of the "good" things we put in place on mainframes for the client/server and PC environment.

Which is why I think you're seeing Intranets play an ever increasing role in running companies. After Y2K, I believe EVERYTHING will be on the Intranet, with centrally managed servers running it. The companies which survive will have learned their lesson.

Glen Austin

-- Glen Austin (gdaustin@aol.com), February 06, 1999.

-- a (A@a.com), February 13, 1999.

-- a (a@b.com), February 13, 1999.

-- a (a@a.com), February 13, 1999.