Does anyone have any further info on a y2k remediation tool from a company called Data Integrity Inc.? There was a glowing article in USA Today on 7/7 stating that this new tool cuts debugging time by 80%. They gave an example a guy who debugged 100k lines of code that would have taken a month in one day. I realize that 'silver bullets' will never be available for this problem (especially considering the embedded systems aspect) but is this product a cause for optimism? Thanks for any insight Steve

-- Stephen Gould (f64@dfw.net), July 08, 1998

Answers

I heard the same story on the radio and found their web site at http://www.dii2000.com. As a former programmer I was able to understand their concept, but I'm afraid it offers little hope (for all the reasons stated here and elsewhere, embedded chips, old languages, starting too late, can't find the source code, etc.) In a shop that has well documented PL/1, COBOL, and ASSEMBLER programs (all of which I have programmed in) it will probably be a great aid. Un fortunately a lot of the stuff that is critical to our countries infratructure uses languages like RPG, PASCAL, FORTRAN, and many, many DoD specialty languages. I believe we will begin to here a lot of similar stories in the next few months all attempting to show that this or that product is the "silver bullet". Sorry. it won't happen.

-- Kevin Lemke (klemke@corpcomm.net), July 09, 1998.

Hi,

I saw some information about this in a Boston Paper yesterday and John Westergaard discusses it today at the Westergaard 2000 site. I responded to his comments by suggesting that he go take a look at Peter de Jager's comments in his article titled "Biting the Silver Bullet". This can be found at http://www.year2000.com/y2karchive.html for anyone who is interested.

There's no such thing as a Silver Bullet and there is no possibility that anyone will come up with one. The best that can be done is to create tools which will help with some of the work, in some of the situations. Rewriting the code is less than half the Y2k job (which mostly consists of planning and testing) and it does not begin to address the problems with embedded systems.

If you are one of the lucky group who are faced with a Y2k situation exactly like the ones which tools like Data Integrity's product, I'm sure that it will be of great help - it might even save you as much as 10% of the work.

Data Integrity has a White Paper which you can order by visiting their web site. The URL is http://www.dii2000.com/

Ed

-- Ed Perrault (EdPerrault@Compuserve.com), July 10, 1998.

I have looked at data integrity's White Paper and their two 'Explainers'. Unless there is something that they are not telling us, their method will only work for a small subset of date problems. and will not work all the time.

- They do not describe their search algorithm for finding dates, so it is not clear if they even find all of the instances where they need to apply a fix.

- They present their patented algorithm as an alternative to either expanding the date field or windowing, but their definition of windowing is defective, and they have to be doing a variation of windowing when they compare two dates.

- They describe their solution to subtracting YEAR1 from YEAR2 if YEAR1 is in a later century from YEAR2 - they perform the subtraction and get a negative number, then add 50, then add 50 again. This will only work if YEAR1 and YEAR2 are defined as two byte zoned decimals, and the result field is a signed two byte field, and you know that the result is supposed to be positive.

- They have no apparent solution to sorting on date fields, using dates in keys, hard coded centuries, leap year calculations, etc, etc,

-- Frances Prevas (ccd.fran@worldnet.att.net), July 13, 1998.