Re: .. Y2K and databases

In article <01bc73d6$0ef89920$803ca8c2@default> "Chris Young" <c.young@virgin.com> writes: >Everyone seems to be concentrating on source code but ignoring the other >major problem - dates in the database. >In light of this I developed a tool to scan most of the major databases >(mainframe, AS/400, Access, etc.) which I am now making available as an >inexpensive utility. >"Datefind" uses ODBC to connect to the database (so works on any database >that has an ODBC driver available) and scans ALL files finding dates which >are NOT Year 2000 compliant. >A trial copy can be downloaded from http://freespace.virgin.net/y2k.html >I hope this tool saves you as much time as it has saved me. Well, Chris, your program might be useful but then again it might not be. Date/Time values are normally stored in relational databases in a binary format, e.g. timer-units since a known but arbitrary "epoch," and they are translated to and from that format during input and output. This format is not the slightest bit ambiguous as to what century it is. I am not addressing your case specifically when I say that it alarms me how many people right now are pressing mis-information on the public, in various degrees and ways, in order to obtain a quick buck. People might not understand how computer databases work, but they trust us. At first. If they find out later that this trust has been violated -- that they were taken advantage of -- then there will be unholy hell to pay. Remember that the public (and their lawyers) tend to think and react in sound-bites and to over-react in most cases. Lawyers make good money from this. An Access database, Paradox or SQL or dBase or what-have-you database, that is designed with the *slightest* thought as to data types and representations (e.g. by a professional) is simply NOT going to have any problems at all with "the data in the database," and not much (if any) problems at all with the rest of the application.