Re: Question: How will the Year 2000 bug impact your work?

Steve Cosner wrote: > > In article <3274EB3D.7CAB@uci.edu>, Hank Liao <hhliao@uci.edu> wrote: > > > >What you said are all correct. 2-digit year display can be changed to > >4-digit year display by simply set century on, however, the problem is > >the screen real estate. > <snip> > > > >Without any modification, these 2-digit year would have problems with > >the year 2000 issue. We are currently adding a function call in the > >VALID of all date fields to call a convertion function to assign 2000 to > >years between 00 and 49, and 1900 to years between 50 and 99. We will > >need to modify the programs again around years 2045, but we will only > >need to modify the procedure to change the assignment values, and > >hopefully the programs are smarter by that time. > > > >Hank Liao > > Why not write a function that will NEVER need to be changed. It is > surprisingly simple, and here's how... > > Get the current 4-digit year, and subtract 90 from it, then extract > two values: the century, CC, and the last two digits, YY. (In 1997, > CC will be "19" and YY will be "07".) Now, compare the 2-digit year > entered by the user on a screen to YY. If the user's entry is greater > than YY then use CC as the century. If the user's entry is less than > or equal to YY, add 1 to CC to get the century. > > Using this method in 1997 returns 19 for the century for years 08 > through 99, and 20 for the century for years 00 through 07. Each > year, the algorithm advances one year. > > This "sliding century algorithm" will ALWAYS work because it starts > with the current year as its base of calculations. You never have to > re-code anything, no matter what year you are in. It forces a date to > contain a century within the range of 89 years BEFORE, and 10 years > AFTER the current year. > > By changing the 90 in the beginning of the algorithm, you can modify > the range of years assigned to each century. > > Hope this is useful. I can send the actual code written in Oracle's > PL/SQL, if you want. > > Steve Cosner NEVER??? What about year 10000??? Never say never !!