Ralph Alvy describes a technique that would probably work for you that eliminates the auto-increment with a self-produced methodology in his book which is here online somewhere. At the moment I don't have the book handy and I can't for the life of me remember the name of it (it's early for me on the east coast, coffee needed) but later today when I'm back in the office I'll get it for you if no one else has spoken up about it by then.
Don *Don FriedmanProfessionalRecords.Com LLCPRS Data Systems* *205 S Main StreetPittsburgh, PA 15215412-784-1600 - 1-800-PRS-FILE 412-784-1615 Fax* On Mon, Apr 7, 2014 at 8:00 AM, Jacques Gaijin <[email protected]> wrote: > Hi All, > > It was from 1986 thru 1993 that I programmed a comprehensive application > in DataPerfect (appr. 80.000 lines of code, comprising 30 interrelated > panels) which administered the complete workflow for an advertising > photography and or movie production studio. A last set of modules that was > added to the application, consisted of an automated bookkeeping add-on, > that would function also as a stand-alone. > > At that time, believing computers would not run any faster than they did, > I programmed each relative link thus, that to obtain an absolute unique > record in the indexes, it had to generate a hidden date and timestamp at > the creation of each new record. When closing the book-year at years end, > through the printer level, the data would be changed as to block the > original data input. This panel replicated the original (separate) > data-input panel and could also be used to recover old data whenever > something would corrupt the financial files. > > Until now the application has continued to work flawlessly. But with the > speed of computers these days, the system does not work any more when > reconstructing the records, because within one second many records can be > generated carrying the same time stamp, thus corrupting the index(es) for > those newly generated records. Reprogramming this time stamp to a > <FORMAT:~GZZZZZZZZZ9::IH~> is not a good option, given that over time > millions of numbers must be generated; all being unique! Thus I kept one > old computer with an AMD 386 40MHz. processor. Having that system processor > running at half speed, would be sufficient to keep generating records at > less than one record a second, as to comply with each record keeping a > unique value in the indexes. > > With this sole computer becoming older and thus more unreliable, Something > must be done to keep the system up and running. I have tried to find an > answer in forcefully slowing the clock speed of present day processors, but > nobody has an answer to that so far. Another option would be a math formula > (applicable to DataPerfect DPIMP compilation) that could generate unique > coding (either numbers, or a combination of numbers and letters). > > Does anyone have a suggestion? > > Jacques > > > _______________________________________________ > Dataperf mailing list > [email protected] > http://lists.dataperfect.nl/cgi-bin/mailman/listinfo/dataperf >
_______________________________________________ Dataperf mailing list [email protected] http://lists.dataperfect.nl/cgi-bin/mailman/listinfo/dataperf
