On Wednesday 14 December 2005 13:52, Martin Vlietstra wrote: > Pierre, > > You have two distinct pieces of information - a sum of money and a > qualifier of what the money will do. If your data can be separated into > fields, you should break your data up into two separate fields - one for > money and the other for "base number of units".
You seem to have misunderstood the question. In the metric system, the prefix symbol M means a million. Using M for a thousand is, to me, wrong, unless you are writing Roman numerals, which this program does not use. The old program doesn't even look at the M - it stores numbers as character strings, and allows any garbage to be put in what should be numeric fields. I am currently converting the data and haven't started on the new program yet. If I see 14M in the old data, should I: *write it to the data file as (14000 00001300), and output it as 14M? *write it to the data file as (14000 00001300), and output it as 14k? *write it to the data file as (14000 00001000), and output it as 14000? The second number is in hex and is a code in a proposal I'm writing at http://phma.hn.org/Metric/unit-codes.html . I started designing this code for this program, but then decided that it could be used widely in databases that hold (possibly disparate) measurements. The other question is: The code for items is 00001000. Should the code for per item be: *00001000, since they're both dimensionless? *00003000, to allow it to have a different symbol while still being dimensionless? *xxxx1000, where xxxx is a number I haven't allocated yet? I'm pretty sure 00001000 is wrong, but I'm not sure what's right. As to the original programmer, I still have to ask him some questions. He wrote this fifteen years ago, and I'm sure he's learned a lot since then. phma
