On Fri, May 11, 2007 at 07:41:10PM +0200, Florian Kulzer wrote: > > Actually I understand the concept well enough to know that it cannot be > applied stringently if one deals with numbers that are divisible by > positive integer powers of ten and which are written down as words. > Obviously, "ten" and "thirty" fall into this category. Even if you write > "30" it is not really clear if you know this value to one or two > significant digits. Sometimes people claim that "30" by convention is > one significant digit and that you have to write "30." to indicate two > significant digits, but this is, as far as I know, not generally > accepted in the scientific community. (Scientists and engineers mostly > use semi-logarithmic notation anyway, which avoids these ambiguities.) > Writing "30" to mean one significant digit and "30." to mean two significant digits is how I was taught. Of course, that may be a deficiency in my public education. > Furthermore, if you want to start patronizing other people about the > concept of significant digits then you should probably be more careful > yourself not to make statements which could be misconstrued as mix-ups > between "significant digits" and "decimals": In your earlier mail you > first give the "one significant digit" argument and then you abruptly > and without necessity switch to numbers that have three (5.00) or four > (10.00 and 30.00) significant digits. Everything you say is technically > correct[1] and it is maybe just a coincidence that these numbers have two > decimals, but at the very least this is unnecessarily confusing. What is > so special about four significant digits when two significant digits are > in fact the threshold for putting "ten" times pi out of range for > "thirty"? > There is nothing special about choosing three or four significant digits. There are, of course, three significant digits in 5.00 and four each in 10.00 and 30.00. It was coincidence that I chose them like that, not intending to be confusing. BTW, I was not patronizing anyone. > [1] You avoided stating how many significant digits these numbers have > in your opinion and, maybe by pure luck, you chose a formulation > which left you enough wiggle room to use more significant digits > than strictly necessary. Being vague enough so as not to be wrong is > not the best way to demonstrate your understanding of a concept, > though. > I did not say how many I thought because there is no question about the number of significant digits in a number with a decimal point. It is only when you have trailing zeros to the left of the decimal point when the situation is ambiguous. This comes from people either being taught incorrectly, being taught something different from the common scientific usage or simply forgetting. Regards, -Roberto -- Roberto C. Sánchez http://people.connexer.com/~roberto http://www.connexer.com
Attachment:
signature.asc
Description: Digital signature