question for all "real engineers" with college degrees, etc on notation.

[Squidly-Diddly]

instead of using "scientific notation" like 4.43500 x 10^^3 to give us 4,435.00 and tell us we have 6 significant digits.....

how about we write 4,435.00s.....with the 's' to show how many sig-dig?

OK, on THAT example we know how many sigs from the ".00" but....




another example....


55,340,000 how many sigs?

55,340,0s00 tells us we have 6 sigs and the last two 00 aren't sigs.

Not only is it less typing, but it would be alot easier for a computer to parse when reading text, and looks better when typed as shown here.

Works fine on non-base 10 numbers, and with or without the "," or the European "."


Any other nominations for the 'sig indicator char' besides "s"?

[ldigas]

Ugh, ... you got so many things mixed up there I don't know where to start.

First, the term significant digit. That is a digit that carries a meaning, carries a precision so to speak, and realistically speaking, you cannot know how many significant digits are there just by looking at a number.
So, to answer your first question ... if it is, let's say a length, I would argue it has 6 significant digits, since you cannot accurately measure ... that value, and still accomplish that precision.
(this can also be argued, but you probably get the point)

This is a common mistake students make, when they increase the number of decimals in Excel, and suddenly "they have more significant digits". Of course they don't ... they just have more digits on screen which
mean nothing (result of calculation, but don't add any value whatsoever).

As far as display formats go, there are several (btw, you got an extra ^ in that first number ... Fortran uses two *'s, i.e. 2.21**3. but I've never seen a program or a language that used two ^'s for exponantion. But, I digress ...

there are several display formats:
- floating point / fractional ... 3.214343, 12323.435 ... usually described with the maximum field width, and the number of decimals after the point
- exponential ... 0.123E+02
- scientific notation ... same as above, except the mantissa is at least one (but less than 10) ... 1.123e+02
- engineering notation ... similar to above, except that exponent is constrained to be a multiple of three (thousant). A non-zero mantissa is greater than or equal to 1, and less than thousant ... 12.345e+06

[Squidly-Diddly]

numbers you are supposed to only give the sigs that your instrument is able to give.

For instance, if you multiply one number with 7 sigs with one that only has 2, you only write down 2 sigs and round off...

and that is supposed to let your co-workers know how many sigs, and the main reason for scientific notation is to indicate how many sigs.

This is in Physics class in a college in Silicon Valley, CA and they tell us this is how the hi-tech companies, etc want it. They say by using Sci Notation as I show that DOES tell you how many sigs.

My way would do the same, without the "e+06"(with is better than my "10^^6") but still more chars.

[ldigas]

I have trouble understanding that last post of yours. Could you explain what is it you're trying to convey?

[gonzo]

Significant digits are the ones where the precision of the instrument can measure them. For example, if you have a micrometer that has a precision of 0.001, if you were to take a measurement of 0.017 and divide by three, the answer with significant digits is 0.006 and not 0.0056666666_

[Poida]

Gonzo you are not a "real college engineer."

You are an engineer in the "real World."

Poida

[Squidly-Diddly]

such as:

2.3453

3232.002

490.0

2330001

but NOT

240

294390000

3910

because you don't know if the last zero was a 'place holder' or WAS a sig and just happen to be a zero, not a 1-9....

and THAT is why sci-notation is prefered.

[gonzo]

Poida: can you please explain where you read my resume with college and technical school degrees details?

[tom28571]

I'm glad I knew how to do measurements before I read any of this.

[jim lee]

This is for boats right? In boat school they told us 1/8" is "good enough". So where's the problem?



-jim lee

[jehardiman]

The importance of significant figures is not in the number itself, but in the calculation. If I have a non-constant number in an equation that has 2 significant figures, the most the answer can have is 2. So I could write that the circumfrence of a circle with a diameter of 1 is 3.141592654; but the circumfrence of a hole with a measured diameter of 1.0 is only 3.1. And then we can add tolerencing...<wicked evil grin>

[Squidly-Diddly]

instead of 2.48e06 or wondering if 2,480,000 has 3,4,5,6 or 7 sigs.......

just write 248s0000 to indicate that figure and number of sig digits.

[jehardiman]

Do it the proper way, not here. Put it to the American National Standard Institue to change ASTM E29 and all the ANSI standard computer code requirements. If you can convince them that your way is better they will change it, otherwise you should conform.

[Squidly-Diddly]

Is anyone even using standard notation to any extent or is everything all in computers?

I have a suspicion they aren't telling us the full truth when they say "this is how industry wants us to teach it".

You mentioned "ANSI standard computer code requirements".

Are current computers all set up to read " 0.123E+02" as a number as well as a string of chars? Or is any program going to need to be told what "E+" etc means if it finds it in that context?

I've taken basic Pacal, Fortran, C++, Basic etc, but it has been a while.

[OverRun]

Gonzo,

In the part of the world that Poida lives in, what he wrote would be interpreted here as a real compliment. It indicates that you are an engineer of knowledge and experience as well as the theoretical learning, and thus to be treated even more seriously.

Originally Posted by Poida

Gonzo you are not a "real college engineer."
You are an engineer in the "real World."
Poida

Umm - this is an unexpected first post by me on any Forum or Board. I usually try and contribute technically first, but I thought this might make a timely contribution.
Cheers
OverRun