How to have presistent physical constants?

Hello,

I know Matlab can do pi. I would also like to add values for other constants such as electron charge, Boltzman constant, etc. and simply have them accessible as ec and kb. However, I don't want them to get wiped out when I do a "clear all" command, so simply defining them in startup.m isn't what I want. Basically, I want them to be treated just like how Matlab treats pi (i.e. they still exist even after I clear all variables). Is there a simple way to do this?

Thanks

[per isakson]

See Fundamental physical constants by Jarek Luberek

Answer 1

[James Tursa]

If you want to do it the way pi is done, then create your own function files. E.g.,

% File ec.m on the path
function result = ec
result = 1.6021766208e-19;
end

and separately,

% File kb.m on the path
function result = kb
result =  5.670367e-8;
end

However, as Walter mentions, it might be better to package all of your constants up into a single variable (e.g., struct) that can be imported via a single function.

[beanbag]

Thanks. I only had a few constants, so I resorted to doing this. But if I need more constants, is there a more compact way to bundle them all together, without having to reference them in a "something.const" format? i.e. keep the variable name super short, like ec and kb.

Thanks

[Walter Roberson]

Create functions for them. You could put all of the functions into one directory and put the directory onto your MATLAB path.

Answer 2

[Walter Roberson]

pi is a function. It even takes arguments.

Some people prefer to use a Package to store constants and import from the package

[per isakson]

"It even takes arguments." &nbsp Not R2016a - afaik

>> pi(3)
Error using pi
Too many input arguments.

[Walter Roberson]

I thought I had once used pi('single') . Perhaps I misremembered.

[Steven Lord]

I don't think pi has ever accepted an input argument. You may be thinking of Inf, NaN, ones, zeros, etc. which do allow you to specify just a class name to obtain an appropriate scalar. Different functions accept different class names. For example if you try to call NaN('int32') it will error but ones('int32') will work fine.