Why do quotes in norm for infinity norm sometimes work

I'm just curious as to why, for some matrix A, you can calculate the infinity norm with

• norm(A, inf)
• norm(A, Inf)
• norm(A, 'inf')but not
• norm(A, 'Inf')

especially when

>> inf
ans =
   Inf

Seems strange to have included 'inf' -- and the command itself doesn't indicate that 'inf' is a valid input either:

>> norm(A, 'Inf')
Error using norm
The only matrix norms available are 1, 2, inf, and 'fro'.

Answer 1

[dpb]

"Because" :)

It appears the author of norm did go as far as to do a string comparison of the inputs but didn't bother to [upper/lower]case the input string before making the comparison internally.

The same case-sensitive comparison fails in the same fashion with 'Fro' as opposed to 'fro'

[Michael Gustafson]

I thought it might be that :) I seem to remember at some point in the past, 'inf' wasn't accepted either... At least, that's how I graded things, but a student this semester pointed out that the lower case version in quotes did work (and now I don't have any old enough MATLAB installations to check The Good Old Days)

[dpb]

It works same way as far back as R12; I don't have anything earlier installed.