Re: Octave should be more consistent in converting automatically char to numeric.

[Sergei Steshenko]

________________________________
From: Juan Pablo Carbajal <address@hidden>
To: Dildar Sk <address@hidden> 
Cc: Help GNU Octave <address@hidden>
Sent: Monday, March 26, 2018 11:37 AM
Subject: Re: Octave should be more consistent in converting automatically char  
  to numeric.



Hi,

I do not speak for the whole community here... that should be obivous.

Octave aims at running "every code that runs in matlab" without
modification. I am not sure there is a stand on code that *doesn't*
run on matlab.
So I guess it is rather a choice to realize that char is meant to be a

value in the ascii table rather than the value itself.

----------------------------------------------------------------------

'Octave aims at running "every code that runs in matlab" without

modification' - suppose there is a Matlab program checking validity of numeric 
user input.

Since the input can be erroneous, and the program shouldn't stop, a good way to 
achieve this is to use 'eval' statement and check whether it was successful 
through the 'CATCH' part of the 'eval' statement.

Under such circumstances Matlab will flag some user input as erroneous while 
Octave will accept it.

So, which of the two behaviors is correct ?

--Sergei.