Re: Octave vs Scilab

[Fotios Kasolis]

--------------------------------------------------
From: "John W. Eaton" <address@hidden>
Sent: Thursday, October 07, 2010 12:12 AM
To: "Fotios Kasolis" <address@hidden>
Cc: <address@hidden>; <address@hidden>
Subject: Re: Octave vs Scilab

> On  7-Oct-2010, Fotios Kasolis wrote:
>
> | However, my point is not to find a good implementation in that case, but
> | instead, to compare the interpreter.
>
> Yes, but everyone knows that for loops are slow and should be avoided
> in array languages, right?
>
> jwe

What makes loops slower in one interpreter compared to another? Or what kind 
of optimizations could be applied to get a reasonably fast/slow interpreted 
for loop? Is it the loop that is slow or accessing the elements (parts) of 
the matrix/vector in an efficient way is not possible?

Sth else I wanted to ask... when you make an assignment in Octave does it 
know (or assumes) a type (default is double I guess if numeric and not 
returned by boolean ) internally to that variable? so does it create a table 
with what variable is of what type internally?

/Fotis