[Scilab-users] Accidentally displaying huge matrices

Samuel Gougeon sgougeon at free.fr
Thu May 28 02:46:02 CEST 2015


Le 27/05/2015 14:15, Vincent COUVERT a écrit :
> Hi all,
>
> What would you think about a feature describe as follows?
>
> We coud add a specific calling sequence for lines function (e.g 
> lines(-1000)) that will switch automatically the display of Scilab to 
> the "short" mode [1000x1000 constant] for a matrix (or hypermatrix) 
> having a dimension greater that 1000.

I agree with Serge that my initial proposal using existing lines() flags 
and just adding a sign convention is not good, for the reason he 
explained, and also because such a convention would be closed, would 
prevent any extension for further paging or overall formating modes.
But i disagree about extending format() instead, because presently 
format() works only on formating numbers (for disp() and string()). This 
would turn its role confused.

I would rather advice to keep lines() to implement new paging modes -- 
because it is already its job, not the format()'s one : Scilab's 
format() is not Matlab's one --, but to extend its size.
Hence, L = lines(); L(3)=100 could be a new flag thresholding 
full/abstract modes.
In the same way, a non-wrapping mode could be tuned through L(4), 
instead of highjacking L(2) with any sign convention.

Finally, if a compact output mode without extra blank lines should be 
implemented & switched, lines() would also be involved, not format().

Samuel

PS : format() could also be extended, in the first place for instance to 
replace xset("fpf",v) -- that has still no set() equivalence ; but this 
is another discussion.




More information about the users mailing list