[Jprogramming] Processing 1GB+ files without chunking

[Dan Bron]

Raul wrote:

>How about ... no garbage in

That was my suggestion.

> a show-stopper resource management constraint.  

The begs the question.  Do I?  That's what I want to know.

> Why do you object to writing code to manage your resources, 

I do not object: I've already written the code.  In Perl.  

So, the only reason to involve J at this point would be if it adds value.  But if I have to write a sufficiently complicated J program that contains an ad hoc,  informally- specified, bug-ridden, slow implementation of a tenth of Perl, it does not add value.  

The value in J is in writing code the J way, and explicitly looping over chunks of a file is incontestably not the J way.  Sometimes it's necessary, just as sometimes it's necessary to write a bit of assembler into a C program, but that doesn't mean assembler is the C way.

Another way for the J program to add value would be if it were fun to write.  Writing code to loop over chunks of files and handle all the edge conditions would not be fun.

Or, to answer your question another way:  I object to writing code to manage my resources because I use J precisely to avoid that. I want to design skyscrapers, not drive nails.

-Dan