The big difference, from my point of view is that atomic
representation is slightly more formal:

There are some word formation issues which can only hit you using
string representation of the code.

And, if you have an atomic representation of a partial phrase, you've
eliminated some possible alternative atomic representations.

And, working with the atomic representation gives you a little insight
into how the interpreter 'sees' the string representations.

(Speedups usually only matter when working with large data sets, or
with particularly inane computational approaches or with real-time
contexts where you're dealing with something happening quickly in the
outside world.)

Thanks,

--
Raul


On Sun, May 3, 2020 at 7:51 PM Raoul Schorer <[email protected]> wrote:
>
> Hi all,
>
> I have been comparing manually built identical phrases (a dyadic verb
> with its arguments) and executing it with ". or `:6. My understanding
> was that the advantage of atomic representation was that it bypassed
> parsing and that the execution model was:
>
> input string -> parsing -> atomic representation -> interpretation -> result
>
> However, I was surprised to see no speedup at all when transitioning
> from string to atomic representation. Of course, my particular code may
> be the issue. What are the usual strengths and weaknesses of string vs.
> atomic representation?
>
> Thanks!
>
> Raoul
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to