Re: Use of Local Variables in Lambdas (Direct Functions) ?

[Dr . Jürgen Sauermann]

    Hi,
      
      in my opinion lambdas should be as close to defined functions as
      possible
      and local variables are one aspect of that.
      
      Many lambdas are simpler with ⎕IO←0 so localizing ⎕IO is a
      frequent simplification.
      
      The primary reason why GNU APL does not support multi-line lambdas
      is that the
      multi-line lambdas of same other APL interpreters have a return
      value semantics
      (like: the first not-assigned value is the return value) that
      differs from the return
      value semantics of defined functions (like: the last assigned
      value to the return variable
      is the return value).
      
      I did not think that having different semantics for lambdas and
      for defined functions
      is a good idea in the first place and therefore I did not
      implement multi-line lambdas.
      There are also other complications when a lambda has more than one
      line.
      
      Best Regads,
      Jürgen

On 7/10/21 1:37 AM, Russtopia wrote:

Hi, reading GNU APL documentation, in the section 2.7 "2.7 Direct Functions (Lambdas)",

It states that lambdas do not support multiple expressions (amongst other features). If this is the case, what purpose does allowing local variable declarations serve?

Allowing

SUM ← { ⍺ + ⍵ ;C;D }

.. but not allowing multiple statements prevents assigning to C or D, or really using them at all, so what is the purpose of allowing their declaration? Please forgive me if I am missing something obvious here.

Thanks,

-Russ