Re: [Bug-apl] [PATCH]: allow using lambdas in ]USERCMD

[Blake McBride]

On Fri, Feb 24, 2017 at 10:37 AM, <address@hidden> wrote:

all the apl2 based modifications would have to be taken out ... it has really become blake-apl2

blake-apl2 === IBM APL 2

If you look at some of my early posts, you'll see that I clearly disagree with some IBM APL 2 decisions.  In the interest of having APL mean APL, and not some other, personal interpretation, I have capitulated.  I am now committed to the IBM APL standard as meaning, for better or worse, APL.

I remember BASIC at the beginning.  You know, GOTO line number X.  If you look at MS BASIC now, it looks more like C than BASIC.  It is some bastardized language that amounts to no more than a MS-only language.  What I am afraid of is this.  In the name of convenience, we add more and more tweaks.  People make use of many of those new features until people with a history with APL like Peter and me can't even recognize the code anymore.

To Elias:  No one expects to create SOAP Web services using APL.  There are enumerable extensions that can be made to APL to accommodate other technologies.  That is not a good reason because you inevitably end up with a bastardized hodgepodge of a language. APL already has a solutions space, and it does it well.  IMO, what is far more important is to make GNU APL take full advantage of multiple processors rather than inching GNU APL towards a bastardized nothing.

The computing world has an increasingly greater need for convenient parallel processing for reason we all know.  APL excels at that notationally.  Rather than spending our time on language extensions, we should be concentrating on two things:

1.  Don't copy arrays unnecessarily internally

2.  Greater use of multiple processors

If these two were accomplished, GNU APL could be at the forefront of the computational tools world.

Blake

 

 

i have tried to bring up things that are wrong but get ignored - i have been using 'real apl' for over 46 years