From Newsgroup: comp.lang.apl

Dyalog Ltd. is likely soon to embark on the development of its own version of a far-reaching enhancement to APL notation.
Referred to as "Array Notation" it's based initially on the simple premise that rather than engendering a SYNTAX ERROR, an opening bracket, "[" closed on a subsequent line or after a diamond, delimits a multi-line expression that denotes a multi-dimensional array wherein each expression defines a cell of that array:
ri| numraE[1 1 2 3
5 8 3 1]
2 4
The principle is naturally recursive:
ri| chrraE[
['zero'
'one']
['two'
'three']
]
2 2 5
Over a number of years the idea was elaborated until Dyalog adopted it and continued to modify it after 2015.
There is a couple of pages in the APL Wiki that attempt to describe it.
It has been modelled in APL several times and has been implemented as a part of the language at least once.
I'm sure Dyalog would be pleased to accept suggestions and comments in their forums to hone their design. I suggest the "Language" forum at https://forums.dyalog.com/viewforum.php?f=13 would be the place to start.
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.apl

I think it is a good idea to allow native array inputs. I'm not sure why overload [] rather than use normal (). I've done this for a while (2004) in my compiler. I like the fact that the input is the same as the output with this approach (for numbers and char strings).
.rho #.is num .is ( 1 1 2 3
5 8 3 1 )
1 1 2 3
5 8 3 1
2 4
.rho #.is array .is (
1 1 2 3
5 8 3 1 )
1 1 2 3
5 8 3 1
2 1 4
.rho #.is chr .is (
<'zero'
<'one'
<'two'
<'three')
+-----+
|zero |
+-----+
|one |
+-----+
+-----+
|two |
+-----+
|three|
+-----+
2 2 1
.rho #.is ca .is ('abc
def')
abc
def
7
.rho #.is ca .is ('abc
defgh')
abc
defgh
9
Sorry for the transliteration, still using ascii; # is quad
Sam Sirlin
On Tuesday, September 6, 2022 at 4:08:50 PM UTC-7, Phil Last wrote:

Dyalog Ltd. is likely soon to embark on the development of its own version of a far-reaching enhancement to APL notation.

Referred to as "Array Notation" it's based initially on the simple premise that rather than engendering a SYNTAX ERROR, an opening bracket, "[" closed on a subsequent line or after a diamond, delimits a multi-line expression that denotes a multi-dimensional array wherein each expression defines a cell of that array:

ri| numraE[1 1 2 3
5 8 3 1]
2 4

The principle is naturally recursive:

ri| chrraE[
['zero'
'one']
['two'
'three']
]
2 2 5

Over a number of years the idea was elaborated until Dyalog adopted it and continued to modify it after 2015.

There is a couple of pages in the APL Wiki that attempt to describe it.

It has been modelled in APL several times and has been implemented as a part of the language at least once.

I'm sure Dyalog would be pleased to accept suggestions and comments in their forums to hone their design. I suggest the "Language" forum at https://forums.dyalog.com/viewforum.php?f=13 would be the place to start.

--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.apl

If you look at the way Adam and I have implemented it in our different platforms - Link and Acre - we've both followed the direction we agreed back in 2016-17 that "broken" brackets and "broken" parentheses denote respectively increased rank and increased depth so your examples would create differently nested vectors rather than multi-dimentional arrays. A mix of the two can then form lists of matrices and matrices of lists in any combination.
--- Synchronet 3.21d-Linux NewsLink 1.2

From Newsgroup: comp.lang.apl

" I like the fact that the input is the same as the output " was one of my two main incentives for all the effort.
--- Synchronet 3.21d-Linux NewsLink 1.2