Re: [unifont] [svfig] <NSF> Crays & APL

[Paul Hardy]

Bob,

My favorite computer song of all time is "APL Blossom Time".

Apart from the commercial APL implementations you mention on your website, there is also GNU APL, written by Dr. Jürgen Sauermann (https://savannah.gnu.org/projects/apl).  I created a special GNU/Linux PSF (console) font of 512 bitmapped glyphs taken from Unifont that supports GNU APL and some other APL versions, so folks could run APL in single-user console mode.  There was no other such PSF font at the time.  Here is a link to the latest version: http://unifoundry.com/pub/unifont-9.0.06/font-builds/Unifont-APL8x16-9.0.06.psf.gz.

In my Unicode forays, I once came across a page that lists interesting Unicode symbols and I just managed to find it again (http://t-a-w.blogspot.com/2008/12/funny-characters-in-unicode.html).  One symbol is from APL with a snide comment; here are a few highlights:

• ⍨ APL FUNCTIONAL SYMBOL TILDE DIARESIS: "Operators from APL programming language got into Unicode too. APL is like 1960s' Perl. This operator doesn't feel too good because it has to program in APL."
• ⋙ VERY MUCH GREATER THAN: "Sometimes it's not enough to be greater than, or even much greater than something else. Oh no, you need to be very much greater than. I think TeX is spoiling mathematicians and they come up with way too many symbols, and then we have to support them."
• ≸ NEITHER LESS-THAN NOR GREATER-THAN: "A polar opposite of the previous character. It's not greater than, neither is it less than. We kinda have a symbol for that already - U+003D EQUALS SIGN. OK, I know it's about partial orders, and it means that two objects cannot be compared, but it's not any less funny for knowing that."

Paul

On Mon, Mar 13, 2017 at 4:20 PM, Bob Armstrong <address@hidden> wrote:

On 2017-03-13 14:23, Paul Hardy wrote:

APL was the second language that Richard Stallman used.  If you don't already know it, he wrote this cute little ditty:

Rho rho rho of x
Always equals 1
Rho is dimension,
Rho rho rank.
APL is fun!

Paul

https://stallman.org/doggerel.html . There were a number of rho rho rho ditties around . Hadn't been aware of Stallman's .

Altho I still use rho as word for the count of a list in CoSy , I've followed Arthur Whitney's structure of lists of lists where simply the count of the list is in the header rather than the rectangularity enforced by traditional APL's header containing rho_rho followed by rho_rho counts of each dimension .

Thus a dictionary , to use K's term , or more generally an association list like the Morse code translation table is

   ` morse Dv@ >t1> rho
2
   t1 ' rho 'm     | ' rho eachMonadic
(
 44
 44
 )

You have to ask for the rho of each of the item lists to get their counts and if they happen to be equal , then you have a rectangular table .

The only operation in K , and in CoSy which intrinsically operates across dimensions is flip or monadic + in K . In K , the lengths of the lists to be flipped have to be equal as in ` morse or any dictionary , else you get a length error .  CoSy with its modulo indexing eliminates that error .

  t1 flip rho
44

I won't do the ' rho 'm  to pump out the 44  2s in the fliped list .

There being no distinction between numbers of dimensions in CoSy  , there's no such thing as a scalar nor a notion like scalar extension ( modulo indexing generalizes that ) . However , I have found it convenient to "special case" singletons so if an argument to a function is of count 1 , then the result is disclosed rather than being returned as an enclosed 1 item list .

The Forth for all of this is fairly small , as you will see in https://github.com/CoSyBob/CoSy/blob/4thCoSy/CoSy/CoSy.f , and will definitely get smaller as it gets cleaned up or translated to another chip .

As I've said various places , there's a lot that's been done in CoSy , and a lot still to do .

Bob A

--

-- Bob Armstrong Vita -- CoSy  A Computing Notebook for the Business of Life I reserve the right to post all communications I receive or generate to CoSy website for further reflection

--

_______________________________________________
http://zork.net/mailman/listinfo/svfig
neither public nor private,
this membership of correspondence