Begun Fri.Nov,001117
Pavel Kocura wrote:
> GUI extensions of K: a K programming environment, GUI for K
Before the US elections , I was too busy with
http://cosy.com/views/travlgat.htm to add my comments on this important
issue .
I want fundamental facilities commensurate with the level of abstraction of
the language itself . I want to build whatever I need from that foundation
in K itself . We`ll really see how fast K is at the EndUser end .
At the most coarse , let me have a map from a character array to a fixed
font window . And , Let me see every keyboard event .
A standard mapping , in open K , of those events for ordinary text input
would be a needed exercise .
Better , give me a map from pixel lists to windows .
And let me see some K-meisters create a little vocabulary to treat it
as partitioned into sub-rectangles and implement the fixed font level
above in open K .
Then I`d be up to the functionality I just about had back in `84 in
CoSy in STSC APL+PC before upgrades to Windows eliminated the direct
screen mapping . ` Never had the performance to interpret all keyboard
events tho .
Oh yea , let me know where the "mouse" is and what the buttons are doing
, too .
I`d like to know more about the nature of current interface between the
existing K gui widgets and the K engine . I remain impressed with with
Tom Zimmer`s WinFORTH which simply taps into some table of the names of
all the Windows entry points . You are responsible for learning what
they all do from MS documentation if you need them .
Unfortunately Zimmer himself apparently recognizes the excessive
complexity his code implementing his editing environment within his
own FORTH upon those bare Win calls . He`s totally stuck in FORTH`s
low level thinking .
> 1. What facilities would you want from a dedicated K programming editor?
> 4. How do you write K programs? (i.e what's the procedure/routine/working
> style?)
I continue to use my own CoSy/APL environment I basically wrote years ago
in open APL as a general purpose note taking and computing interface .
One of its key aspects is that tapping F6 sends the line under the
cursor to the APL interpreter and space is opened in the following text
for any result generated . There are several variants on this action .
I`ve constructed a few words to communicate , thru files , with a little
hosting kernel in K and have programmed F10 to execute K expressions
in the same manner as F6 executes in APL .
Altho the old DOS environment is dead-mouse , the environment has a number
of facilities , coming from it`s open code nature , I still have not
seen elsewhere . A basic point is to never have to write anything twice
and to be able to edit and re-execute any line I want until I get it
right .
It is the continual redefinability of its open code that has kept it
alive .
Bottom line , I am generally living in an HTML string which may have
all sorts of other stuff , notes , programming , whatever , in
< ! - - - comments - - > .
An HTML browser based interface might be a very practical direction
to go as Jim Brown and James Wheeler have with their SmartArrays
Java/C++ function library .
Finally , I have never seen what these Emacs based interfaces are
like . The closest I got to emacs was a thing called
FINE Is Not Emacs which U. of Rochester had .
I would be very interested in seeing one if some one here around NYC
might be so kind . Actually , I think it would be interesting if
some occasional User meetings were held here again . I think there
is enough evolving info to be exchanged .