[aplusdev] A+ vs. K

[stevan apter]

off the top of my head ...

K is ascii.

A (the core language) follows Sharp APL semantics for boxed arrays.
K is based on lists (arrays of scalars are equal length lists of atoms).

given this basic difference in datatypes, you'd expect to see the
primitive sets diverge, as indeed they do.

A has contexts (one level, non-first-class dictionaries).  
K has dictionaries.

A distinguishes assignment and function definition.
K functions are first-class data; functions can be variables.

K has attributes; triggers and dependencies are attributes.

A triggers are cyclic.
K triggers are not.

A has itemwise dependencies.
K does not.

A has adap.
K has 3: and 4:

A has writeable mapped files.
K does not.

K has a non-extensible native gui and an extensible Java gui.

the A+ gui (MSTK) and the K java gui are roughly comparable in 
features, although the A+ gui is faster and handles larger data.  

the native K gui is much faster and much smaller than MSTK, but 
not nearly as feature-rich.

the K java gui is much smaller than MSTK.


----- Original Message ----- 
From: "Olivier Lefevre" <Olivier.Lefevre at genedata.com>
To: <aplusdev at d13.com>
Sent: Tuesday, January 30, 2001 3:27 PM
Subject: [aplusdev] A+ vs. K


> Hi,
> 
> This is probably a loaded question but could someone give me an idea of
> how much and in what areas K and A+ differ? Also, has the A+ core been
> under active development since Arthur left MS?
> 
> Regards,
> 
> -- O.L.
> 
> 
>