[comp.lang.smalltalk] MVC RIP ?

neil@dcl-cs.UUCP (01/15/88)

Dear All         You will no doubt remember my  appeal  for  help
with  regard  to creating good user interfaces to applications in
Smalltalk (That was back in October).  Well a number of you  good
people  did  reply  and advised us to read up on MVC, and explore
the system's MVC classes.

Such replies included one by Ken Auer, who very kindly offered to
come over to England, with his lovely wife, and tell us all about
MVC if we could pay his expenses . (har har) :-) .

At present, with the help of Trevor Hopkins at Manchester,  we've
learned  how  to produce simple MVC type interfaces, but are con-
stantly frustrated by how difficult it is to produce very complex
interfaces  (which  include state, or modes), or interfaces which
are generic or should I say reusable .

Obviously apple have decided to move away from the MVC idea,  for
reasons  known only to themselves, and I don't think they will be
alone in  trying  to  find  some  other  interface  paradigm  for
Smalltalk cos MVC is no panacea .

But there should be scope for upward compatibility to some extent
surely  ?  For example our latest version of Smalltalk [ParcPlace
VI2.2 VM1.1] includes a multiple inheritance facility, which none
of the standard system classes use, but it keeps working.

Any comments gratefully received

Neil
-- 
EMAIL:	neil@comp.lancs.ac.uk		| Post: University of Lancaster,
UUCP:	...!mcvax!ukc!dcl-cs!neil	|	Department of Computing,
					|	Bailrigg, Lancaster, UK.

de@comp.lancs.ac.uk (David England) (01/18/88)

In article <465@dcl-csvax.comp.lancs.ac.uk> neil@comp.lancs.ac.uk (Neil Haddley) writes:
>
>Obviously apple have decided to move away from the MVC idea,  for
>reasons  known only to themselves, and I don't think they will be
>alone in  trying  to  find  some  other  interface  paradigm  for
>Smalltalk cos MVC is no panacea .
>

MVC is fine for prototyping/hacking in Smalltalk but its locked into
the idiosyncracies of the Smalltalk "machine" implementation. Sounds like 
you need a more abstract specification (esp. for "generic interfaces) or 
a graphics UIMS.  See Silbert et al. SIGGRAPH '86, Jacob ACM trans. on
graphics, Oct 1986 or Coutaz Interact 87 for stuff in that direction.
Who is this guy anyway :-)

Dave de@lancs.comp.ac.uk@ucl.cs.nss
-- 
Dave uucp(Europe):   ...!mcvax!ukc!dcl-cs!de
     uucp(Atlantic): ...!uunet!comp.lancs.ac.uk!de
     arpa/janet: de@comp.lancs.ac.uk