Andrew Greenberg wrote:
> Surprise!  Another newbie arrives to inform us that Squeak is 
> deficient 
> because it doesn't look like Windows.  The archives are replete with 
> these arguments, which have been discussed substantively again and 
> again.  Bottom line: don't like it, fix it.  Many have implemented 
> various forms of "work-alike"-ism in the past -- including direct GUI 
> access, and interestingly enough, nobody really used them -- there 
> simply wasn't the need and demand for it that the newbies 
> anticipated.  
> The projects ultimately laid fallow after completion and dropped into 
> the bit bucket.

I think you're drawing the wrong conclusion from the evidence in this
case.  The reasons that nobody really used (IMO) them are many fold, but
here a few few of them:

1.  It's risky.  The base Squeak has not yet reached a state where the
core system is very stable and many desirable features are still being
added.  Thus, there are likely to be changes to the core Squeak system
that you'll want, and there's no guarantee that the author of the GUI
system will keep up to date with those changes.  And, everyone knows
that without a good modularity solution, the work of upgrading is at
best unknown, and at worst monumental.

2.  Few real world applications (native widgets).  There are a whole
range of applications that would need native widgets.  Not emulated
widgets, but actual native widgets.  However there is absolutely no need
for them in prototypes, non gui works, research activity, etc.  There is
actually one very popular use of native widgets in Squeak that people
tend to forget...Swikis (and other web serving applications).  The
argument that they are not needed is bogus in the absence of a
discussion about the application in question.

3.  Few real world applications (emulated widgets).  With emulated
widgets, you can *in fact* reach out to your potential users more
effectively by offering them a familiar interface.  It's debatable
whether or not the emulated interface is better or worse than Squeak
native interface, but it is not debatable whether or not the emulated
interface is easier to learn for people that already know the emulated
interface.

More real world applications would certainly surface a greater demand
for native or emulated widgets.  And, this is not to say that a real
world application needs native or emulated widgets.  There are plenty of
examples of real world applications that do not need either native or
emulated widgets, but it's the nature of those applications which
determine the acceptability of using a custom (or Squeak) UI paradigm.

So...this whole argument is pointless, not because native or emulated
widgets are not needed.  But because the whole argument is being debated
in the absence of the context of any particular application. 

(also, it is not by intention to belittle any of the great work being
done in Squeak by using the term "real world applications"...with that
term, I simply mean applications targeted at users among the majority of
people using computers in the world today) 

- Stephen