tim Rowledge wrote:
> Exactly. And it's about time we used unicode properly and had the
> appropriate char for our assignment. This ':=" crap is just vile. What
> is this, Pascal?
> 
> I quite like the leftwards sqiggle arrow"⇜" Unicode 21DC and RTF8 E2 87
> 9C but even plain boring  ← Leftwards Arrow unicode 2190 would be ok.
> And how about ⤴ arrow pointing rightwards then curving upwards unicode
> 2934 for return?

I like Unicode, but if you're not joking, consider how hard are Unicode
characters to type for a commonly used expression on a common keyboard?

And then at what point does Squeak start sliding into APL once you go down
that route? Arbitrary Unicode unary or binary selectors?
  http://en.wikipedia.org/wiki/APL_(programming_language)
Putting a core operator into Unicode just seems like asking for trouble to
me. Like APL, maybe *fun*, but still *trouble*. :-)

Having said that, of course great Unicode support is a must, and there is no
reason an expert could not define things to their personal liking. But do
you want to be the one writing the newbie tutorial on Squeak including a
long section on how to construct a backarrow or uparrow on an arbitrary
keyboard across all platforms? I guess you could just suggest they cut and
paste one from code somewhere or pull up a virtual keyboard or rely on auto
completion in a browser? :-)

The leading colon of ":=" actually makes a little syntactic sense in a
Smalltalk-y way; assign is kind of like a message to self, only different
(but maybe I have just looked at too many. :-)

Again, what real benefit to whom for what cost to whom of doing it
differently from every other Smalltalk as the *default*?

By the way, the last character you included did not render for me in my
email client (using Thunderbird, with a lot of fonts loaded). That should
not be the sole decider, but it is a data point on the complexity of
introducing Unicode for *core* syntax.

--Paul Fernhout