>The reason, I think, would be the same as in Python: indented code 
>is easier to read.  I'm not saying that a decent indentation-based 
>syntax exists for Smalltalk, but wouldn't it be nice?  Currently, 
>you can't *really* trust the indentation of code you are looking at, 
>because the indentation cues might be wrong.  You have to scan code 
>character-by-character looking for delimeters.

So now the question remains, are there any other considerations that 
must be weighed against such "ease of reading" in deciding between 
one proposal or another?  Consider, for example, "ease of learning," 
"ease of writing," and "ease of debugging."

In particular, Smalltalk has an awesomely simple syntax and 
correspondingly simple semantics related to that syntax.  Nearly 
everything is a message.  Accordingly, an analysis of "ease of 
reading" for an indentation structure syntax for #ifTrue:#ifFalse: 
must hold equally true for an indentation structure syntax for 
#copyReplaceAll:with: and every other two-parameter keyword message, 
or we must begin to invade the present syntax and semantics of the 
language.

I agree that Python code has a versimilitude that, however stilted, 
makes is somewhat "easier to read."  I have also seen straightforward 
and elegant Python code that could be made easier to read (IMHO) 
somewhat stifled by the same indendation obligations.  I'm not sure 
that "easier to read" as used here is more than a matter of personal 
taste.

Its for that reason that I was hoping we could, as a threshold 
matter, come up with some objective criteria for making decisions.