The assignment character in 3.9 and onward

> On 6/8/06, Ralph Johnson <[hidden email]> wrote:
>>
>> On 6/8/06, Klaus D. Witzel <[hidden email]> wrote:
>> > On Thu, 08 Jun 2006 13:33:55 +0200, Lic. Edgar J. De Cleene wrote:
>> > >
>> > > I wonder why not use <- , two characters as := , but closer to
>> original
>> > > idea.
>>
>> This is a bad idea, because you would be different from other
>> Smalltalks for no good reason.  The right solution is to make the
>> editor smarter.  That way you get compatibility AND a good user
>> interface.
>
>
> I agree 100%.
>
> I'm surprised that no one mentioned it, but for me, a big reason why I
> missed having the good ol' left arrow is because := just takes so many  
> keys
> to type.
>
> I've been working on a patch to add a new keyboard shortcut to
> ParagraphEditor to allow control-= to insert assignment. For now, I'll
> definitely make it insert ':=', because that's the way things are going.
>
> Duncan.

> On 6/8/06, Klaus D. Witzel <[hidden email]> wrote:
>> On Thu, 08 Jun 2006 13:33:55 +0200, Lic. Edgar J. De Cleene wrote:
>> >
>> > I wonder why not use <- , two characters as := , but closer to original
>> > idea.
>
> This is a bad idea, because you would be different from other
> Smalltalks for no good reason.  The right solution is to make the
> editor smarter.  That way you get compatibility AND a good user
> interface.
>
> -Ralph Johnson
>

>Duncan Mak wrote on Thu, 8 Jun 2006 11:40:01 -0400
>  
>
>>I'm surprised that no one mentioned it, but for me, a big reason why I missed
>>having the good ol' left arrow is because := just takes so many keys to type.
>>    
>>
>
>Exactly - there are two entirely separate issues relating to the
>assignment character: how to type it and how it is displayed by the
>various hardware/software combinations.
>
>I would just like to take a small detour into the history of character
>sets for those who keep complaining that Squeak "hijacked" the underline
>character. As we can see in
>http://homepages.cwi.nl/~dik/english/codes/stand.html there was no
>underline character in the 1963 proposal for ASCII. In 1965 the up and
>left arrows were replaced by the now familiar caret and underline
>characters, but by then the uppercase-only older version was already in
>use in printers and popular teletype terminals. It is easy to say that
>when Xerox PARC was founded this was already ancient history and they
>should have adopted the final version of ASCII but things don't easily
>change that fast - at the computer center where I worked in 1982 all our
>teletypes still had no lowercase, caret nor underline characters. And
>this is why we call a method #helpWithTopic: instead of
>#help_with_topic: like many other languages.
>
>It is interesting that the original IBM PC video hardware did include
>characters for all four arrows, with 16r18 as the up arrow and 16r1B as
>the left arrow. So when Digitalk introduced their Methods text based
>system in 1985 they could have used these. But they would have been very
>awkward to type (specially with the left arrow having the same character
>code as ESC) and since their target audience was Pascal programmers (all
>examples in the first part of their manual include the Pascal equivalent
>of the Smalltalk code) they decided to change the assignment to
>something that would be more familiar to them and would look good when
>printed on the ASCII dot matrix printers of the time.
>
>Now we have Unicode and several nice arrows to choose from while letting
>the underline be just an underline. Certainly it is far from trivial to
>make this change without messing up old code but I think it would be
>worth the effort. But even if this is done there still wouldn't be a
>good way to type in these characters, so Duncan's proposed patch would
>be very important.
>
>-- Jecel
>
>
>
>  
>

> Squeakers,
>
> >*yes*, the "pascalish" ":=" assignment is ANSI and it should be used as
>
> It stems from Algol 60. And Algol 60 inherited it from mathematics, where it is
> used to "define" something. Algol 60 was not only used as a programming language
> but also as a mathematical notation to describe algorithms concisely and
> unambiguously.
>
> As a mathematical notation Algol 60 used strange symbols. So something that
> resembles a little "v" is the boolean "or" operator. These symbols were employed
> regardless of the limitations imposed by keyboards and display devices.
>
> Nowadays Smalltalk offers all the flexibility needed. The scanner in my image,
> that originated from Squeak 1.18, does not accept ":=" any more. I simply do not
> need it. In the tradition of mathematics and Algol 60 I have the human reader in
> mind when composing a method, and if a computer program cannot cope with my
> preferences, I prefer to adapt the program to my taste instead of adapting my
> taste to the limitations of some "standards".