The story of E, part 2 (fwd)

[Mark S. Miller]

At 07:41 AM 10/13/98 -0700, Dean Tribble wrote:
>A few notes that I couldn't resist.
>>...
>
>If you only have one, make it a double.

Is this about numerical representation or drinks?  ;)


>> This looks to me like a battle worth fighting.  Better for us to
>> curse and grumble about the explosion of types in Java while
>> implementing E, than for generations of E programmers after us
>> to curse *us* for it. :)
>
>The same could be said about getting integer mod(division) right.

I know.  The most painful issue in E design is when to go with Java
conventions vs when to do something sensible.  If Java had adopted fewer
C++ conventions, and if C++ had adopted fewer C conventions, they each
might have been better languages that never caught on.

Sadly, there are no absolutes in this part of the design space, but a
high-priority principle is: If an expression is accepted by both languages,
and the data types operated on exist in both languages, then the expression
must mean the same thing in both languages.  As another example, my
peculiar precedence & associativity rules result from simplifying Java's
subject to this principle.  

Whenever I violate this principle (as in my treatment of "/") I
substantially decrease E's utility for rapid prototyping code that will be
re-written in Java.  


>If you can stand some Java non-conformity, "count" has always seemed the best
>generic term.

I can stand some, but not that much.  Think also of E being used as an
interactive shell by a java programmer, poking at his java world while
writing new java code.  (This is currently rather pleasant using Elmer.)
Sort of a one-line-at-a-time rapid prototyping.

	More, asynchronously, later,
	--MarkM