[Grace-core] Pomona Colloquium & meetings

[Kim Bruce]

Thanks, I've run into instances of both of these in the past, but just needed to have an example ready on a slide so I could show it.

I suppose if we had the semantics of simultaneous assignments then we could get around the uninitialized variable problem, but I suspect it would be more complex than it was worth!

Kim



On Oct 6, 2012, at 4:35 PM, Marco Servetto wrote:

>>> I had a number of complaints about dropping null from the language - they didn't see the advantage as we still detect uninitialized >>variables.  I'll need to have a good example ready for such objections in the future.
> you could say that is like using a proxy pattern: the get method can
> trow exception on reading, thus "null" can never be propagated, and,
> as example, no method could ever receive a "null" argument.
> 
>>> One student also wanted to require that all variables be initialized on declaration.
>>> I'll need another example ready to show why that can't always be done (at least not gracefully).
> The most common example is where circular initialization is required.
> (I'm working of something to solve this somehow... in my hearth I
> agree with that student.)