Formal model for defender method resolution

[Paul Benedict]

Agreed. I looked at the document for 10 seconds and closed it. It looks
written for those with a strong mathematical set theory background.

On Thu, Jan 20, 2011 at 8:01 AM, Collin Fagan <collin.fagan at gmail.com>wrote:

> Okay, I'm officially out of my depth. This looks like math and I was never
> good at math. Is the English/java for this really so verbose as to require
> this massively abbreviated syntax? It's a six page document, if the syntax
> is ten time more expressive/terse then I can see not wanted to write 60
> pages. But if it's only twice as abbreviated the would not a 12 page
> document reach more people?
>
> Collin
>
> On Wed, Jan 19, 2011 at 11:48 AM, Brian Goetz <brian.goetz at oracle.com
> >wrote:
>
> > At
> >
> >
> http://cr.openjdk.java.net/~briangoetz/lambda/featherweight-defenders.pdf<http://cr.openjdk.java.net/%7Ebriangoetz/lambda/featherweight-defenders.pdf>
> <
> http://cr.openjdk.java.net/%7Ebriangoetz/lambda/featherweight-defenders.pdf
> >
> >
> > I have posted a draft of a formal model for resolution of defender
> > methods.  This is written in the style of "Featherweight Java"
> > (Igarashi, Pierce, et al), in which a number of real-world language
> > concerns are abstracted away, in order to simplify the formalism for the
> > portion of the language of interest, notably the typing and resolution
> > of defender methods.
> >
> > Hopefully this will serve as a basis for discussion of the proposal.
> >
> > The T- and S- rules are implemented by the compiler and are used for
> > typing; the R- rules are implemented by the VM to do method selection.
> > (The primary computed item of interest is mres(C), which is the method
> > resolution for a given class -- while this is not used in any further
> > production (it would be used if the operational semantics were
> > specified), it is in fact the whole point of this exercise.)
> >
> >
> >
>
>