This is the mail archive of the
java-discuss@sources.redhat.com
mailing list for the Java project.
Re: CFP: USENIX Java Virtual Machine Symposium
- To: per at bothner dot com (Per Bothner)
- Subject: Re: CFP: USENIX Java Virtual Machine Symposium
- From: Godmar Back <gback at cs dot utah dot edu>
- Date: Sun, 15 Oct 2000 00:43:55 -0600 (MDT)
- Cc: gback at cs dot utah dot edu (Godmar Back), java-discuss at sourceware dot cygnus dot com
Here's some pointers (now I haven't read all of these, mind you)
Look at:
http://www.pmg.lcs.mit.edu/papers/bidirectional/node21.html
in particular DMSV89, Dri93, PW90 and Ros88 could conceivably
apply.
I would also check out http://www.csr.uvic.ca/~nigelh/pubs.html
I looked specifically at
http://www.csr.uvic.ca/~nigelh/Publications/cdt95.pdf
>
> I don't know. I don't know exactly what you mean by "selector
> indexed dispatch tables". The phrase seems to encompass standard
> single inheritance, C++-style multiple inheritance, and many
> other techniques. Do you have a reference that defines the term?
>
I'm not sure who defined it.
See how the pdf above defines/uses it.
> > (Except that you don't spend any time trying to
> > compress the selector tables?)
>
> Well, isn't that fairly valuable: Getting compact and efficient
> tables without doing global compression? Maybe it isn't all that
> original, but I haven't seen anything similar. I'd be grateful
> for pointers.
>
It seems to me that you get the advantage of dense
partial interface dispatch tables simply by virtue of
Java's simply inheritance scheme.
Compacting the per interface vectors would require some of
the optimizations you proposed; I don't know if Bryce did any
of these.
I believe that compacting is more important if STI is applied
in other contexts (dynamically typed languages, for instance)
- Godmar