[kepler-dev] Re: zero-length arrays
ellen_zh at eecs.berkeley.edu
Fri Aug 13 12:44:26 PDT 2004
At 11:48 AM 8/12/2004 -0700, Bertram Ludaescher wrote:
>I agree -- the immutability is a good thing in general, and your other
>email in this thread brings up some deep issues.
>No, we shouldn't take this lightly.
>An interesting problem (maybe for Yang too? hint, hint.... ;-) is what
>happens to immutable tokens in a distributed/Grid environment:
>There tokens might be just handles to remote files. So it's a good
>thing that the handle is immutable. However how can we make sure that
>the object being represented by the handle itself doesn't change?
Thanks for the hint;-) This is an interesting question, but could making
token mutable help(may be i missed the point here)? I think it is up to the
grid environment to provide some mechanism to make sure when the object can
> From a conceptual/semantic point of you, the handle should be as good
>as the object, but then again there is probably no way to enforce
> >>>>> "EAL" == Edward A Lee <eal at eecs.berkeley.edu> writes:
>EAL> At 08:53 PM 8/11/2004 -0700, Tobin Fricke wrote:
> >> Right now
> >> there's no way to append to an ArrayToken w/o copying the whole thing.
>EAL> This is not an implementation detail...
>EAL> A key principle we've followed is that tokens are immutable.
>EAL> There are many benefits to this (e.g., you can broadcast a token
>EAL> to many actors without introducing a hidden source of nondeterminacy).
>EAL> I don't think we should change that policy lightly...
>EAL> Edward A. Lee, Professor
>EAL> 518 Cory Hall, UC Berkeley, Berkeley, CA 94720
>EAL> phone: 510-642-0455, fax: 510-642-2739
>EAL> eal at eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal
>EAL> kepler-dev mailing list
>EAL> kepler-dev at ecoinformatics.org
>kepler-dev mailing list
>kepler-dev at ecoinformatics.org
More information about the Kepler-dev