This is an issue that we are currently addressing in Ptolemy II by
allowing galaxies (called "composite actors" in Pt II) to be
instantiated "by reference" or "by copy". An interesting twist
is that if you instantiate by copy, then when you look inside,
you will see the instance of the galaxy, and you can then modify
that instance without affecting other instances of the galaxy.
Edward
At 04:23 PM 11/2/99 -0500, you wrote:
>Version control of stars is fairly straight-forward, since one or multiple
>developers can check-in/out the text files of the stars, but when you
>consider galaxies it becomes stickier. For example, a developer may change
>a star that is called by many galaxies, some that the developer is aware of
>and some he is not. This is analagous to a developer changing a function
>that is called by many main programs. If the star is dynamically linked
>into the galaxies, pre-existing galaxies may break. But if they were
>statically linked, the result may be that pre-existing galaxies are giving
>incorrect (or out-of-date) answers, unknown to the user or the developer.
>How is the issue of version control addressed in the context of galaxies?
>
>Thanks
>Beth Savoy
>Corning, Inc
>
>
>----------------------------------------------------------------------------
>Posted to the ptolemy-hackers mailing list. Please send administrative
>mail for this list to: [EMAIL PROTECTED]
------------
Edward A. Lee Professor
phone: 510-642-0455 EECS Dept., Cory Hall
fax: 510-642-2739 University of California
email:[EMAIL PROTECTED] Berkeley, CA 94720
<http://www.eecs.berkeley.edu/~eal>http://www.eecs.berkeley.edu/~eal
----------------------------------------------------------------------------
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
