[kepler-dev] PtolemyII 5.0beta release

Bertram Ludaescher ludaesch at ucdavis.edu
Tue May 24 12:00:32 PDT 2005


But the long start-up time is of concern, too. Maybe there could be a
switch that determines whether (pre-packed) components are "validated"
at start-up time or not. Once the KSW and object manager capabilities
outlined in Matt's email will be available, several of these problems
will disappear (at least that's the goal.)

Seems that we all want "on-demand" loading of actors (including
composite ones/workflows) and more flexible packaging capabilities (as
outlined by Tim). So I some convergence :-)

Bertram


Shawn Bowers writes:
 > Edward A. Lee wrote:
 > > 
 > > It seems that search and browse is to have an index file and a way
 > > to update it...  The index file should be updated lazily, and definitely
 > > not on startup...
 > 
 > Actually, the way it currently is in both Ptolemy and Kepler, I don't 
 > really see a difference, except for initial start up time.  Basically, 
 > in Kepler you know when you try to start it up that there are actors 
 > that don't work (i.e., that don't exist or that didn't compile).  In 
 > Ptolemy, you find out they don't work when you open the folder that 
 > contains them ... which throws an exception and basically crashes the 
 > system.
 > 
 > I think we need robust ways to handle all of this stuff, which all the 
 > other emails are getting towards.
 > 
 > shawn
 > 
 > 
 > 
 > 
 > 
 > > 
 > > Edward
 > > 
 > > 
 > > At 05:07 PM 5/23/2005 -0700, Shawn Bowers wrote:
 > > 
 > >> Christopher Brooks wrote:
 > >> > Also, I think because of the super cool Kepler search mechanism
 > >> > all the actors get instantiated at start up time.  This is why the
 > >> > python actor(s) caused such a problem at start up.  I suspect that
 > >> > if the JAI actors are added to the Kepler configuration but JAI
 > >> > is not present at runtime, then there will be problems at startup.
 > >> >
 > >> > The way default Non-Keplerized Ptolemy II works is that the
 > >> > configuration is not expanded until it is actually browsed.  This
 > >> > allows us to have optional packages in the configuration at the
 > >> > expense of not having a search mechanism.
 > >>
 > >> I think that we really need to support something like this behavior.
 > >> One, we don't really want to "materialize" the actors at start up -- but
 > >> we did that because it was easier to implement on our first experiments.
 > >> And two, we want the ability to support *gracefull* mechanisms to
 > >> include "proxies" for actors that aren't available ...
 > >>
 > >> I think in some cases, these "proxies" might even contain some
 > >> information that can tell someone how to go and get the actors (i.e.,
 > >> contact XYZ for the actor), or that the actor hasn't been "dowloaded"
 > >> yet (from a kepler repository) and prompt the user to determine if they
 > >> want to actually download it, or even prompt one to install any needed
 > >> dependencies (libraries), etc.
 > >>
 > >> But, we need a way to search and browse, without having to always "load"
 > >> the actors.
 > >>
 > >>
 > >> shawn
 > >>
 > >>
 > >>
 > >>
 > >>
 > >>
 > >> >
 > >> > I could be wrong about some of the above because I have no tried
 > >> > this yet.
 > >> >
 > >> > _Christopher
 > >> >
 > >> >
 > >> > --------
 > >> >
 > >> >     Bertram Ludaescher wrote:
 > >> >     > Ok, I guess what Shawn is saying... (I have the benefit of 
 > >> sitting
 > >> >     > next to him .. so I can pick his brains ;-) is this:
 > >> >     >
 > >> >     > The actors haven't been actually deleted in the Kepler
 > >> >     > repository.. they are just not shown in the browser menu.
 > >> >     >
 > >> >     > BTW -- Why are they not shown????
 > >> >
 > >> >     Because you haven't annotated them yet ;-)
 > >> >
 > >> >     >
 > >> >     > thanks
 > >> >     >
 > >> >     > Bertram
 > >> >     >
 > >> >     >
 > >> >     >
 > >> >     > Shawn Bowers writes:
 > >> >     >  >
 > >> >     >  > Carlos,
 > >> >     >  >
 > >> >     >  > We ultimately want to set up a uber-repository for all of 
 > >> the kepler a
 > >> >    nd
 > >> >     >  > ptolemy actors and workflows ...
 > >> >     >  >
 > >> >     >  > But currently, we only include a subset of those by default 
 > >> in the
 > >> >     >  > "library".
 > >> >     >  >
 > >> >     >  > However, you can, just like in Ptolemy, open any moml file 
 > >> in kepler -
 > >> >    -
 > >> >     >  > so if you know the path to those files in ptolemy, you can 
 > >> also open
 > >> >     >  > those in kepler.  In fact, you can also add them directly 
 > >> to the
 > >> >     >  > hierarchy as well (a little more work, but possible).
 > >> >     >  >
 > >> >     >  >
 > >> >     >  > shawn
 > >> >     >  >
 > >> >     >  >
 > >> >     >  >
 > >> >     >  > Carlos A. Rueda wrote:
 > >> >     >  > > Hi Dan, cvs admin, and all,
 > >> >     >  > >
 > >> >     >  > > I would like to start using some Ptolemy actors that are 
 > >> based on
 > >> >     >  > > the Java Advanced Imaging system (JAI), but they are not 
 > >> included
 > >> >     >  > > in the released kepler 1.0.0alpha6. Perhaps the current 
 > >> CVS version
 > >> >     >  > > of Kepler already includes them but I apologize I have 
 > >> forgotten my
 > >> >     >  > > cvs password (username: rueda)
 > >> >     >  > >
 > >> >     >  > > So my questions are:
 > >> >     >  > >
 > >> >     >  > > 1- Is the current CVS updated with the new actors in 
 > >> PtolemyII 5.0be
 > >> >    ta,
 > >> >     >  > >     particularly those related with JAI?
 > >> >     >  > >     If not, i) is there a time frame for that?, or ii) 
 > >> how can I ins
 > >> >    truct
 > >> >     >  > >     kepler to load them?
 > >> >     >  > >
 > >> >     >  > > 2- Would the cvs admin please contact me about my password?
 > >> >     >  > >
 > >> >     >  > > Thanks.
 > >> >     >  > >
 > >> >     >  > > carlos
 > >> >     >  > >
 > >> >     >  > >
 > >> >     >  > > Dan Higgins wrote:
 > >> >     >  > >
 > >> >     >  > >>Hi All,
 > >> >     >  > >>
 > >> >     >  > >>    The Ptolemy project has released version 5.0beta of 
 > >> PtolemyII at
 > >> >     >  > >>
 > >> >     >  > >>http://ptolemy.eecs.berkeley.edu/ptolemyII/index.htm
 > >> >     >  > >>
 > >> >     >  > >>This link is to the source code version which can by used 
 > >> to build
 > >> >     >  > >>Kepler. You might want to use it instead of trying to get 
 > >> the CVS
 > >> >     >  > >>version of Ptolemy. It seems to work fine with the 
 > >> current Kepler CV
 > >> >    S
 > >> >     >  > >>code; it won't reflect the latest changes in Ptolemy code 
 > >> but it is
 > >> >     >  > >>probably more stable and doesn't have a lot of 
 > >> experimental stuff th
 > >> >    at
 > >> >     >  > >>the Ptolemy CVS head does. You can also see the latest 
 > >> changes made
 > >> >    in
 > >> >     >  > >>Ptolemy
 > >> >     >  > >>
 > >> >     >  > >>Dan
 > >> >     >  > >>
 > >> >     >  > >
 > >> >     >  > > _______________________________________________
 > >> >     >  > > Kepler-dev mailing list
 > >> >     >  > > Kepler-dev at ecoinformatics.org
 > >> >     >  > > 
 > >> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler
 > >> >    -dev
 > >> >     >  >
 > >> >     >  > _______________________________________________
 > >> >     >  > Kepler-dev mailing list
 > >> >     >  > Kepler-dev at ecoinformatics.org
 > >> >     >  > 
 > >> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-d
 > >> >    ev
 > >> >
 > >> >     _______________________________________________
 > >> >     Kepler-dev mailing list
 > >> >     Kepler-dev at ecoinformatics.org
 > >> > 
 > >> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
 > >> > --------
 > >>
 > >> _______________________________________________
 > >> Kepler-dev mailing list
 > >> Kepler-dev at ecoinformatics.org
 > >> http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev
 > > 
 > > 
 > > ------------
 > > Edward A. Lee
 > > Professor, Chair of the EE Division, Associate Chair of EECS
 > > 231 Cory Hall, UC Berkeley, Berkeley, CA 94720
 > > phone: 510-642-0253 or 510-642-0455, fax: 510-642-2845
 > > eal at eecs.Berkeley.EDU, http://ptolemy.eecs.berkeley.edu/~eal 
 > 
 > _______________________________________________
 > Kepler-dev mailing list
 > Kepler-dev at ecoinformatics.org
 > http://mercury.nceas.ucsb.edu/ecoinformatics/mailman/listinfo/kepler-dev



More information about the Kepler-dev mailing list