Home » Eclipse Working Groups » Science WG » Triquetrum Components(Description of initial components for the Triquetrum workflow project)
Triquetrum Components [message #1697925] |
Tue, 09 June 2015 15:36 ![Go to next message Go to next message](https://www.eclipse.org/forums/theme/Solstice/images/down.png) |
Jay Billings![Friend of Eclipse Friend](https://www.eclipse.org/donate/web-api/friends_decorator.php?email=billingsjj%40ornl.gov) Messages: 54 Registered: July 2011 |
Member |
|
|
This thread describes the individual components of the initial contribution for Triquetrum, a new workflow project that is being proposed to the Eclipse Foundation.
Jay Jay Billings
Oak Ridge National Laboratory
Twitter: @jayjaybillings
|
|
|
Re: Triquetrum Components [message #1697926 is a reply to message #1697925] |
Tue, 09 June 2015 15:36 ![Go to previous message Go to previous message](https://www.eclipse.org/forums/theme/Solstice/images/up.png) ![Go to next message Go to next message](https://www.eclipse.org/forums/theme/Solstice/images/down.png) |
Jay Billings![Friend of Eclipse Friend](https://www.eclipse.org/donate/web-api/friends_decorator.php?email=billingsjj%40ornl.gov) Messages: 54 Registered: July 2011 |
Member |
|
|
From Christopher Brooks on the contribution from Ptolemy:
In Ptolemy II, we have the notion of configurations. The Ptiny configuration is a small, useful configuration with a minimum of third party licenses. Ptiny includes the kernel, common polymorphic actors, common models of computation and Vergil, the UI. Vergil is AWT and is what we are looking at replacing with something that is more based on Eclipse.
The Ptolemy II source code is primarily BSD, so contributing it should not be that difficult. We've done a reasonably good job managing the licenses, the dependencies in the core are fairly straightforward. A tricky part could be tracking down certain contributors and getting an assignment of copyright. Fortunately, the core is fairly stable and not that many people have worked on it.
Hallvard Trætteberg created some OSGi modules that use the core Ptolemy II classes. One issue that we ran in to is that Ptolemy II uses the class path to add actors, thus we don't always know in advance the class names of all the actors that are to be used in a system. I believe that this is a bit contrary to how OSGi works, where OSGi expects to know all the dependencies in advance. There are workarounds to this though.
Jay Jay Billings
Oak Ridge National Laboratory
Twitter: @jayjaybillings
|
|
| |
Re: Triquetrum Components [message #1697946 is a reply to message #1697934] |
Tue, 09 June 2015 17:37 ![Go to previous message Go to previous message](https://www.eclipse.org/forums/theme/Solstice/images/up.png) ![Go to next message Go to next message](https://www.eclipse.org/forums/theme/Solstice/images/down.png) |
Christopher Brooks![Friend of Eclipse Friend](https://www.eclipse.org/donate/web-api/friends_decorator.php?email=cxbrooks%40gmail.com) Messages: 6 Registered: July 2009 |
Junior Member |
|
|
We don't necessarily have to reuse the ptII code core code, though if we don't use it, then we end up reimplementing the wheel. There is quite a bit of engineering there, and some particularly subtle issues.
Looking at the ptII code and using it as the basis for triquetrum would probably be a win. One big issue would be splitting things in to interface and implementation. It could be that in the new code, the package names, class names and method names might change from the names in ptII, but the bodies of a number of the methods would be similar.
It is ok if we lose backward compatibility between ptII and triquetrum. It would be nice if we had either a conversion script or had only a few types of changes to start with. Delaying the split for awhile could be a good thing.
In this message, I had a bunch of text with links to the triquetrum wiki that would help get the conversation rolling, but the forum says:
"You can only use links to eclipse.org sites while you have fewer than 5 messages."
Feh.
Anyway, triquetrum.org links to a wiki that has some info.
Also, try searching for
Ptolemy Package Dependencies
the first link is a page on the kepler site that discusses the packages that are there.
BTW - I don't find forums that easy to work unless I get email about updates to the forum. I checked the 'Post Notification' button for this post, so we will see. If I'm not responding to a future post then it is because I probably because I have not seen it, so feel free to email me directly and I'll reply within the forum. Does this forum have an email gateway that would allow me to read and post?
Is there a wiki that we can use? As an alternative, I can update the package dependency work on the triq wiki and give out accounts to anyone who wants one.
|
|
| | |
Re: Triquetrum Components [message #1698266 is a reply to message #1698259] |
Fri, 12 June 2015 16:17 ![Go to previous message Go to previous message](https://www.eclipse.org/forums/theme/Solstice/images/up.png) ![Go to next message Go to next message](https://www.eclipse.org/forums/theme/Solstice/images/down.png) |
Erwin De Ley![Friend of Eclipse Friend](https://www.eclipse.org/donate/web-api/friends_decorator.php?email=erwindl0%40gmail.com) Messages: 52 Registered: August 2013 |
Member |
|
|
Christopher,
I've continued on the work of Hallvard. You can find the current status of "bundled" ptolemy on https://repo.eecs.berkeley.edu/svn/projects/eal/ptII/branches/osgi-2-0 for :
- diva
- ptolemy.actor.lib
- ptolemy.actor.lib.gui
- ptolemy.core
- ptolemy.domains.process
- ptolemy.domains.sdf
- ptolemy.gui
- ptolemy.moml
I have added 2 new bundles, with code taken from Passerelle :
- org.ptolemy.commons : defines a VersionSpecification for model elements + a FutureValue implementation
- org.ptolemy.testsupport : some simple tools to write unit tests on models
- org.ptolemy.core.test.actor : illustrates how to add new actors in a dynamic way to Ptolemy in an OSGi application
- org.ptolemy.core.test : some unit tests using the testsupport utils and that extra actor
- org.ptolemy.modulebuilder is a small java tool that uses the pt-jar.files content lists (as done by Hallvard) in the module projects to create linked sources/resources.
Remark that as I wrote to you before, this all works fine for testing from inside your eclipse, but the PDE export doesn't seem to include linked sources.
So this approach seems to be a bit of a dead end for "real" builds I think.
If you are willing to "fork"/replicate ptolemy core stuff into triquetrum, this problem will be gone, as we will have "real" sources then from the start,
and no longer linked sources from the ptII source tree.
regards
erwin
[Updated on: Fri, 12 June 2015 16:32] Report message to a moderator
|
|
| | | | |
Goto Forum:
Current Time: Sat Nov 11 06:10:28 GMT 2023
Powered by FUDForum. Page generated in 0.02359 seconds
|