Turadg wrote:
> I like the plan Scott has laid out.  I think Tony can move ahead on
> developing the tool and the user interactions above the abstractions
> in PasCurnitUtils.  For example, there are a lot of quick editors to
> write.
>
> I'm looking forward to seeing this skeleton.  In the meantime I'd like
> to know more about this statement, if there's time to articulate it.
>
> "The move to OTrunk/EMF will be more fundamental than just redesigning
> the Pod system.  In OTrunk/EMF there might not be a need to have
> something called a Pod.   Most of the functionality would be there, it
> would just be accomplished by multiple parts of the system working
> together."
>   
I think it is easiest to see with an example. 
> Would OTrunk/EMF allow for distributed authoring and versioning?  Re-
> use and remixing?
>   
OTrunk/EMF does allow this.  I think the way to think of it is like the 
.java files for Java.  The java compiler and runtime does not explicitly 
support distributed authoring and versioning.   However there are 
multiple ways to do distributed authoring and versioning of .java files.

Similarly to Java, OTrunk/EMF provides the adressing and id'ing supports 
to enable distributed authoring and version.  It is up to an external 
system to actually manage it.  We are currently using Subversion to 
maintain a repository of OTrunk examples, this is available online so 
people can easily run them.  We will probably extend this so there will 
be a repository of OTrunk files which are used in schools.  We are also 
using the DIY to store OTrunk files, and that functionality will be 
extended in the near future.

For reuse and remixing, this part of OTrunk hasn't been pushed much 
yet.   It is integral to the design.  However just like Java it is up to 
the person who is designing the system which uses OTrunk to make sure it 
is reusable as possible.   The key piece that hasn't been pushed is 
splitting the content into multiple files, and the mechanism to find 
dependent files.  But that part is coming and I hope it will be 
available before October.  

The part I'm working on now is the "overlay" system.  User data is 
currently stored as an overlay on top of the authored data.  I'm going 
to extend this overlay mechanism so it can support multiple layers of 
overlays.  We will then use that to support localization.  The bottom 
layer will the be initial authored content, the next layer will one of 
many localization overlays, and the top layer will the user data.   This 
will make it possible to localize any part of the content, but still be 
able to change the "unlocalized" parts in a single place.


Here is a continuation on with the comparison to Java, I wrote it and 
realized it wasn't really relevant.   You might find it interesting.  
The key parts that Java has to support distributed authoring, 
versioning, resuse, and remixing is the package naming and class 
naming.  As can be seen after you've used Java a while, these things do  
provide enough control to maintain over time an application consisting 
of interdependent parts.   The main things missing are packaging classes 
together, and fine grain control over dependencies between packages.  
The OSGI system used by Eclipse adds these features.  Maven tries to do 
something similar but it is missing the fine grain control over 
dependencies between packages.

> Also, I understand Scott's statement, "The longer we hold off
> switching to OTrunk the easier it will be" but want to add that on the
> other hand the longer we wait the deeper and broader the impact of the
> switch.  So let's keep this ball rolling.
>   
I agree.


--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"SAIL-Dev" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/SAIL-Dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to