Print

Print


>I've started developing a small prototype WebLVC server based on the draft_0.5.

Great!

>As I understand Filters apply to objects and interactions, however there is no "ObjectName" in
>the interaction messages, and so the "ObjectBounds" filter cannot be applied
>to interaction messages. This makes filtering interactions within a range of a designated object difficult.

Good point. The ObjectBounds filter implicitly assumes you're filtering an entity object of some kind. This filter could use a lot more definition, couldn't it? For example what exactly does "... Interaction messages outside this range..." mean, anyway?

Maybe this should be dropped from the specification, and left to be implemented as proprietary extensions?

Here's a problem I'm thinking of: suppose a WebLVC client decides to publish a new object or interaction of which the server is unaware. It would be nice to avoid having to re-write the server to understand the new object or interaction. So how do we write filter definitions which don't require implicit assumptions like the ObjectBounds filter?

We're still re-working the subscription and filtering stuff, so be warned we could be in for changes. The change between 0.4 and 0.5 was quite drastic, but we make no promises about changes until we get to a first full release.


>Given a client requesting the Configuration of CoordinateReferenceSystem to be ECEF, does that
>mean that all attributes updates and interactions will be sent by the server in
>ECEF and will therefore be subject to possible DeadReckonning.

Yes, that is the intention. It's quite a convenience for a client developer. However, this also has the implicit assumption that the object is an entity, or that the server "knows what to do" for other interactions and objects.

>With regards to StatusLogRequest messages, a client has no way of knowing
>how many StatusLog messages the server has, and
>so cannot meaningfully set an Offset and a Length values.
>Similarly, cannot ask for the last 100 StatusLog, for example.

Perhaps we should make negative offsets reference the most recent log object, such that length=10, offset=-10 would get the 10 most recent log objects.

>As part of the specifications, will there be an "official" mapping
>between DIS and the Standard WebLVC Object Model?

I'm not sure. There is definitely an implicit mapping, we're using DIS semantics but adapting to JSON format. Whenever there is a question about the nature of a data value, you should refer back to DIS standards whenever necessary.

An official mapping could be written
a) inline in the WebLVC object model
b) as an appendix or separate section in the WebLVC spec
c) as a separate companion document like the RPR GRIM

Thanks for your contribution!
And good luck, please keep us up with your progress.

>Regards,
>R Wathelet
>
>########################################################################
>
>To unsubscribe from the SAC-PDG-WEBLVC list, click the following link:
>https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-WEBLVC&A=1
########################################################################

To unsubscribe from the SAC-PDG-WEBLVC list, click the following link:
https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-WEBLVC&A=1