New model relay facilities: FLUID-5361, FLUID-5371

Denis Anson danson at misericordia.edu
Thu May 15 19:34:56 UTC 2014


This is exactly the issue that audiophiles have dealt with since the invention of stereo.  The tape "output" of the amplifier connects to the "input" of the tape deck, and vice versa.  It would be so much easier if there were discrete terms that made logical sense for the two channels.

For GPII, it might make sense to use meteorological language.  You could call the information coming from the cloud to the local computer the "downstroke," and the information from the local computer to the cloud the "upstroke."  This is mostly tongue in cheek, but not entirely.

Denis

On May 15, 2014, at 1:16 PM, Antranig Basman <antranig.basman at colorado.edu> wrote:

> On 15/05/2014 07:42, Justin Obara wrote:
>> Hi Antranig,
>> 
>> Thanks for the summary, I have some comments inline below.
>> 
>> Thanks
>> Justin
>> 
>> On May 15, 2014, at 3:46 AM, Antranig Basman <Antranig.Basman at colorado.edu> wrote:
>> 
>>> This work is being driven by cindyli's requirements for upgrading our metadata editing component for the FLOE project to use the new model relay system. We're trying to ensure for the Infusion 1.5 release that the system is usable for at least one use case beyond the ones that were already in place when it was first designed.
>>> 
>>> http://issues.fluidproject.org/browse/FLUID-5361 implements global ordering for notification of modelListeners across the entire component tree. The "priority" element which has been supported as standard for all Infusion listeners and change listeners was ignored in the initial delivered implementation of model relay since it was not immediately clear how it should be honoured. Model listeners were notified in an indeterminate order. With FLUID-5361 listeners are now notified in priority order, sorted across all components/models participating in the update transaction - that is, for example, all listeners with priority 1 will be notified definitely before all listeners with no priority, which will be notified before all listeners annotated as "last".
>>> 
>>> 
>>> http://issues.fluidproject.org/browse/FLUID-5371 introduces new configuration options accepted in "modelRelay" blocks under the names "forward" and "backward" - please comment on naming and apparent sanity.
>> 
>> How about "in" and "out" since I'm not sure if forward means relays coming into the component or going out of it.
> 
> It means neither - either end of the relay may be on any component or on the same component.
> 
> "forward" represents the direction which carries the model material referenced at "source" to "target", and "backward" represents the opposite direction which carries the material at "target" back to "source". I don't believe it's possible to represent this intelligibly with "in"/"out". A possibility is replacing "backward" with "reverse" or "inverse". Or "forwards" + "backwards". Any preferences/suggestions?
> 
> Cheers,
> A
> 
> _______________________________________________________
> fluid-work mailing list - fluid-work at fluidproject.org
> To unsubscribe, change settings or access archives,
> see http://lists.idrc.ocad.ca/mailman/listinfo/fluid-work




More information about the fluid-work mailing list