Re: [vwrap] thinking about LLIDL and the future

[Meadhbh Hamrick <[e-mail]>]

> EVENT is not a resource verb  but an asynchronous transport
> property of a resource, and if that property is enabled by a CRUD operation,
> then EVENT messages are emitted under control of that resource.
>

hmmm.... okay... point taken. true enough, EVENT describes not a verb,
but a message type.

my motivation here is to identify unsolicited messages. i think it's
unrealistic to assume that EVERY message will be properly modeled as a
request / response. consider the situation when someone wants to open
a chat session with you? is it better to model that interaction as an
unsolicited message (like you receive an event saying: "hey. someone
wants to open a chat session with you") or as a response to the
request (like "this is my request that you alert me on my event queue
when someone wants to chat with me.")

so yeah, i agree that EVENT is not a verb in the way CRUD verbs are,
but i don't agree that EVERY interaction in a virtual world is a
synchronous request / response, which is why the EVENT was added.

we could, however, say that the model implies a reverse client-server
relationship over an event queue and in that case, things that were
events would wind up being resources the server accesses using normal
CRUD message types.

it still doesn't help us if we want to model an object update with the
abstract resource definition or interaction definition scheme.

i think we're dealing not only with whether a message is synchronous,
but also dealing with whether the message has been solicited. i think
REQUIRING interactions to be both synchronous and solicited is a big
mistake, cause it means we can't use the system to model object
updates.

-cheers
-meadhbh
--
meadhbh hamrick * it's pronounced "maeve"
@OhMeadhbh * http://meadhbh.org/ * [e-mail]