[ALLY] Re: event! experiment...
From: rgaither:triad:rr at: 21-Jun-2001 13:11
I've replied to both lists because I think the event driven
paradigm is important for core as well as view.
>> This question is probably for Holger but feel free to respond ;-)
>>
>> Holger some time ago said that there may be possibility to create/change
>> /View events. I played with this idea a bit but with a small succes :((
>>
>> How can I create my own event!, I cannot understand the structure of this
>> datatype...
>
>At the moment you cannot.
I'm going to be optimistic and read into this that RT is
at least considering putting that option in. :-)
While on the topic of event driven programming I'd like
to share some "experience" from another language that
grew "badly" into the event driven paradigm.
List of mistakes I hope REBOL can avoid -
1. Event driven is not just for user interfaces. Publish
and subscribe for example work for all kinds of design
problems.
2. A language should provide the plumbing for event
handling but leave the usage/definitions open for
developers to extend. Fits well in the current native
versus mezzanine layers REBOL uses.
3. Event handling needs to be designed with external
connections in mind. For example - while connections
from a client to an application server were already
possible the initial event mechanism was limited to each
session and could not use the existing communications
pipeline.
Off the top of my head I'm thinking we need system level
event handling and an event datatype that developers
can use however they want.
Thats enough to get started. :-)
Thanks, Rod.
Rod Gaither
Oak Ridge, NC - USA
[rgaither--triad--rr--com]