Re: Re: Standalone xPL Hub Summary, Mark II

[Gerry Duprey <gerry@xxxxxxxxxxx>]

Howdy,

> I don't think so - to my mind, the application should act as if there
is
> a hub present anyway. I think that an app should *always* assume that
a
> hub is present and in essence, not give it any other thought.

I think this is an OK assumption, as long as the app isn't going to break.
Since you can always send xPL messages, hub or not, nothing it technically
in the way.  However, the underlying framework should provide some sort of
check to see if a hub has been detected or not (like you can ask if you've
been configured or not) so for an app that chooses not to send messages
until the hub is active, it has something to make that decision from.

> One thing also - the suggested timings have to be worded as
"suggested
> values", not as a protocol breaking rule. There would be a lot of
work
> involved in re-writing existing code, and I think that time would be
> better spent on new developments!

I'll make it suggested in the next version.  However, I have to say, I was
surprised at how easy this change was to make to the xPL4Java framework.
Only took about 30 minutes to code it and do initial testing.  I want to do
more testing, but it was pretty easy.

I've also re-done my xPL hub (for linux) so it now has a supervisory
function that insures it's running and if it stops, relaunches it.  That
took about 30 minutes too.

It's one of the things I really appreciate about this solution.  The change
is pretty easy to accomplish.

While I understand if folks don't have time to implement it completely in
various frameworks, I would hope that an version with any automatic hub
removed could happen quickly -- once this gets "official", the
clearer the
apps work, the quicker the "transition" period will be.

Gerry

--
Gerry Duprey
Ann Arbor, MI 48103
http://www.cdp1802.org