RE: Re: xPLHubs (again!)

["John B" <home-automation@xxxxxxxxxxxxxxxx>]

> In fact, there's no reason to include a timestamp - it
> doesn't really matter if you leave a client on the list for a
> few minutes longer than necessary.  If the hub just assumed
> 2*interval as the timeout when it started up (interval would
> do, but 2* allows for missing one message), that would be OK.

So, are we saying that when a hub starts up, it assumes that all clients
are still alive unless a period of 2 * interval elapses and it hasn't
heard from them?

The only issue with this is that for users who don't use services, it's
quite possible that a regular Windows app could be the hub.
It's also quite possible that the role of the hub could switch between
applications quite rapidly (i.e. if users are opening/closing xPL apps
that act as hubs)

Consider a device with an interval of 9 (the maximum permitted value).
If the hub is never left running for more than 18 minutes, the client
will never disappear, because the hub will never be up long enough for
it to timeout.

Xpllib only saves the hubs when it's shutting down, so I see no problem
with you adopting the same approach in your own library.

Regards,

John


xPL Links: http://www.xplproject.org.uk http://www.xplhal.com http://www.xpl.myby.co.uk
To Post a Message: ukha_xpl@xxxxxxx
To Subscribe:  ukha_xpl-subscribe@xxxxxxx
To Unsubscribe:  ukha_xpl-unsubscribe@xxxxxxx