RE: [Fwd: Re: [xAP_developer] RE: Win a WebBrick w orth £350!]

["Patrick Lidstone \(Personal E-mail\)" <patrick@xxxxxxxxxxxx>]

=20

> -----Original Message-----
> From: Kevin Hawkins [mailto:lists@xxxxxxx]=20
> Sent: 23 February 2005 21:13
> To: xAP_developer@xxxxxxx
> Subject: [Fwd: Re: [xAP_developer] RE: Win a WebBrick worth =A3350!]
>=20
>=20
>=20
> Could xAP be implemented as a TCP connection to James' XServer ?=20
> Requires a PC helper app but has the potential advantage that=20
> messages could be prefiltered if needed.
>=20
> Kevin

Remember I have also have my equivalent of the X-server. I haven't had a go
with Jame's (shame on me), but it would be nice to standardise on the tcp
transport for xAP.

One of the major issues is how an x-server client would discover it's
serve=
r
- I don't think users should have to configure IP addresses (and this may
not be possible on GUI-less devices anyway), so I rate it as a
"must".
Discovery without application layer support for broadcast isn't impossible,
but is potentially clumsy. The most obvious solution is a port scanning
algorithm. This could work reasonably quickly given that most home networks
today use a 255.255.255.0 netmask I guess? Or we could reverse roles, and
the x-server could detect the presence of the client (perhaps by detecting
=
a
new MAC address?) and initiate connection, which has the benefit of
allowin=
g
for really simple clients but may be a security nightmare?=20

Any thoughts? I'd be happy to host a mini-meet for a mini-brainstorm if the
Londoners of us fancy a beer and there is interest in thrashing these ideas
out a bit further?

Patrick



=20