Re: [Flightgear-devel] Proposal for storing the mp server information in DNS instead of a web service

[Geoff M. <ub...@ge...>]

On 12/11/16 21:16, Torsten Dreyer wrote:
> One thing can be said for sure. There will be a requirement updating 
> the fgms installation at some point in the future. Improvements in the 
> protocol are long overdue.

Hi Torsten,

Thank you for the very pleasant reply...

We all agree what is contained within each
`packet` sent to a `fgms` server needs to
change... but is that really a `protocol`
change?

For instance, it is absolutely agreed how
char strings are encoded/transmitted is a
vast waste... thus adding new 'types` is a
needed possibility...

But this has nothing to do really with `fgms`!

Although it is recognized that the fgfs sender,
and then other fgfs receivers of such changed
packets, would have to account for this... as
a `protocol` change...

But in a way, that is a minor `fgms` change to
accept say a new `protocol` value... and treat
the packet as valid... and relaying it as it
does... no big problem...

Strangely, what about part of the packet be
gziped/compressed data, aside from the headers...
sent simply as a byte stream...

We are /all/ interested in a better MP experience
for users... absolutely!... and even reduce the IO
felt by all our very generous `fgms` server
providers... And we always hope this list GROWS!

As to forking `fgms` source, and embedding it in
a FG controlled repo, I am not so sure... Why is
this needed? We, of `fgms` development, can accept
any RFC, and respond quickly to change...

How quickly 'fgms' providers then update is another
thing... out of our control...

One of my pet problems, having played with it a
/LOT/, is the encounter with a 'tanker'... having
explored other MP ideas, when two aircraft
are in such very close proximity, we need to switch
to some sort of one-on-one packet exchanges... or
something...

Again, I repeat, fgfs MP code needs to be changed,
improved, to support new desires of the FG community,
and that I am /all/ for... and any needed `fgms`
code changes will quickly follow...

So, for sure, updating each `fgms` installation is
long overdue...

But, as always, just my thoughts on this...

Very kind regards,
Geoff.