[Dxspider-support] Spider Protocol (was Problems with command " dx by ")

[Ian J Maude]

On Mon, 2003-12-29 at 16:03, Dirk Koopman wrote:

> But what is it we actually, really, want to achieve with lists of
> users
> (and nodes)? Why do we actually *need* to know who is out there? Have
> we
> not now got to the stage where needing a huge (and probably
> inaccurate)
> sh/c listing is no longer a sysop requirement?
This is a great question!  One that I hope will generate some debate at
least.  How often do I use sh/c or sh/c/n?  Probably very little.  Am I
interested how many users a particular node has, probably not.  Would I
want to see the users on a particular node? Possibly, from time to
time.  Users like to look to see if a friend is connected, that could be
furnished with a 'where' type command I guess.

 
> If you have a system that will (or at least: could) tell you that a
> talk
> (for instance) has been delivered, what more do you need to know?
That would be fine for me?  Other views anyone?

> Not having any user and node state information, provided that
> everybody
> doesn't poll the system the whole time for user info, would reduce the
> bandwidth requirement considerably.
Agreed.

> Also, if you reduce the information that you send out to just
> (possibly)
> 'useful', instead of state information, then I reckon you could just
> flood route (with some simple optimisations) everything.
In truth, I see hundreds of nodes here at gb7mbc and I am really only
interested in a few, as I guess are my users.  Mainly UK for example. 
If I could look for a user and it reported them on a Spanish node for
example, that would be enough.  It is only then a matter of education.

> I want to make 'chatting' between interest groups to be the standard
> way
> of communicating. Announces, DX and talk messages just represent
> different aspects of 'chatter' where DX is interesting to all,
> announces
> to some and talks just 'chat' groups of two people.
I agree totally with this.  The only issue is ann/full for talks and I
know you have some ideas about re-routing these if a callsign or
something is seen to force the chat.  Having said all that, with no PC
protocol to worry about, it should be easier to implement as there would
be no hidden users in the network anyway.

My 2 Euros worth, anyone agree/disagree?

Ian

-- 
Ian Maude G0VGS Morecambe Lancs UK | ian at gb7mbc.net
Sysop of GB7MBC, the Morecambe Bay Cluster
Running Linux and DXSpider
 
Find out about Summits on the Air!  
Visit the SOTA web site at http://www.sota.org.uk