Subj : New Format  vs  DNS
To   : Johannes Lundberg
From : Peter Knapper
Date : Sun Jul 21 2002 11:10 am

Hi Johannes,

 JL> Resulting in a solution 
 JL> providing all the information in the current nodelist, 
 JL> except the changed transport information. A completly robust solution.

Agreed.

 JL> On the other side, we have a a New Nodelist Format. A tag-based format, 
 JL> containing a list of nodes. And along with each node, a 
 JL> list of 0 or more transport defitiontions. Service 
 JL> type, and service specific data. For POTS(?), it will 
 JL> contain Phone number, along with modem-flags. For 
 JL> BinkP, IP and port. And so on. And in order for current 
 JL> nodes to prevail(?), public domain conversion tools 
 JL> will be provided, making it impossible for current 
 JL> nodes to even notice they are using a nodelist based on 
 JL> the new nodelist. A robust solution as well.

I can certainly see the perspective for keeping a Fidonet nodelist, however I 
still think we need to be EXTREMELY careful about attempting to modify the 
current nodelist. Like it or not, we are likely to have POTS only nodes around 
for a while, so we CANNOT break any existing S/W they may be using. We MUST in 
all cases remain completely compatible with current practise.

 JL> There is one major thing making me believe a New 
 JL> Nodelist Format would be the best option for FidoNet. 
 JL> And I think you could call it network independency. You 
 JL> will be able to add new transfer methods as time 
 JL> passes. The New Nodelist Format could be used even 
 JL> after the death of the Internet. A solution that will 
 JL> prevail.

While I can also see your rationale here, I still think there are better ways 
of doing this. I don't believe we need a new nodelist format, simply because I 
do NOT believe we need a NEW Nodelist. What I think we really need in addition 
to the connectivity issues for IP nodes (such as using DNS for contact info), 
is a Fidonet MEMBERSHIP list, IE something that tells a Fidonet node what 
"other" Fidonet nodes actually exist. It seems (to me) that this is what people 

are really talking about. Somehow we need to marry that need up with the 
technical information need, without breaking the current Nodelist. Thats no 
small task in itself.

I see the simplest "solution" for the Nodelist, is to provide a method for 
"listing" IP nodes, that does not break current POTS nodes. All that needs to 
happen to the Nodelist is for Fidonet to come up with an agreed way of pointing 

people to the DNS for contact info.

 JL> Yes, I am a dreamer. But also a programmer and FidoNet 
 JL> SysOp that will work hard for the future existance of 
 JL> this lovely phenomena(?).

And yes, surely most of us are dreamers too....;-) I am VERY grateful that we 
still have (a few) S/W developers around Fidonet. Without them, Fidonet can do 
very little. 

My IP node works by using all the little bits I have put together myself 
(mainly OS/2 Rexx coding as a wrapper to executables), the only thing missing 
is a "common" IP Nodelist type format, and all that needs is an agreement 
between all of Fidonet over how its going to handle this part of the problem.

Cheers...........pk.


--- Maximus/2 3.01
 # Origin: Another Good Point About OS/2 (3:772/1.10)
 * Origin: Baddog BBS (1:218/903)

.