Hello,

People seem to be dodging this, so let's take it into another thread.

There seems to be an assumption that provably unique SL addresses would be 
needed.

Can you tell me at least one reason (note: don't say globally routable)
where this would be required instead of "nearly enough" unique identifiers
(chance of collision about 1/2^38)?

Why I prefer "nearly unique"?  Some reasons from the top of the head:
 - 2^38 is big enough to be in practise collision-free
 - there is no need for _any_ registration infrastructure <=== !!!!
 - automatical generation is easier, also for e.g. disconnected sites
 - people don't get funny ideas they could be totally collision-free and 
thus routable in the internet; or use them as global identifiers
 - gives people some reason to use _globals_ too, when they're really to 
be used

But then again, my take is that we shouldn't be advocating these unique
SL's everywhere anyway, most should just be able to stick to globals.

-- 
Pekka Savola                 "Tell me of difficulties surmounted,
Netcore Oy                   not those you stumble over and fall"
Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords

--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to