Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!wuarchive!kuhub.cc.ukans.edu!anu-news!list
From: munnari!csc.anu.oz.au!gih900@UUNET.UU.NET (Geoff Huston)
Newsgroups: news.software.anu-news
Subject: Re: RE: Information needed
Message-ID: <8909260021.AA11199@uunet.uu.net>
Date: 26 Sep 89 15:09:02 GMT
Sender: ANU-NEWS Discussion 
Reply-To: Geoff Huston 
Lines: 61
X-cc:         GIH900@uunet.UU.NET

Following on with this topic about the differences between the USENET NEWS
model and that used by VAXnotes:...
     
Mats Sundvall writes:
     
>This also make it similar to Usenet News in the sens that you hook up postings
>to a linear list that was sampled some time ago. All other users that may reply
>to the same item may or may not have your replies in their list. If they
>sampled the database BEFORE you updated the database they will not have them
>available. The only difference is the amount of time it will take until you
>get the "complete" reply list.
     
but you will NEVER get a "complete" reply list (unless the newsgroup is very
inactive) - new postings are being added continually.
     
>Is it possible to use the timestamp to order entries? If you assume that
>all nodes have a valid timestamp and a valid timezone header line you
>could use this to order entries.
     
Have you looked at the timestamp fields in the headers? They are a mess!!!!
Even if a program could make some sensible effort at translating the ascii
timestamps into a binary post date, the information doesn't help - the
asychronous nature of USENET postings makes efforts at linearizing a
conversation of limited value.
     
> Sure DEC has to do something similar
>with postings into VAX Notes. I have never used VAX Notes but as you describe
>it they rebuild a tree into a linear list. The information you lose may not
>be that big in a small conferense but in a worldwide VAX Notes conference
>you would loose a lot.
     
VAXnotes does not have to rebuild a tree into a list: as all postings within a
conference occur on the node hosting that conference there is a unique time
order of postings into the conference. USENET does not have that ability.
     
>> So in answer to the original question - no - NEWS is not going to do much
 more
>> than it does at present in the topic/reply area. If thats what you are after
 -
>> then buy VAXnotes!
>
>I hope you are wrong. If Usenet News is going to survive it will have to
 evolve.
>This is one of the areas where it need improvements.
     
Its a big problem - part of the assets of the USENET network is the ability to
use a wide range of network tools to distribute news - mail, ip connections,
DECnet, X.25, and anything else at hand. The diversity of NEWS platforms and
the diversity of systems and networks certainly presents challenges - the task
for each NEWS implementation is to be able to correctly interpret data that is
fed to it, and feed data to adjacent NEWS sites in a format which is as
reasonable as possible (this implies adherance to the RFC, certainly, but also
the conservative use of additional functions in other implementations (e.g. the
Supersedes: header interpretation).
     
Any topic/reply tree construction implementated by a NEWS system is going to
have to fit within this USENET environment. By camparison Digital's VAXnotes
was easy - one platform - one implementation - one network architecture. USENET
NEWS faces a much more significant set of challenges!
     
Geoff Huston