[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: post + comments
This gap doesn't exist with wfw:commentRss since the server provides all comments in response it knows about when requested by the client.
Basically all I want to know is if Atom's threading model will focus on the 1:N scenarios or the M:N scenarios. It if focuses on the M:N scenarios it is likely I won't implement support for it since it will conflict with the threading that RSS Bandit already does for items that reference each other by linking. On the other hand if it focuses on the the 1:N scenarios I definitely would want to implement it since the ability to read comments in the aggregator is a feature that lots of my users like having and often pester other aggregator authors to add this feature as well.
Ideally I'd prefer if the distinction was made between both scenarios and both a concept of <parent> and <in-reply-to> made it into the spec.
--
PITHY WORDS OF WISDOM
There are always two solutions to the problem: yours and the boss's.
________________________________
From: Ziv Caspi [mailto:zivca@xxxxxxxxxxxxxxxx]
Sent: Thu 6/24/2004 6:37 AM
To: Dare Obasanjo; bobwyman@xxxxxxxxxx; 'Isofarro'; 'Atom-Syntax'
Cc: 'Asbjørn Ulsberg'
Subject: RE: post + comments
The problem you've raised regarding the M:N model also exists in the 1:N
model: if there are "gaps" in the comments feed (your snapshot of the feed
shows C-point-to-B, but B-points-to-A has already expired) all you can say
is that C-point-to-B is some descendant of the original A post, but nothing
more than this. The M:N model can easily achieve a similar functionality, by
using more than one ancestor indications (you can use the immediate parent
and the ultimate parent, for example, and have a reasonable model that works
in most cases).
(*) The attractiveness of this model is that it is naturally resistant to
spam -- you only get content from people whose opinions you want to hear.
Ziv Caspi
cell: +972-53-668-751
web: http://radio.weblogs.com/0106548/