Or.. perhaps <i:index>1</i:index>
http://www.ietf.org/internet-drafts/draft-snell-atompub-feed-index-01.txt
- James
Henry Story wrote:
It's late here, but since I have been called...
Clearly the order of the entries could not be deduced from the tags
themselves.
One would need to have an extra tag in there such as this <position> tag
<feed>
<entry>
<id>tag:first-in-list</id>
<ext:position>1</position>
<title>The Graphes of Wrath</title>
<update>1 July 2005</udpated>
...
</entry>
...
</feed>
to specify the order of the entry. I think at the time I was talking
in the context
of that being proposed.
Again it is late, but I think my idea of having a entry always stay
at the same position
was that I was thinking that from one week to another one could just
change the content
of the entry
<feed>
<entry>
<id>tag:first-in-list</id>
<ext:position>1</position>
<title>Of Mice and Men</title>
<update>8 July 2005</udpated>
...
</entry>
</feed>
So that one could still think of a feed as a sequence of changes to
resources. One
would then, I thought, not need to republish the whole list of
entries, just
those that changed from week to week. As a result there was I
thought, no need for
a <stateful> tag.
But the problem with that suggestion was that it could not cope well
with mistaken
feed updates. So that if I mistakenly entered "Of Mice and Men" but I
meant to
enter "The Restaurant at the End of the Universe" then adding the
following entry
<feed>
<entry>
<id>tag:first-in-list</id>
<ext:position>1</position>
<title>The Restaurant at the end of the Universe</title>
<update>8 July 2005</udpated>
...
</entry>
</feed>
would tend to indicate that the change was insignificant, whereas a
change
of date would tend to indicate that the there was a new entry in
first position.
Some proposed that one should distinguish therefore between the
update time of the
entry and the update time of the object of the entry...
Henry