How about an approach that is specific to the defaulted values?
1. The feed metadata is split into feed-meta and entry-meta-defaults.
2. Each item in the entry metadata is specified as defaultable or not.
3. Entry metadata can be explicitly "unset" to not have a value, perhaps
by using an empty element with an attribute to make it clear.
Perhaps this has been proposed at one point. If so, I missed it amongst
all the other discussion.
This is orthogonal to whether feeds and entries are kinds of each other.
If they are, the feed-meta and entry-meta will need to be coordinated.
wunder
--On Wednesday, November 17, 2004 09:30:26 AM +0100 Henry Story <[EMAIL
PROTECTED]> wrote:
Ok. I agree that there are arguments in favor and arguments against.
I will go back to working on BlogEd[0], keeping in mind Anotone Roundy's points
[1][2], and see if the simplification that I am proposing here, and that is
implemented in BlogEd, really do end up creating any problems or confusions. I
have not come across any yet.
In the meantime I'll leave the pace open. I think it should be revisited
later. The way things are going [3] things are going to be getting a lot more
complicated in this spec really fast.
[0] https://bloged.dev.java.net/
[1] http://www.imc.org/atom-syntax/mail-archive/msg11577.html
[2] http://www.imc.org/atom-syntax/mail-archive/msg11673.html
[3] http://www.imc.org/atom-syntax/mail-archive/msg11662.html
On 17 Nov 2004, at 04:23, Tim Bray wrote:
On Nov 16, 2004, at 11:45 AM, Henry Story wrote:
But as Antone specifies in that mail, the differences are not
insurmountable. I have tried to answer some of his most difficult
questions in an e-mail [2].
Since they are different on the surface, and most people are going to
easily understand that they are different, I just don't see the
benefits from peering deep to find the essential sameness. -Tim
--
Walter Underwood
Principal Architect
Verity Ultraseek
