On 14 Nov 2004, at 19:36, Robert Sayre wrote:
I don't understand anything you are writing. The effect of the Pace is negligible, yet it is awful? We need preserve the ability to differentiate the spec for entry and head, yet they are almost identical? The changes proposed are so small and people are flipping out. This is getting funny.
I will withdraw the Pace, and keep the Iron Curtain between head and entry. You will feel safe. Plus, I won't have to answer emails from you about it anymore. Everybody wins.
Please don't withdraw the Pace. I am 100% behind it.
Others who think this is not such a bad idea please add your +1s to give Robert a little boost of confidence!
Robert Sayre
Graham wrote:
This Pace is awful. No one disagrees that head is very-much like an entry. But the thing is, that doesn't make it one.
What makes it one? Do you have special access to some Platonic realm where you can see exactly what a head structure is? Please let us know how and why it is not an Entry even though for all purposes it looks exactly like one.
I don't even think you believe it is really an entry, otherwise you wouldn't have called it headentry.
That is a complete misunderstanding of why Robert called the tag a headentry:
1- to distinguish normal entries from this special one
Bill de Hora explained that very well recently as feed(head,stream) [1]
2- to distinguish the headentry property proposed from the previous head
one with which he wishes to replace it. He even acknowledge that it would make
no difference in the end if we returned back to calling it "head".
This Pace does nothing, literally nothing, in that afterwards we'll still have a head and a bunch of entries, we'll just be calling them by really non-sensical terminology. The conceptual argument is completely moot.
Why not adopt it anyway then? Because the real effect of the Pace is to arbitrarily remove our ability to differentiate the spec for an "entry" and a "headentry", and there's no reason to.
I think that is exactly a very good reason to merge them. Let us simplify the spec, and if we find that a lot of problems end up surfacing because we failed to make a distinction, we can always go back to the previous state!
It seems a little silly to avoid simplifications to the spec, for some hypothetical distinction we may want to draw in the future.
Graham
[1] http://www.imc.org/atom-syntax/mail-archive/msg11563.html
