Re: Quality factors
Koen Holtman (koen@win.tue.nl)
Thu, 9 Oct 1997 20:48:54 +0200 (MET DST)
Graham Klyne:
>
>The following thoughts emerged from an off-line discussion about the use of
>quality factors in content negiatiation.
>
>The premise for what follows is the assertion that the only practical use
>for a quality factor is to rank some set of alternatives according to
>preference.
I don't agree with this premise, but we have discussed this before. I
do want to comment on the reasoning below however:
>Simple sequencing of of alternatives (e.g. per Multipart/Alternative) may
>not be possible because the sender may not be able to locate (hence
>present) the alternatives in order of quality. Therefore some separate
>ranking mechanism is required.
>
>I suggest that in this case a 3-digit (max) number is insufficient, as with
>a significant number of alternatives an implementation will soon run out of
>space within which to slot further entries between existing entries. I
>estimate that a perverse presentation would run out ranking space after
>about 10 entries (log2(1001)).
Your reasoning assumes that the sender has to send out the source
quality value of a variant as soon as the variant itself is sent. If
you wait sending source quality values until all variants have been
sent, the problem does not exist anymore. And I see no logical reason
why you could not wait: the recipient can't do anything until the last
source quality value is received anyway.
>GK.
Koen.