W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 2014

Re: Dependency based properties

From: Roberto Peon <grmocg@gmail.com>
Date: Thu, 27 Mar 2014 12:04:18 -0700
Message-ID: <CAP+FsNc7_rAvd-pDYKtiZdOOYM51V_ZkuWnJdL_EkGYVf_bviA@mail.gmail.com>
To: Martin Thomson <martin.thomson@gmail.com>
Cc: Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com>, HTTP Working Group <ietf-http-wg@w3.org>
On Thu, Mar 27, 2014 at 11:56 AM, Martin Thomson
<martin.thomson@gmail.com>wrote:

> On 27 March 2014 01:50, Tatsuhiro Tsujikawa <tatsuhiro.t@gmail.com> wrote:
> >> A <-- B <-- D
> >>    \- C
> >> I'd assume C would make progress before D given it is closer to the
> root.
> >>
> >
> > Thank you. 'Closer to the root' is a good reason.
>
> I had assumed an implementation that wouldn't respect this "closer to
> the root" property, but that isn't the only implementation I can
> imagine.
>
> Note that you haven't really said anything at all about the relative
> priority of C and D.  It could be that a "longer" tail is actually
> more interesting to the client, but either way, the client has the
> means of expressing an actual preference using prioritization.  In all
> other cases, the server is left to guess.
>
> >
> > If B is closed, then do C and D have the same precedence?
> >
> > A <-- D
> >      \- C
>
> Same as above.  C and D are assigned no relative weighting or
> dependency, so the order in which the server allocates resources is
> indeterminate.
>

It depends also on how the server decides to do cleanup.
It could keep knowledge around that 'b' existed, and so act more closely to
what the client probably intended, but it could also immediately modify the
deps and execute on it, as you've assumed above.
-=R
Received on Thursday, 27 March 2014 19:04:45 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:14:25 UTC