- From: Kulkarni, Saurabh <sakulkar@akamai.com>
- Date: Thu, 7 Aug 2014 16:59:34 -0500
- To: Martin Thomson <martin.thomson@gmail.com>
- CC: "Simpson, Robby (GE Energy Management)" <robby.simpson@ge.com>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Well even if say content-type: text/html is used 10% of the time, isn’t it good to have one of these rather than having an empty entry. I am not proposing enumerating all the combinations, its just to replace the content-type: empty-string with an entry with non-empty string value. So for values different from text/html you can still use the index with a different value, but for text/html you can just refer to the entry. - Saurabh On 8/7/14, 2:50 PM, "Martin Thomson" <martin.thomson@gmail.com> wrote: >On 7 August 2014 14:27, Kulkarni, Saurabh <sakulkar@akamai.com> wrote: >> Since now we have switched the static table to the lower indices, we >> should try to see if we can populate the entries if possible with their >> most-popular values. For e.g. Œcontent-encoding: gzip¹ is better than >> Œcontent-encoding: <empty>¹. Having an empty value made sense when we >> always added entries to header table before, but now it just seems like >>we >> are missing an opportunity to further get some compression thru >>populating >> all the headers. > >I just went through the list; there really isn't much that we can add >that is unambiguously good like that. > >Content-Encoding seems like the only good example. Cache-Control: >no-cache too. > >But see previous comments regarding changes.
Received on Thursday, 7 August 2014 22:00:05 UTC