Re: Negotiating compression

On 2014–06–01, at 1:30 PM, Amos Jeffries <squid3@treenet.co.nz> wrote:

> As a result we need a halfway operation to prevent the existing huffman
> table being used. The best proposal for that so far is the
> disable-huffman SETTING. Which downgrades it from the existing table to
> none rather that changing to an arbitrary alternative table.

I’m not sure how this relates to the previous message, which was about the idea of allowing upgrades to a better table. We already have a facility to disable compression per header for security reasons.

> The simplicity gained for naive implementations is an added bonus.

Naive implementations may wish to hard-code request frames. If Huffman may be disabled for servers, then hard-coded clients must never use it. But, small implementations would often benefit from minimizing transmitted data.

So, it might not be a good idea to allow servers (origin or intermediary) to disable Huffman.

As for complexity, my C library compiles to under one kilobyte and I have placed it in the public domain. In terms of implementation effort, it’s only a matter of finding such a library in one’s chosen language.

Received on Sunday, 1 June 2014 08:21:21 UTC