Re: [hybi] RSV bits in extensions

[John Tamplin <[e-mail]>]

On Wed, May 23, 2012 at 2:47 PM, Bruce Atherton <[e-mail]> wrote:

> I think this is a little pessimistic. People writing extensions only have
> to accept that if they use a particular RSV bit, they are automatically
> incompatible with any other extension that uses it. So long as everyone is
> really clear in their documentation about what RSV bits they use, users can
> decide how to deploy them so that no conflicts occur.
>
> Admittedly, though, if this WG defines any other RSV bits, that will
> effectively make them off-limits to others. I wouldn't recommend to any
> extension writer that they use the RSV1 bit for their extension, for
> example, unless it was an extension for compression.
>
> So although users manually managing extensions for RSV bit conflicts is
> complicated, and although extension writers run a risk of being booted off
> their chosen RSV bit by the WG, I still think using them is doable for a
> couple of reasons. First, I don't see that there will be a flood of
> extensions that require an RSV bit, and second I don't see the WG rushing
> to create standard uses for them either.
>
> Of course, I could be wrong. If there is a sudden rush for RSV bits then
> that might be a reason for the WG to define one of the RSV bits as an RSV
> bit extender, or to have an extension for reserve bits as you suggest.
> Let's see what happens.


I suspect that if we came up with a bunch of extensions that needed
reserved bits, we would come up with a way of dynamically allocating them.
 So, I don't think it is really a problem.

-- 
John A. Tamplin
Software Engineer (GWT), Google