Re: discovering "alternate" representations?

Folks;

first off, thanks a bunch for your comments, greatly appreciated.

Am 05.10.2012 02:06, schrieb Erik Wilde:
> RFC 6596, maybe? cheers, dret.

Thanks for pointing me there. The approach outlined in there seems
definitely reasonable, and I will definitely take a look in implementing
this, eventually based upon header fields to work with binary resources,
too. However:



Am 05.10.2012 03:38, schrieb mike amundsen:
>
> 1) use link headers with the response that return the list of all
> other representations for the current response (work on rel values
> for this)
>
> 2) support for an OPTIONS call for the current URI that return the
> list of possible reprsentations for the current response
>
> 3) arrange a client-drive negotiation model where the server returns
> 300 w/ a list of possible representations and the client picks which
> it wants this time

So far I thought about (1) and (3) already. Is there any of these 
options, including RFC 6596, to be preferred from an interoperability 
point of view, talking about various software systems accessing these 
resources, including browsers? I.e. it seems people are pretty much 
reluctant to using / implementing OPTIONS, for whichever reasons.

TIA and all the best,
Kristian