Re: #296: 203 Non-Authoritative Information: deprecate?

I think that like it or not, the meaning of the term "proxy" in the 
context of HTTP has been somewhat relaxed from a legally-based definition.

Apart from translating content, agents called proxies tend to do all 
sorts of things without the express agency of their client.

Like challenging for authentication, or returning an access denied 
message, or a page stating the content contained a virus, as well as the 
features you mention.

It is of course completely unrealistic to expect either

a) vendors to desist from providing such features which are demanded by 
customers; or
b) vendors to abandon the term "proxy" in relation to their product.

Therefore a semantic argument about the use of the term is of little 
purpose.   Educating the entire world to a new term which may more 
appropriately describe the behaviour of the "proxy" is an expensive 
thing to do, and I'm not aware of a compelling need (user confusion).

History is full of cases where an original meaning for a term is changed 
(even radically) over time or by passage between languages.

So, my question is, what do you hope to achieve with this argument?


Adrien de Croy

On 9/06/2011 7:19 p.m., Poul-Henning Kamp wrote:
> In message<>, Mark Nottingham wri
> tes:
>> We're rewriting the text to clarify them; they've always been there,
>> just badly documented (mostly in the caching section, of all places) as
>> "non-transparent proxies," a term which eventually got a different
>> meaning in common usage.
> I don't quarrel with their existence, but I disagree that a service
> which translates contents from english to french (a "semantically
> meaningful transformation" if I ever saw one) is a "proxy" of any
> kind.
> In RFC-2616 non-transparent proxies are quite clearly devices which
> aid transmission but do not disturb the content as presented:
>        A "non-transparent proxy" is a proxy that modifies
>        the request or response in order to provide some added service to
>        the user agent, such as group annotation services, media type
>        transformation, protocol reduction, or anonymity filtering.
> Notice that it says: "service to the user agent", not "service to the user".
> We are talking about converting PNG to GIF to cope with old browsers,
> not translation of languages, insertation of advertisements and similar.
> The reason why I think the distinction is important, is that the
> transformations covered by the RFC2616 text can be done without
> a lawyer, no matter what origin server you are talking to:  There
> is no change in artists control over their images, there is no
> change in authors control of their texts.  The proxy simply does
> not in any way change the users perception of the web-object she
> is being presented.
> But once you start translating works to different languages, remove
> bits that are offensive to taste and so on, the term "proxy" simply
> does not cover your activity, no matter how you qualify it, and
> therefore I think allowing "semantically meaningful" translations
> for proxys should not be allowed.
> Such transformations can also result in non-trivial HTTP header
> rewrites, for instance "Content-Language", a field which is not
> approved for non-transparent proxy modification in RFC2616.

Adrien de Croy - WinGate Proxy Server -

Received on Thursday, 9 June 2011 07:52:09 UTC