Robert Collins wrote:
> Exactly my point about getting the attached clients to be one
> step 'above' the HTTP data stream in terms of abstraction.
>
> Then it should become very clean. Different TE requirements &|
> http versions is an argument for encoding on client side and decoding
> before or in the store.
I dislike the idea of TE decoding before the store only for the cause of
TE decoding.
The majority of these objects will not get referenced again, and any
work spent on this decoding for such objects is overhead.
Also, if we have had one client requesting an object in such a way that
it made sense for the server to TE encode it, it is quite likely that
any cache hit will have the same client capabilities.
/Henrik
Received on Sat Jan 13 2001 - 16:39:20 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:13:18 MST