Thomas Broyer wrote:
> 2006/3/23, James M Snell <[EMAIL PROTECTED]>:
>> One approach that I've taken in our server implementation is to use
>> Last-Modified / If-Modified-Since HTTP headers to provide a kind of
>> sync'ing mechanism.  Specifically, when a client requests the APP feed
>> using a If-Modified-Since header, the server will return ONLY those
>> entries that have been modified since the specified date.  If no entries
>> have been modified, the client will receive the expected 304 Not
>> Modified status response.
> 
> Hmm, not sure this is not an abuse of those headers…
> It was 
> The "If-Modified-Since" section of HTTP/1.1 reads:
>       b) If the variant has been modified since the If-Modified-Since
>          date, the response is exactly the same as for a normal GET.
> 
> If it had been that simple easy, there wouldn't have been a need for
> RFC3229/feed ;-)
> 

Hey, I never said I *wasn't* abusing the headers ;-) Can you say
serendipitous hack? ;-)

What we're doing may not be 100% per spec, but it doesn't seem to cause
any problems either... other than some nominal wierdness in feed readers
that support conditional gets but don't keep a feed history (like
Firefox Live Bookmarks).  In other words, assuming that some offenses
are more egregious than others, this is kind of like a rolling stop
through a four way intersection in the middle of the desert when there
aren't any other vehicles around.

- James

Reply via email to