Re: dilemma of cache: two types of image

David,

My concern is how the naive-user is able to negotiate the  
possibilities for cache and <use>.

How is the user to know whether an icon that they have incorporated  
into a page, is one that metamorphes, and how often?
there is a screencast: http://www.youtube.com/watch?v=SiV-hkuKr9o
and demo file to experiment with: http://www.openicon.org/feeds/test.svg

this is in addition to the issue previously raised regarding the  
relative difficulty of setting up a browser and server with different  
cache rules.

regards



Jonathan Chetwynd

j.chetwynd@btinternet.com
http://www.openicon.org/

+44 (0) 20 7978 1764


On 17 Jul 2008, at 22:38, David Woolley wrote:

>
> Jonathan Chetwynd wrote:
>
>> Organising this to work well for a single UA and server** is  
>> already far too difficult.
>
> Given that you use .htaccess, I don't see why.  Even without trying  
> to manipulate cachability, it would be normal to put the library  
> images and editorial ones in different directories.  All you then  
> need is a .htaccess files in each directory.
>
>> ExpiresDefault A0 Header set Cache-Control "no-store, no-cache,  
>> must-revalidate, max-age=0"
>
> This sort of sledge hammer approach to cache control is actually why  
> people are unlikely to put much effort into the area (a large part  
> of HTTP/1.0 is about optimising caching, but all that people seem to  
> want to know is how to most throughly defeat it).  In practice,  
> there tend to be two cachability classes:  never cache, and "I don't  
> care about how it is cached".  Typically HTML gets the first  
> treatment and images the second (in that case not having any control  
> headers).
>
> Specifically, in the above, I  believe:
>
> no-cache has no (specified) effect when sent from the server;
> must-revalidate is only meaningful if storing is allowed, so is  
> incompatible with no-store;
> no-store is intended to protect confidential data, not to guarantee  
> freshness; it may do so, but it is too aggressive.
>
> must-revalidate is probably enough here, although I think it would  
> be better to set an expiry time of, say, ten minutes.  One may need  
> Expires, for HTTP/1.0 caches.  I haven't re-reviewed the  
> specification to look for what fine tuning might be appropriate.
>
> In any case, it seems to me that this is about server and user agent  
> aspects that are orthogonal to the SVG statndards, and about  
> management and authoring policies.
>
> -- 
> David Woolley
> Emails are not formal business letters, whatever businesses may want.
> RFC1855 says there should be an address here, but, in a world of spam,
> that is no longer good advice, as archive address hiding may not work.
>