Re: Last-Modified vs. Expires

--MimeMultipartBoundary
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT

> From: Duane Wessels <wessels@nlanr.net>
> Forwarded by: squid-dev@nlanr.net

> dgaudet-list-squid-dev@arctic.org writes:
> >
> >Were there images marked with Last-Modified but no Expires? In that case
> >does squid requery on every hit, or does it implement its own staleness
> >timer?
>
> I don't know if the images had expires. They certainly had last-modified.
>
> But the point is that even with expires or last-modified, the
> 'pragma: no-cache' request causes a cache miss. A surpsising
> number of requests include no-cache. I'm not sure why, maybe
> everyone hits reload all the time, or maybe some user agents
> always include it.

    Salute goes to Mr Bill. Internet Explorer is dumb in two ways: in its
 default config, it is unable to do Reload - upon pressing Refresh it simply
 rerequests page from cache. Users are complaining about some URLs being
 stale and NOT being able to reload them. Solution to this is really MS
 style: check a box in prefs ala "check page every time" and bingo, IE
 passes "no-cache" along with EVERY single request! This has been tracked
 for IE at least up to ver. 3.01.
    More over, few our customers have stated that after enabling proxy
 cache in IE, IE stops using its disk cache! This might well explain why
 too many cache hits for a given URL are in a time frame of few hours.
 The worst result is that dialup users HATE caches, and I could imagine
 why - IE requests every object from cache, and every object with no-cache!
 Even when pressing Back button... every object they browse comes via
 modem link, and from the source. They simply MUST hate caches...
 Although I cannot confirm that IE stops using its disk cache completely,
 as I haven't bothered to install one, but I'd not be surprised at all...

> So what I'm after is a way to safely ignore no-cache. To give
> the cache a way to say "I'm going to ignore your stupid no-cache
> request because I KNOW this object will never EVER change."

    Be that fatal is dangerous. Consider some site once giving you a page
 marked so for their homepage? Then finding oops, they want to change
 their homepage from time-to-time (Microsoft for eg?). Now a cache may
 ignore every request to reload the page? Remember MS home page last year
 that circulated in world caches for 2 months after it was obsolete?
 Although case was different, possible nuisance is similar. The browser
 must have a chance to enforce reload. Cache must itself detect when
 these reloads are too often and ignore these part time, and should
 issue IMS when it feels unsure.

 IMHO,
 best regards,

 ----------------------------------------------------------------------
  Andres Kroonmaa mail: andre@online.ee
  Network Manager
  Organization: MicroLink Online Tel: 6308 909
  Tallinn, Sakala 19 Pho: +372 6308 909
  Estonia, EE0001 http://www.online.ee Fax: +372 6308 901
 ----------------------------------------------------------------------

--MimeMultipartBoundary--
Received on Tue Jul 29 2003 - 13:15:42 MDT