How precisely does the Automatic cache setting work in IE6?

  • Thread starter Thread starter Sarah Howarth
  • Start date Start date
S

Sarah Howarth

Can anyone point me towards a detailed description of how
the Automatic setting for Temporary Internet Files works
in IE?

I am implementing a predictive caching tool, and but am
concerned that even if I push a bunch of my pages into
the cache, will IE
a) display them,
b) display them and send an IMS request at the same time,
or
c) completely ignore them and request the page from the
origin server?

Thank you in advance!
 
Can anyone point me towards a detailed description of how
the Automatic setting for Temporary Internet Files works
in IE?

Did you see this?

<title>KB263070 - How Internet Explorer Cache Settings Affect Web Browsing</title>
<quote>
Automatically (Internet Explorer 5 and later only):

This is the same as the previous setting, but with a logic algorithm to
understand the habits of Web page behavior. This setting specifies that
when you return to a page you viewed previously, Internet Explorer should
not check to see whether the page has changed since you last viewed it.

If you select this setting, Internet Explorer checks for new content only
when you return to a page that you viewed in an earlier session of Internet
Explorer or on an earlier day. Over time, if Internet Explorer determines
that images on the page are changing infrequently, it checks for newer
images even less frequently.
</quote>

(MSKB Boolean search for
cache AND automatically AND "Every Visit"
)


FWIW my interpretation of the Automatic setting is that it would be
an adaptive heuristic which would try to reduce the number of checks
which would be done for particular repeatedly used files, IOW a setting
which tries to give the accuracy of the Every Visit... setting but with
a more acceptable overhead. Note that this interpretation is not what
the precise wording of the article implies. It makes it sound more as
if it is trying to improve on the overhead of Every time you start...
which I think would be unnecessarily risky.

IMO Every Visit once had horrendous performance which made
Automatically (the default) desirable. However, lately I think that
Every Visit works much better and use of Automatically is not worth
the risk of lack of accuracy. The main thing that I can think of that
might have been done to make such a difference is preventing multiple
checking for repeatedly used files within one rendering.


In case it helps here is a previous reply about the cache checking
mechanism which contains a link to a relevant MSDN article.
I suspect that the implementation of this feature might be used to
implement the heuristic that the Automatically option effects.

<paste>
Here's a reply I gave to someone who found the relevant
page at the MSDN site. (Just my interpretation of what
I was reading.)

---

The only "No-check" in that article looks to me like an abstraction
to label the Green line in the diagram. I would say that

Post-check=mm
defines the end of the "No-check" interval
(green line in the diagram)
and the beginning of the "Pre-check" interval
(yellow line in the diagram).

Pre-check=nn
defines the end of the "Post-check" interval
(yellow line in the diagram)
and the beginning of the (indefinite) Pre-check interval
(red line in the diagram)

I wish that the article made clearer if HTTP/1.1 is a requirement
for this feature. There is a loose implication that it is.
For example here is an extract from the RFC referenced:
http://www.cis.ohio-state.edu/cgi-bin/rfc/rfc2068.html#sec-14.9
<extract>
Note that HTTP/1.0 caches may not implement Cache-Control
and may only implement Pragma: no-cache (see section 14.32).
</extract>


HTH

Robert Aldwinckle
 
Back
Top