What Does Refresh Do?

[TC]

I always thought that Refresh (F5) went out to the internet to get the
latest version of a page. However, a simple experiment tells me that
isn't right: I have google.com open in another window, and when I
click Refresh, nothing changes, even though I've disconnected from the
internet. If Refresh were loading the latest version of google.com, I
should be getting a "The page cannot be displayed" error.

So, if it doesn't reload the page, what does Refresh do? Conversely,
is there any way to reload a page?


-TC

 

[Rob ^_^]

Hi TC,

F5 = Refresh from cache
Ctrl+F5 = Refresh from web

Regards.

 

[TC]

Hi TC,

F5 = Refresh from cache
Ctrl+F5 = Refresh from web








- Show quoted text -

Why would you ever want to refresh from the cache? Wouldn't that just
give you the same page which is already loaded? I thought the purpose
of refresh was to see if the page, as served over the web, has
changed.

-TC

 

[Robert Aldwinckle]

Why would you ever want to refresh from the cache? Wouldn't that just
give you the same page which is already loaded?

It the cache is current that would be sufficient.

I thought the purpose of refresh was to see if the page,
as served over the web, has changed.

It does. But the real purpose of caching is to save bandwidth.
E.g. instead of sending a request which says: Download this page
(which may be large), cache-checking asks: I've got this page
already downloaded; is it still current? If it's current the server
gives a (very small) response which tells the client to go ahead
and used the cached copy.

A problem with this scheme occurs when there is an intermediate
caching server which does not forward the cache-checking request
to the real host and instead just responds with an Ok to the check
of the (stale) content. That's when you need the Ctrl-F5 which then
sends an HTTP request as if the file was not cached. Usually that
is enough of a clue to the caching server to send the request all the
way through to the real host. Another thing which may work is changing
the protocol to https (provided the host server supports that protocol)
since then the transmission is encrypted and can't be cached.
Another thing which may force the request through is tacking a question
mark on the end as if you were making a CGI request, since again
caching should not be done. And again, a caveat would be that the
host server would have to be tolerant of this misuse of the question mark
when requesting the particular page you wanted to have downloaded.


HTH

Robert Aldwinckle
---