We do photographic processing and have had no problem, after using this
technique,
of deciding which images to get from the cache and which from the server
(retouched photos).
We haven't found any further GETs to the server if the images are already in
the array!
The preloading (or better still postloading) of images has been standard
technique for rollovers.

One could even create the hidden frame without an array but straightforward
IMG tags and just copy
the images across from the tags instead. Nothing would be reloaded here!!!

We have pages with up to 80 images/page and need to control this very finely
for maximum
efficiency, only downloading the images that have been altered between
redisplays.

To force server side loads it is necessary to add a random parameter to the
image source to
prevent the image being loaded from the cache, but then that is the opposite
problem to that
being discussed here (which was actually our problem).

I can't see how there could be any roundtrip checks when reloading a frame
WITHOUT images.
The other frame already has the images (in its array) and it is not being
reloaded.
It's images were downloaded the first time the frame was created.

Where is the roundtrip supposed to be occuring?
(Netscape has a problem if images of different sizes are switched in an image
tag.
But that is a different problem from the one being addressed here.)

Avron Polakow wrote:

> In both browsers IE and NS, this is a browser preference setting!!
> The user can override your own programming techniques.
> In IE you set this with: Tools > Internet Options > General > Settings >
> Check "Every Time you start Internet Explorer"
>
> WORKAROUND
> ~~~~~~~~~~~~~
> An easy workaround is to use an hidden frame with the technique of reading
> the images into an array.
> Then in the target frame create images with NAME attributes and empty
> <SRC>s and
> then copy the images from the hidden frame array of images after the
> target frame as loaded
> (from the onLoad event or by creating a user-defined postevent).
>
> I don't know if the browsers actually place images placed in arrays in
> their caches or keep them in their intestines.
> Check it out!
>
> BTW, you can use this technique to choose which images you want to
> download again.
>
> You will find that IE is VERY SLOW using arrays. Benchmark this technique
> against NS.
>
> Larry Serflaten wrote:
>
> > I've looked high and low for a method to force images to be cached.
> >
> > I have images displayed on an HTML page that I want to cycle other
> > images into.
> >
> > (Using IE 5.5) I have already tried these methods
> >
> > - Using the EXPIRES meta tag, setting to expire at a later date
> > - Using hidden IMG objects, setting the src property
> > - Creating a 'global' array of IMG objects, setting their src property
> > - Using multiple attributes in the IMG tag:
> > <img src="pic.gif" src1="pic1.gif" src2="pic2.gif">
> > Then attempting: myimage.src = myimage.src1
> > - Defining an image class with a public src property
> >
> > In each case, IE will try to download the image again, slowing down
> > the response time.
> >
> > How is this done, so that IE does not try to download an image on
> > each change (click)? I have several places on the page that will
> > all use these same images so I can't just place them all on top of
> > each other, I have to be able to assign them to the various images
> > on the page.
> >
> > I am trying to do this with VBScript. My current test page is at:
> > http://www.usinternet.com/users/serf...games/test.htm
> >
> > Any help is appreciated!
> >
> > LFS