Re: Attaching docs, Virtual images

Jason B. Bluming (jbluming@chivalry.eit.COM)
Thu, 16 Dec 1993 16:32:06 -0800

>> I'll pick up the second for you. Your problem seems to be not on the
>> idea, but on the choice for implementation. While you could try
>Using a temp file for this is a very ugly kludge and has many problems.
>It's very simple and elegant to just return the image on the fly
>when it's requested. That way you don't have to worry about the
>numerous cases when people request the document but don't request
>the image (like delayed image loading, or browsers that don't support

I don't think we're disagreeing on these issues. I was just pointing
out that before you decide upon a particular implementation, it would
be beneficial to consider how it is being used. For example, say I
want to generate a graph of server utilization for the past 24 hrs --
for any arbitrarily complex figure being generated, it is not unreasonable
to assume that the file generation (and associated data compilation)
may be non-trivial, and separating the access time from the generation
time might be less irritating to the user (ie. by activating delayed image
loading if THEY so choose). In either case, I won't argue with
advocating an elegant solution, as long as it addresses the problem being