Definitions: caches, history buffers, etc.

Koen writes:
    The central point of your message, which seems to be that the 1.1
    draft does not define `cache' in a clear enough way, is better
    addresed by Roy.

    For the record, I do think you have some point when you say that
    browser authors can confuse caches with history buffers while
    reading the 1.1 draft.  But if we start calling history buffers
    history caches, we will only add to that confusion.

My opinions:

(1) History buffers, whatever you want to call them, are different
enough from HTTP caches that we need to continue to make this explicit
in the specification.

(2) Using the word "cache" to apply to both things seems to be
a recipe for confusion.  On the principle that confusion is bad,
and that we ought to be able to agree on almost any term (even
a made-up word), I suggest that we use the term "history buffer"
unless someone else has a better suggestion.  In particular,
let's avoid "history cache."

(3) We need definitions of "cache" and "history buffer" to
put into the spec.  I'm happy with Roy's definition of cache,
but in the absence of an equally concise definition of "history
buffer", I can see how some people could be confused.  So I
invite people to try to write a definition of "history buffer"
for use in the spec, and (if necessary) minor tweaks to Roy's
definition of "cache" to address any possible confusion.

-Jeff

Received on Monday, 8 January 1996 19:33:39 UTC