feature bonanza

marca@ncsa.uiuc.edu (Marc Andreessen)
Date: Mon, 4 Oct 93 01:36:07 -0500
From: marca@ncsa.uiuc.edu (Marc Andreessen)
Message-id: <9310040636.AA06764@wintermute.ncsa.uiuc.edu>
To: www-talk@nxoc01.cern.ch
Subject: feature bonanza
X-Md4-Signature: e74131e827577dedc9aa2015dfc0dc47
I notice a certain enthusiasm on this list for lots and lots and lots
and lots of little features :-).  Unfortunately, clients like Mosaic
are already approaching the dreaded state of kitchen sinkhood.  So
with that in mind...

How much would a feature like image prefetching or packaging or ...
really add, relative to the overhead and complexity involved in
supporting it?  One would get a marginal performance increase on
low-latency, high-bandwidth networks (which are rare -- in life, it
seems to be either high/high or low/low, excluding certain very rare
special cases) but not much else.  Probably wouldn't even be

So I might suggest that design discussions take into account
practicality and usefulness in a system as complex as (e.g.) Mosaic
2.0 plus Plexus 3.0, and try to triage below a certain level, in

Between HTTP/1.0, fill-out forms, and the other recent/coming
advancements on both client and server ends, we're nearing a really
powerful general-purpose information environment with considerable
extensibility; at this point we may want to focus on applications and
advanced functionality rather than "under the hood" types of things
like image packaging.

Just a thought,