[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: fiat mode on regarding WWW and documentation (fwd)

Forgot to cc the list on this one.

---------- Forwarded message ----------
Date: Mon, 30 Jun 1997 19:33:21 -0500 (EST)
From: branden@apocalypse.sequitur.org
Reply-To: branden@purdue.edu
To: Bruce Perens <bruce@pixar.com>
Subject: Re: fiat mode on regarding WWW and documentation

On Mon, 30 Jun 1997, Bruce Perens wrote:

> Bruce:
> > Huh? Why not uncompress them on the server and serve them to the browser
> > as HTML? That would have the advantage of working well :-)
> From: <branden@apocalypse.sequitur.org>
> > Bandwidth/speed.  Though I know that would require hacking the browsers to
> > re-examine the files for parseable content after they're uncompressed
> > locally.
> In the case of browsing Debian documentation, the system is the local
> host or is on your local network. Compression isn't going to improve
> your browsing speed on a loopback or a 10 MBPS ethernet.
> Anyway, this is a properly a function of the data-link layer.
> Compression is done for you by the LZW compression in V.42 modems,
> and for ISDN users by LZS compression in their hardware or by Berkeley
> compression in PPP. And of course you know that sending compressed data
> over a compressed link doesn't make it faster , so you don't want to
> duplicate the compression at the application level.

Okay, I yield the point -- but I still notice file transfers over my PPP
line going faster when I gzip the stuff first.  I can only assume that the
LZW/Berkeley compression involved is not maximally efficient because it
only has so much time to work on the data before it much ship it out the door
(or lose the beneifit of compression by introducing waits).

I guess I just don't trust the data-link layer to compress with the
ruthless efficiency that an agressive piece of software tied to a powerful
CPU can.

I'll admit I don't have any hard figures on this, though it seems to be
true from careless anecdotal experience.  Would anyone care to enlighten

At any rate, Bruce, your take on the documentation policy is in agreement
with what I feel with the above exception, which is a nit, so I at least
won't be screeching at you about it.  :)

G. Branden Robinson                 |  Somebody once asked me if I thought sex
Purdue University                   |  was dirty.  I said, `It is if you're
branden@purdue.edu                  |  doing it right.'
http://www.ecn.purdue.edu/~branden/ |  -- Woody Allen

TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .

Reply to: