Subject: Re: Proprietary vs. Web-based From: Rex Ballard Date: Thu, 24 Oct 1996 01:55:02 -0400 (EDT)
How the Web Was Won
Subject: Re: Proprietary vs. Web-based From: Rex Ballard Date: Thu, 24 Oct 1996 01:55:02 -0400 (EDT)
In-Reply-To: <199610200157.VAA07123@mailnfs0.tiac.net>
Message-ID: 
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII



	Rex Ballard - Director of Electronic Distribution
	http://cnj.digex.net/~rballard


On Sat, 19 Oct 1996, Curt A. Monash wrote:

> The question of proprietary vs. Web-based services has reared its head again.
> 
> In my opinion (unchanged for some months now!), the winning model will be a
> hybrid between the two.

I would like to make the case for the Open Standards that made the
internet and the World Wide Web possible in the first place.

> AOL has *not* succeeded in dramatically improving or changing the nature of
> its content: Chat, chat, chat, and more chat, generally mindless (in fact,
> generally focused about 60 centimeters below the mind); generic connectivity
> (their technical problems are different than the Web-based ones, and more
> tolerable to many people); a couple of other communities (e.g., Motley Fool,
> one of the most successful notegroup offerings out there); stock prices; a
> *little* bit of news.  (They don't even have Rosalind any more!)

AOL's chat and board groups are more like these prepackaged amusement
park type environments.  Its like comparing an olympic size swimming pool
with the Pacific Coast.  In the olympic pool, you have to contend with
thousands of people, clorine, people's disgusting habits (taking a leak in
the pool), and the occasional outbreak of violence.  On the other hand you
don't have to deal with jelly fish, syringes washing up on the shore, or
undertows.  What's possible on the pacific coast though is the ability to
surf for extended periods, the fun of diving into a wave, and the fun of
being carried by the waves, currents, and tides.  In the olympic pool,
it's covered so you don't have to worry about rain or bad weather, but you
can't go windsurfing either.

On AOL you can participate in limited discussions in relative privacy, and
not have people flaming you for something you posted out of ignorance.

On the Internet, a news group posting bay be distributed and read by
thousands of people all over the world.  Cencorship is not an option so
you are flamed when you don't have your facts correct or when you post
something inappropriate.

> Except for the chat, this is pretty much what Prodigy offered.   

I still use prodigy for electronic banking.  It's about the only reason
I use it.  My bank doesn't trust the internet yet.

> Thus, as an aggregator of -- well, of whatever they aggregrate -- they're
> within the reach of Web-based services.
One big difference is that Content on Prodigy isn't archived by
government, business, financial, and media databases.  I have freqently
received requests for followup on postings that I distributed months ago.

> Meanwhile, the ISPs (especially Netcom), the search/directory sites (led by
> Yahoo), and, yes, the newspapers are all broadening their aggregations
> (content, software, and branding alike).

Let's break this down a bit more carefully.  Within the umbrella of
service provider we have distinctions such as:
	 Independent Access Provider (IAP),
	 Independent Service Provider (ISP) - Mail, News, Web Hosting.
	 Independent Content Provider (ICP) - Publishers
	 Internet Clearing Services Providers (ICSP) - ClickShare, Digicash
	 Internet Search Services Providers (Yahoo, Infoseek...)
	 Redistribution Services (Hosts which serve corporate content).
	 There are "Home Pages", "Infomercial Pages", and
	 "Commitment Pages", which may include independently hosted
	 Artwork, CGI Forms, Text content, and Java streams.

A "simple" web page may contain content from 20 different sources and
traverse 50 servers, routers, firewalls, and gateways in the process.

This is only possible because of the open standards and GPL reference
models which constitute the foundation of the Internet.  Companies like
Netscape succeed because they provide an edge in the quality of service.
The same document that takes 5 minutes to load using Mosaic is available
within a 10-20 seconds on a Netscape engine.  The Netscape server isn't
substantially different from other servers, but it can be configured using
simple menu choices instead of having to look the choices up in a book
somewhere and edit a text file by hand.  A simple Linux or Unix box has
all of the "Standard Ingredients", but a Netscape Publisher server can
parse input feeds, nntp feeds, and e-mail to create classy looking
displays from "Flat Text" ascii files.

While Lotus and Microsoft tried to beat each other to death over the
e-mail market, the open systems approach created the possibility of
taking the "MIME" packets used for Open Systems Standards email
and create a different "read-on-demand" client/server protocol (HTTP)
which could pass the same packets much more efficiently and with less
system overhead.  The big difference was that Open Systems only had
to jump a few "inches" where has Lotus Notes and Microsoft Exchange
had to designed from the beginning meaning they had to travel "Miles".

Apache and Arena are already creating the possibility of commerce over
internet using Open Standards.   The source code for the implementation
is available and developers can simply offer patches to support clearing
services such as Digicash, FirstVirtual, ClickShare, and e-cash.  It also
can be used to protect high value news feeds similar to the way scramblers
protect sattellite signals.  Simple encryption systems like DES can
implemented with relatively low risk.

Internet Explorer and Internet Information Server contain hooks for
proprietary content that can be hazardous to your network.  Since it is
possible for a Microsoft Word Document to contain imbedded macros and
executibles which can to amazing things like reformat your hard drive or
sent all of your software serial numbers, disk drive serial numbers, and
web browser access history to the author.

> What's more, as I've argued before, aggregation is technologically
> unstoppable.  Most content will NOT be available in static HTML, but rather
> available through various sorts of programs, database-oriented or otherwise.  

HTML si merely a presentation protocol, a restricted subset of SGML.
Attempts to promote full SGML on the Web have failed due to the lack of
supportable GPL reference implementations.
> 
> Why does this matter?  Because unless the major service providers cache your
> material, access to your site is likely to be horribly slow and unreliable
> for several years.

You underestimate the resourcefulness of the collective intelligence of
the "Internet Organization".  The current bandwidth problems can be quite
easily solved.  Ironically, the earliest solutions will come as upgrades
to General Public License Reference implementations.  The interesting
thing is that GPL cannot be used by companies like Microsoft.  All of the
original autheors and promoters would have to be compensated.  Even "black
box" reengineering would be a copyright violation.

>  But they can't "cache" a program without your active
> cooperation.  Similarly, no search service can index your database-generated
> dynamic content, without your cooperation.  Thus, SITE OWNERS WILL HAVE TO
> WORK WITH AGGREGATORS, at least if they want access to their site.  

We are coming full circle back to the original concept of the original
WAIS server and the original Z39.50 protocol.  In this model, a single
query can result in a simultaneous search of an entire group of search
engines.  Also, content can be managed using smart domain name servers
that can determine which coast or geographical area you came from before
giving you the IP address of the nearest server (kismet).

> With this technical imperative, I expect all sorts of business arrangements
> to blossom.  Web-based aggregators will have plenty of chances to succeed.
> 
The original possibility of the WEB was to make economic opportunity
available to the widest number of people.  Any teenager with a 486 PC
can create a Web Server using Linux.  The infrastructure is sufficient
to support about 100 users.  Beyond that, if he has managed his revenue
stream, he can upgrade to Sun, SGI, or HP with a minimum amount of
"Retooling". and support several hundred, even a few thousand, users.

Tools like PERL enable the output from search engines and news feeds to be
transformed into HTML content in near-real time.  Alerts such as a
scrolling headline display can be fed using a PERL script and a JAVA or
JAVAScript engine. 

How mush is a web server?  How much do you want to pay.  The Linux server
on a P75 can run as little as $1000.  The superservers that can do
real-time Virtual Reality Modelling (VRML) to JPEG/MPEG conversions are
budget-busters.  A full-integrity ethical server based on NT and Oracle
can run several million dollars.

> 
> CAM
> 
> 
> Posted to ONLINE-NEWS. Made possible by Nando.net - http://www.nando.net
> 


From rballard@cnj.digex.net Thu Oct 31 02:14:01 1996
Status: O
X-Status: