Subject: Re: Copyright, The Internet & Low Cost Universal Commercial Info From: dowjone!rexb (Rex Ballard) Date: Fri, 25 Feb 94 23:59:04 EST

         Services.
Cc: smw@veritech.com, jpalme@dsv.su.se, fig@eff.org, proyse@peeras.demon.co.uk,
        online-news@marketplace.com, mhsnews@ics.uci.edu
Sender: jvncnet!marketplace.com!owner-online-news
Content-Length: 4907
X-Lines: 99

> Lee,
> my personal opinion is that 5 years after Public-Key cryptosystems
> become common, people will have the same opinion of unsigned
> information on the Net as they have of leaflets handed to them
> on Main Street.

The internet provides three arenas.  It shouldn't lose it's identity
as a public forum.  The internet provides one of the few equivilents
to an open forum for open debate of national issues.  It's a great
deal of noise, but for every megabyte of flames, there is usually
one or two lines that will eventually become policy.

> Unfortunately, starting the clock on those 5 years probably requires
> passage of bill H.R. 3627 in the US Congress (disemboweling the NSA).

For 14 years, the IRS and Treasury have been resisting the use of DES
encryption, making it a crime to export it.  They do this while hiding
behind the motives of wanting to catch terrorists, drug dealers, and
child molesters.  What they're really after is unreported taxable income,
insider trading information, and import/export tarrifs.

> Remember that there are 3 kinds of information around:
 
> - Things that the originator wants anyone to have (like advertising,
>   or FAQ publications). There, the problem is solved when the
>   signature is verifiable, because the only real check is "did the
>   sender really say this".
Fortunately, there is plenty of this to go around.  We might even find
a way to charge admission to the forums.  This "Advertizing" could bring
substantial revenue.

> - Things that the recipient wants to have, and the originator wants
>   to get paid for, but does not care who the consumer is.
>   I agree - this is a BIG problem.
Only when using a micromodel instead of a macromodel.  Saying you want
$4 bucks every month I call your server, even if I only call it 3 times/year
is a nuisance.  What am I getting?  What is the service being performed?
 
>   Among other things, one needs:
>   - An efficient means of referencing information *without* quoting it
Accessing WWW or WAIS servers using an acession numer would help.  This
is still a micromodel approach.  Of course, it also begins to take on the
flavor of OLE.  You have to wait until the whole object is retrieved before
you can view it.

>   - An efficient means of accessing information through a reference,
>     available to all parties

Considering the bulk of information most people deal with today, the
"Net Traffic" will probably only be cached to the disk.  The user is
willing to pay $30/month for a gigabyte of information he won't use,
and a kilobyte that he will.  Pictures, sound, and video make the
"cache factor" even more significant.  Most people won't even cache a whole
2 hour movie locally.  They would just pick the channel, start caching
the compressed video, and get 20 or 30 meg queued up before starting the
view.

The internet represents about 200 million people trying to access 200
tarabytes of data.  It's value is directly proportional to the time
it spends in the cache.  The user will pay for access to the right megabyte.

>   - A "net.ethic" that regards copying without proper chargeback as
>     an immoral action

We just need to manage the access to the network the same way we manage
access to the public roads.  People buy gas, oil, vehicles, and tires.
On the internet, they will buy access time (months), media (short term
storage), Computers and CSUs, and electronic purchasing.

>   - The opinion I mentioned before, giving zero trust to anything
>     where the path of information copy cannot be verified.
Again, the pure bulk of the information available today almost mandates that
the information needs to be loaded directly.  The average workstation
probably won't have 20 gig disk drives to keep 30 or 40 "sound bytes".

Being able to get that from a locally cached system makes sense.  OSF provided
secure versions of Telnet and FTP, perhaps versions of WAIS, MIME, and
WWW can support transparent registration of documents by acession number.
A classic model is Source, Date, sequence number for this source at this at this date.  A more precise model would be IP Address, Date, Time, and Sequence
Number.  This could be matched with the IP Address, Date, Time, and Sequence
number of the Respondant.

> - Things where the sender has a definite opinion of who should read
>   this stuff. This is probably simpler, because a finite number of
>   readers allows you to operate key distribution with finite cost.
>   Again, a proper identification scheme is vital.

As mentioned before, this is the "gold" which will actually pay for the
publications you mentioned above.  People have credit cards which can be
verified using a reader at the grocery store.  The equivalent would be
DES keys passed to servers which then call the "banks".  The NASD has
been using something like it for a few years now.

>                        Harald T. Alvestrand


	Rex Ballard



From jvncnet!marketplace.com!owner-online-news Mon Feb 28 10:20:36 1994