Subject: Re: Competition by internal expertise for F/OSS vendors
From: Thomas Lord <lord@emf.net>
Date: Wed, 03 Sep 2008 14:31:05 -0700
Wed, 03 Sep 2008 14:31:05 -0700
Chris DiBona wrote:
>
>  
>
>
>     He does argue that there will be a backlash against, e.g., Google
>     and Facebook.   Things like Facebook will fall quickly out of
>     fashion (in a 10 year view).   Google's ad-broker business will
>     collapse when the pricing bubble for ads breaks and when users
>     begin demanding a curtailment of the creation of a privatized
>     secret police.   Meanwhile, billions more humans will personally
>     own and control computers with plenty of speed and storage
>     to do quite a bit on their own.  That personal-computing-on-the-
>     edge bit is where he and Stallman see the strategic action.
>
>
>
> So I honestly don't understand where you are going here. You are 
> mixing up auction based ad markets with secret police with license 
> politics with computer storage trends. Care to clarify?

Auction based ad markets compete among one another
in the quality of the demographic computations used in
pricing and placing ads.    The ad broker gets ahead of his
competitor by knowing his ad recipients better, either or both
as individuals or in terms of aggregate trends.  The ad broker
gets ahead by inserting stochastically successful control
signals into its ad recipients' environments.

Therefore, the ad broker is a kind of professional stalker-
paparazzi, always collecting data on users and keeping
it as trade secret.  So rich is this data that we have as a
result current struggles regarding government's rights
to copy that data.

Think about that:  off of one loading dock the ad broker
is selling "social influence" derived by stalking and
harassing users;  off the other loading dock the ad broker
delivers intel to the use-of-force monopolist.

Secret police are in exactly the same business.

So that is how ad markets and secret police relate.

License politics relate this way:  the ad broker's
use of software is entirely an exercise of the rights
we call software freedom.   Thus, if we value
software freedom highly, we must conclude that
we ought do nothing at all to take away the ad-broker's
right to use software that way.

One specific conclusion would be that the
Affero variation on the GPL is the wrong idea.
It attempts to curtail three expressions of software
freedom:

(1) to run a program without restriction;
(2) to modify a program without the obligation to
report your modifications to anyone.
(3) to develop a program without sharing it.

Since we wish to leave those rights in tact,
it makes no sense to try to use licensing
to forbid those practices.

That is how licensing relates to ad brokering
and secret police.

Technologically speaking:

Laptops and personal computers aren't actually
going away anytime soon.  On the contrary,
user populations are exploding in size.  At the
same time, these devices continue to become
more powerful (e.g., more storage, more cycles,
more pixels).

The *capacity of the edge* is growing by
leaps and bounds, even when compared relative
to the growth of capacity server-side.   Sure,
Google is putting up a lot of server farms but
have you seen global laptop sales?

P2P and similar application architectural approaches
can do a lot with all that power at the edge and while
it can't make "drop-in replacements" for all centralized
control web services -- nevertheless, a distributed
computing approach *can* create economic *substitutes*
for most centralized control web services.

For that reason, the FSF wing of the free software
movement remains focused on software freedom in
the same way it has been all along -- it regards the
current influence of centralized web services as a
self-limiting, transient phenomenon.

Thus, the possibility for resistance against the trend
towards secret police cum ad brokers is in part assured
by advances of edge-node hardware.

That is how storage trends relate to the other topics.

To that view from Moglen I would add only that
centralized services are sometimes technologically
necessary -- we do want there to be some big server
farms.  What I have been proposing are technologies
aimed at virtualizing such server farms in ways so as
to give users control and privacy over what runs,
server-side.


-t






>
> Chris



Chris DiBona wrote:

 

He does argue that there will be a backlash against, e.g., Google
and Facebook.   Things like Facebook will fall quickly out of
fashion (in a 10 year view).   Google's ad-broker business will
collapse when the pricing bubble for ads breaks and when users
begin demanding a curtailment of the creation of a privatized
secret police.   Meanwhile, billions more humans will personally
own and control computers with plenty of speed and storage
to do quite a bit on their own.  That personal-computing-on-the-
edge bit is where he and Stallman see the strategic action.


So I honestly don't understand where you are going here. You are mixing up auction based ad markets with secret police with license politics with computer storage trends. Care to clarify?

Auction based ad markets compete among one another
in the quality of the demographic computations used in
pricing and placing ads.    The ad broker gets ahead of his
competitor by knowing his ad recipients better, either or both
as individuals or in terms of aggregate trends.  The ad broker
gets ahead by inserting stochastically successful control
signals into its ad recipients' environments.

Therefore, the ad broker is a kind of professional stalker-
paparazzi, always collecting data on users and keeping
it as trade secret.  So rich is this data that we have as a
result current struggles regarding government's rights
to copy that data.

Think about that:  off of one loading dock the ad broker
is selling "social influence" derived by stalking and
harassing users;  off the other loading dock the ad broker
delivers intel to the use-of-force monopolist.

Secret police are in exactly the same business.

So that is how ad markets and secret police relate.

License politics relate this way:  the ad broker's
use of software is entirely an exercise of the rights
we call software freedom.   Thus, if we value
software freedom highly, we must conclude that
we ought do nothing at all to take away the ad-broker's
right to use software that way.

One specific conclusion would be that the
Affero variation on the GPL is the wrong idea.
It attempts to curtail three expressions of software
freedom:

(1) to run a program without restriction;
(2) to modify a program without the obligation to
report your modifications to anyone.
(3) to develop a program without sharing it.

Since we wish to leave those rights in tact,
it makes no sense to try to use licensing
to forbid those practices.

That is how licensing relates to ad brokering
and secret police.

Technologically speaking:

Laptops and personal computers aren't actually
going away anytime soon.  On the contrary,
user populations are exploding in size.  At the
same time, these devices continue to become
more powerful (e.g., more storage, more cycles,
more pixels).

The *capacity of the edge* is growing by
leaps and bounds, even when compared relative
to the growth of capacity server-side.   Sure,
Google is putting up a lot of server farms but
have you seen global laptop sales?

P2P and similar application architectural approaches
can do a lot with all that power at the edge and while
it can't make "drop-in replacements" for all centralized
control web services -- nevertheless, a distributed
computing approach *can* create economic *substitutes*
for most centralized control web services.

For that reason, the FSF wing of the free software
movement remains focused on software freedom in
the same way it has been all along -- it regards the
current influence of centralized web services as a
self-limiting, transient phenomenon.

Thus, the possibility for resistance against the trend
towards secret police cum ad brokers is in part assured
by advances of edge-node hardware.

That is how storage trends relate to the other topics.

To that view from Moglen I would add only that
centralized services are sometimes technologically
necessary -- we do want there to be some big server
farms.  What I have been proposing are technologies
aimed at virtualizing such server farms in ways so as
to give users control and privacy over what runs,
server-side.


-t







Chris