More on walled gardens
----------------------

My previous entry on walled gardens appears to have created quite a bit of
interest.  So far I've seen two phlog entries written in response and have
received two or three emails.  Thanks to everybody who shared their thoughts and
sorry it's taken me so long to continue the conversation.

First of all, I just wanted to clarify that in no way was I proposing any kind
of overall moral equivalence between SDF and Facebook.  The scope of the entry
was supposed to be entirely limited to the "walled garden" aspect of SDF.  Even
if we were to take it as a given that being a walled garden is fundamentally and
invariably evil, this is of course just one of many dimensions on which the
morality of on online service can be judged.  Facebook do plenty of other evil
things which I am supremely confident SDF does *not* do (such as closely
tracking the online activity of people who are not even users of Facebook), so
there is no question in my mind as to which is the more respectable and
trustworthy system.

Jandal responded to my post remarkably quickly with one of his own, which I
highly recommend you read if you haven't already:

gopher://grex.org/0/%7ejandal/phlog/communication-expectation-violations

Jandal's entry actually has applicability well beyond the discussion I started
and I would say it presents a very solid conceptual framework for thinking about
internet communication systems in general.  The entry taxonomises communications
as being public, aimed at an in-group or private (and SDF presents services
facilitating all three kinds of comms) and argues that sevices become evil
whenever they violate some specific user expectations about how each kind of
communication works.

I think I essentially agree with everything Jandal has to say.  Implicit in this
is that a walled garden (i.e. a system for communicating with an in group) which
does a good job of being a walled garden (i.e. restricts access to the intended
in group and maintains a relatively fixed definition of said in group) is a
perfectly respectable system.

I am having a strangely difficult time digesting this.  When I ask myself "is it
reasonable to want to communicate exclusively with some in group?" the answer I
get immediately is "yes, of course, why is this question interesting?".  But in
the original formulation of "Are walled gardens ever okay?" the situation seemed
far less clear.  I suspect I have to some extent fallen victim to the fact that
the term "walled garden" is exclusively in a pejorative sense and so has become
a kind of thought terminating cliche - if something can be accurately described
as a walled garden, it is automatically bad.  Slugmax is bang on the money when
he calls this "lazy thinking":

gopher://sdf.org/0/users/slugmax/phlog/walled_gardens

But "walled garden" would presumably not have become a slur amongst nerds if
there wasn't a history of walled gardens having other undesirable properties.
So the real matter of importance here is to identify those properties and
thereby distinguish harmful from harmless walled gardens.

Slugmax suggests that one important dimension on which walled gardens vary is
the amount of anonymity they afford their users, and notes, correctly, that SDF
performs exceptionally well in this regard.  I agree with this entirely, but I
suspect it is just one of many factors important to the acceptability of walled
gardens.

Another might be the cost of becoming an in group member - cost in a general
sense, not just financial.  The service might send you an email once a month
asking for a donation to keep things running.  That's not so bad.  Or, as a
result of signing up, you might have every single thing you do online closely
scrutinised and sold to third parties.  That's awful.

However, both of these issues are nothing unique to walled gardens.  A
communication system designed to facilitate purely public communication may not
permit anonymous sign ups or may engage in tracking, intrusive advertising or
all sorts of other unpleasant things.

At this point, I am starting to consider the possibility that while there is no
logical connection between being a walled garden and featuring various other
undesirable properties, there has historically often been a strong correlation
between the two.  Restricting access to content to an in group is a more
difficult problem that permitting public publication, and solving the problem
may often lead implementers to solutions which tend to make it easier to
implement other unpleasant features as well.  I'm not really convinced that this
is the core of the issue(s), though...