So much of the time, when we talk about network services,
technological/software freedom, and this idea of "Cloud" computing,
there's a bunch of debate: "is it a good idea?" "are we giving up too
much freedom?" "how does this work out economically?" "what about
privacy in the cloud?" While these are important questions, without
doubt, I fear that they're too ethereal, and we end up tussling with a
bunch of questions about the future and present of computing that might
not be entirely worth debating (at least for the moment.)
Lets take 2 assertions, to start:
1. There are some applications--things we do with technology--that work
best when these applications are running on high performance servers
that have consistent connections to the Internet, that we can access
regardless of where we are in the world.
2. The only way to have control over your data and computing experience
is to be responsible for the administration and maintenance of these
I mean to say, that if we care about our autonomy, and our freedom as we
use computers in the contemporary age (i.e. in the era of cloud
computing), the only thing to be done is to run our own services. If the
fact that Google has all of your data scares you: run your own mail
server. If the fact that all of your microblogging output is on twitter,
run your own status.net instance. And so forth.
If we really care about having power over our technological experiences,
we must take responsibility for services on the Internet. We can say
"wouldn't it be nice if service providers weren't such dicks with our
data," or "wouldn't it be nice if software developers wrote networked
software that respected our freedom." And while it would be nice,
these convinces don't in and of themselves
Control over technology and autonomy in the networked context ultimately
means that we as users have to:
- Administer networked servers that provide us with the services that
we want and need to do whatever it is that we do.
- Participate in some exchange for networked services (i.e. pay for
service, either in cash or by way of access to data.)
That's hard! Computers should get easier to use not harder, right?
Leading question there, but...
Yes. One of the leading arguments for consumer-"Cloud Computing" is that
by accessing computer services (software) in the browser, developers can
provide a more structured and "safe" user experience. At least that's
how I understand it.
While this is a great thing in terms of making computers more
accessible, no argument from me, I think we must be careful to avoid
confusing of use" with technologically limiting. I fervently believe
that its possible to design powerful software that is also easy to use,
and I think that as often as not, a confusing technology is an
opportunity to provide a teaching experience as much as it presents an
opportunity to improve a given technology.
And if it comes down to it, there are situations where it doesn't matter
so much if you're the one entering the commands into the server. It
doesn't much matter if you are the one managing the server or if
you've hired someone to configure it for you. As I think about it,
there's probably something of a niche here for people to offer
management services in a very boutique sort of style.
If we have to contract to people to do our administration for us, is
that really a step in the right direction?
I think, it is. At the moment we pay for our networked computing
services (i.e. gmail) by looking at google's ads next to our mail and
giving Google access to the aggregate of our mail spools so that they
can mine it for whatever data they need. The other price that we pay for
these services is "lock in:" once we commit to using a service it's
quite difficult to change to an alternate provider. Since these are
"real costs," it seems reasonable to expect and want to pay (money)
for services that don't have these costs. Which is where cooperative and
boutique-style services make a lot of sense.
I'm not a systems administrator, I just want to do [the thing that I
do] and not have to tinker with my computer. This is a lousy idea.
And that's a lousy question.
To dig in a bit further. I don't think that "doing the [whatever you
do]," would necessarily require a lot of tinkering. It might, of course,
and the chances are that we've all had to tinker with our technology at
one point or another. In most cases tinkering is an upfront rather than
ongoing cost. Ideally, the other thing that having full control over
your network services you'll be able to use have services which are
more tailored to [the thing you do] than the one size fits all
application provided by a third party.
Ok, so what's the stack look like.
I'm not sure. There's clearly a common set tasks that we currently use
in the networked context. I'm not sure what the application is, exactly,
but here's a beginning of what this "application stack" looks like.
- An XMPP Server like Prosody.im, with PyAIMt and other convectional IM
- Some sort of Email Service: Citadel springs instantly to mind as an
"all in one solution," but some
postfix+procmail+fetchmail+horde/squirrelmail seems to make some
- A web server, either for hosting personal websites, or with some sort
of authentication scheme (digest?) for sharing files with yourself.
The truth is that web servers, are pretty darn lightweight and it
doesn't make sense to not install one. Having said that, people see
"web hosting," and probably often think "Well, I don't really need
web hosting," when that's almost beside the point.
- SSH and some system for FUSE (or FUSE-like) mount points, so that
they can use and store remote files.
- There's probably a host of web-based applications that would need to
be installed as a matter of course: some sort of web-based RSS
reader, wiki-like note taking. Bookmarking. Some sort of notification
- [your suggestion here.]