Re: Control is Highly Overrated and Overpriced

Ken’ Hess has posted a blog article on ZDNet about how control over your own computer is overrated. This sentiment I feel is an attempt to embarrass people into moving their computing further onto the cloud.

This type of thinking also deeply effects the free and open source culture. Since one of the reasons for using FOSS is ultimate control (and responsibility).

From an individual perspective the goal of personal control is simple: You have this responsibility to provide this service and you do it with this property running this configuration. It’s human nature to want to control directly the service you’re responsible for. The other option is to pass over control to a good friend who you have a good positive relationship with (company or individual is irrelevant).

I think the failure of a speedy transition to “cloud computing” has been a failure in relationship building, but I’m sure that will come along in due time as the industry matures.

From a social perspective, having everyone on the same centralised system can introduce a fragility which can cause some interesting cascading and simple root failures which would be very bad for economy should enough businesses all move to the same few providers.

A lot of the people who would want their services taken care of are already not in a good mood from the 20 years of bullshit from the likes of Microsoft, as providers go we’ve had some fairly nefarious characters in control of everyone’s desktops.

I think it will take a while to turn that around, of course I’m putting my bets on distributed computing using things like the sheva plug or the free software router currently in development, because distributed resources that are properly designed can be much more interesting that centralised service prevision.

What are your thoughts?

Ubuntu One and FOSS Services

My good friend S.Gerguri asked me to talk about the nature of the Ubuntu One services offered by Canonical Ltd. and has sent me his thoughts by email, I’ve quoted him here and responded with my own thoughts. Full disclosure: I briefly worked on the team that develops the Ubuntu One service at Canonical and so I’m going to be careful since I’ve seen code and talked about strategy while on the team.

I just saw Martin Albisetti’s post on Ubuntu One Mobile and went to check the service. It appears to be a rather neat way of accessing your music while not having to lug around your whole catalog when you’re away just with your phone. … It is also priced quite fairly, even though I think the cloud capacity could be larger for the asking price.

The amount of space has to be very carefully calculated between the amount offered for free and the amount offered for a price. The servers the files are stored on are Amazon’s cloud, so the money doesn’t all go to Canonical. This is something of a business decision though and it’s really up to the market to decide at what price it thinks the services on offer are worth paying.

The service is also competing against companies for who this service is a loss leader, hoping to attract a large enough user base to sell their companies. This is precisely the opposite of Canonical that wants UO to be self sustaining and ultimately paying for developers to work on Ubuntu.1

The client code is open source, while the server is apparently under full control by Canonical. Apparently though, there are still people that have a problem with this, as evidenced by some of the comments in Martin’s post.

It’s true that the server component is proprietary, actually it’s also not even available in binary form. It might as well just be a magic wall that talks a certain language. My feelings on services is that the user has already made the decision to give their destiny over to the service provider regardless of whether the server code is FOSS or not.

For the client software, this is more interesting because the protocols are documented and well known so creating a server component is a matter of guile and not hard work. Accepting code in the FOSS client so it can connect to other servers using the same protocol and shipping that by default in Ubuntu is perhaps where Canonical’s community and open market principles will be shown either way.

The service in question here is the cloud storage along with _open source_ retrieval mechanisms (I am talking about the app). … So about the only possible super-wild argument against having Canonical complete control over the server side (including source) is code inspection for security issues, and even that one falls short because the storage software is just part of the server stack.

The security of the server side is more likely to be better than the client side anyway. But I don’t think there is any rationale to holding onto the server code, there are bigger sticks and better ways to use them. At the moment the closed server code is used as a weird proxy for user’s unsettling feeling that Canonical is making money from Ubuntu in ways where it doesn’t invite anyone else to make money too2.

Some people are also concerned that it’s a matter of principle. If you don’t as a company _believe_ in free and open source, then why advertise and promote free software principles at all? If you do believe them; then why the hypocrisy? That’s a social element which basically boils down to fear being the greatest eroder of principle and it’s fear of being out-competed on the Ubuntu platform which keeps it closed.

I actually find this pretty insulting from the people that complain about it. Canonical gives out Ubuntu for free, provides 2 GB of free cloud storage which, mind you, is not forced on the user, and provides the client side in full open source.

The 2GB is a loss leader, it’s only partly there to improve the Ubuntu desktop as a feature. It’s a win-win and besides you couldn’t sell anyone music if they needed a paid for account first.

The thing that showed the worst side of UbuntuOne service was the decision to hide the purchased music folder instead of using a standard sub folder for it. There were technical reasons and design reasons but I and others are still uncomfortable with the lack of access users have to their music files and the ability to move them out of the UbuntuOne cloud.

Actually, I think this would be a great idea for any cloud-based open source service – let the parties that participated in developing the server code keep it for a competitive advantage, and provide the client in full open source. It’s the service that’s important anyway, and this way you have a chance to recoup on your development costs by providing the service first for some time (until the other interested parties get their server code and modified client code ready).

I’m generally not fond of cloud based services, I think they’re needlessly grabby with people’s data and access rights. The first priority must be for the user to weigh up the cost of doing it themselves with the cost of loosing control. The one good thing I can say about online services is that general users tend to recognise that trade-off better than with normal software.

It’s frustrating that users consider a lack of control over their own computer to be something to be agreeably ignorant about. Users weighing up and making an informed decision about how to solve a practical immediate problem with a solution that may have bad long term implications is better. At least with online services it’s more likely users will be burned earlier and in a recognisable way that makes them cautious.

What are you thoughts dear reader?

1 I’d be happier if Canonical just asked for money to work on user feature requests and bugs, but hey I’m just not in a majority on that thinking yet.
2 A known bone of contention; general thinking is that creating space for a marketplace is a good thing that attracts investment. But I feel sometimes that Canonical is more concerned with filling all holes with it’s own services than opening up the market and really benefiting Ubuntu.