There's an interesting, brief article in the Guardian Technology section today: Cloud computing is a trap, warns GNU founder Richard Stallman. In it, Richard Stallman is quoted as saying about cloud computing:
"It's stupidity. It's worse than stupidity: it's a marketing hype campaign."Later in the article he elucidates further:
"One reason you should not use web applications to do your computing is that you lose control," he said. "It's just as bad as using a proprietary program. Do your own computing on your own computer with your copy of a freedom-respecting program. If you use a proprietary program or somebody else's web server, you're defenceless. You're putty in the hands of whoever developed that software."I don't think it would surprise anyone that I respectfully disagree with this statement. I'm very supportive of his concern about cloud computing, and I agree that it's something that the Free Software and Free Culture communities need to address. But in rejecting all network computing, I think RMS has thrown out the baby with the bathwater. I don't believe loss of absolute control means that you lose your autonomy completely. And I think that exchanging some control in order to participate in social, collaborative computing is ultimately enriching for individuals and for society. Here's an admittedly overstretched metaphor: I live in a house where I control everything* -- the temperature, where the furniture is placed, how much and what kind of food is in the cupboards. I can go in any room in the house whenever I want, and I can change whatever I want. Great. I wouldn't want to spend any time in jail. In jail, I have very, very limited freedom, and there are hostile fellow inmates and in some jails interrogations and beatings. It is a really bad place to spend any amount of time. But I do like to go visit my friends' and family members' houses. I don't have absolute freedom to do whatever I want at their house, but I get to spend time with people I like, enjoy their hospitality, and also see the way other people live for a little while. By having an informal custom of hospitality interchange, I and my friends and social network get to enjoy more of the world than we would just in our own houses. If friends' houses were more like jail, I wouldn't want to go. If a friend told me that I couldn't talk about politics in her house (say), or another required everyone who visited to be strip-searched at the door, I'd of course not visit (and hopefully would be allowed to leave). But I usually can expect a certain level of autonomy in my person and in my effects that is acceptable and comfortable. Going places I don't individually control -- restaurants, museums, retail stores, public parks -- enriches my life immeasurably. A definition of "freedom" where I couldn't leave my own house because it was the only space I had absolute control over would not feel very free to me at all. At the same time, I think there are some places I just don't want to go -- my freedom and physical well-being wouldn't be protected or respected there. Similarly, I think that using network services makes my computing life fuller and more satisfying. I can do more things and be a more effective person by spring-boarding off the software on other peoples' computers than just with my own. I may not control your email server, but I enjoy sending you email, and I think it makes both of our lives better. And I think that just as we can define a level of personal autonomy that we expect in places that belong to other people or groups, we should be able to define a level of autonomy that we can expect when using software on other people's computers. Can we make working on network services more like visiting a friends' house than like being locked in a jail? We've made a balance between the absolute don't-use-other-people's-computers argument and the maybe-it's-OK-sometimes argument in the Franklin Street Statement. Time will tell whether we can craft a culture around Free Network Services that is respectful of users' autonomy, such that we can use other computers with some measure of confidence. * For hypothetical purposes. My wife and daughter would probably dispute this claim.
I've long ignored anything RMS has to say about how normal people should use their computers. He doesn't experience the web like most people, so I don't think he can comment in a relevant way.
I think what would be nice to see is a combination of self-hosted and proprietary web apps all talking to each other in standard formats while still maintaining privacy controls and allowing me to own my data.
My current worry with cloud computing is that I don't have private backups of my data that's in all these proprietary systems. I'm worried about a service shutting it's doors (Seems likely given the current state of the US economy) and me being locked out of relationships and data that used to provide value to me.
Anyone know if there's a self-hosted "cloud backup" application out there? Or will I actually turn an idea into code for once.
Do you take The Guardian article at its face value as a representation of what RMS meant? Below I've included a copy of an email that I sent to The Guardian earlier this evening and cc:d to RMS. I am immediately suspicious of any article that suggests that Stallman's comments "echo those made last week by Larry Ellison, the founder of Oracle." I don't think I've ever heard Ellison championing the cause of software freedom; on the contrary, a quote from this article http://www.haaretz.com/hasen/spages/893074.html clearly shows that he does not remotely sound like RMS: "Open source is not something to be feared. Open source is something to be explained. Open source wins not because it's open and not because it's free. Open source wins only when it's better," he says.
I would suggest digging deeper beyond The Guardian article to find out the truth. I am happy to forward you a copy of any reply from them that I might receive in response to this email that I sent:
To: tech@guardian.co.uk, userhelp@guardian.co.uk Subject: Request for clarification of Sept 29, 2008 Bobbie Johnson artcle Date: Mon, 29 Sep 2008 19:25:31 -0400
I'm writing because I have concerns about potential misrepresentations or distortions in an article published today (Sept 29, 2008) by Bobbie Johnson entitled: "Cloud computing is a trap, warns GNU founder Richard Stallman". I would like to request a transcript of the interview or correspondence between Stallman and The Guardian so that I can review the full context of the quotes attributed to RMS in the article.
Cloud computing is a nebulous term with little agreement concerning its precise definition. Both the content of the article and its title strike me as sensational and misleading; the wording makes it sound as if cloud computing as a whole is a trap, whereas I expect what RMS was saying is that cloud computing built upon a proprietary, non-free platform is where the trap lies. Most of the industry players such as Google, Microsoft, and Amazon are pushing forward their plans to deliver proprietary, non-free information and proprietary, non-free software over the net. The emphasis is on it being a proprietary platform, not clouds in general. There was a similar trap with Java as a proprietary platform, however Sun has begun disarming that particular trap: http://www.gnu.org/philosophy/java-trap.html
Not all cloud computing platforms are proprietary!
For example Nimbus, the Globus Alliance cloud computing project http://workspace.globus.org/, uses the Apache License, Version 2.0 which is a free software license that is compatible with version 3 of the GNU GPL. While it does provide a backend to Amazon's proprietary EC2 platform, the bulk of the project focuses on allowing a person to create a compute cloud on one's own hardware.
Another example of a seemingly misleading statement in the article is this:
<quote>The concept of using web-based programs like Google's Gmail is "worse than stupidity", according to a leading advocate of free software.</quote>The article then goes on to talk about cloud computing in general, which in my reading makes the previous quote seem to apply to web-based/cloud-based programs in general, instead of proprietary web-based programs like Google's Gmail. A good way to make a clear distinction would be to offer an example of a non-proprietary free software alternative to Google's Gmail, such as the web-based SquirrelMail program http://squirrelmail.org/wiki/SquirrelMailGPL that can be hosted on a server by one's self or by a trusted group known to the webmail user.
Given the reaction to Bobbie Johnson's article that I have seen on quite a few blogs there has been a large amount of confusion generated by this article. At best it is an unfortunate mistake that I hope will be corrected by The Guardian. I would like to take that more positive attitude rather than believing that either the author or The Guardian intended to distort Stallman's words for sensational effect or to spread FUD.
I have looked for Stallman's own statements on cloud computing, but searching the archives at stallman.org, fsf.org, and gnu.org I can find nothing written about cloud computing. If you could send me a copy of the correspondence or full transcript of your interview with RMS or, even better, publish it in full as another article, it may help to clear up the confusion.
I have to disagree with your disagreement, because 1. you (intentionally?) misunderstand RMS in claiming he does not want computers to communicate wich is easy to seperate from cloud computing 2. your comparison with places to go is a nice story but is not applicaple to computers. Cloud computing is about flexibly finding hardware to run your software on. There is no incentive (contradictory to your statement) of software to live in many different places to "enrich its life". If you restate your saying in terms of computer science may be I will get the point ... 3. Cloud computing as it is is a jail for your software (if you're the developer), because there is no interoperability between vendors. Changing the landlord of your cloud computer requires rewriting your software. Unless interoperable cloud computing is available no sane business model should rely on a single company offering it.
Also you should not mix cloud computing and web applications wich have not much in common (although the first is often used to deliver the second) in your argumentation. I myself agree with RMS that I do not want to deeply rely on software on someone elses computer. But knowing that most poeple do not care about their dependencies and privacy, I believe in APGL and Franklin Street Statement endorsing writing applications capable of open integration of external data sources: I can host the service on my own, everybody else uses big vendor xyz and nobody is locked out.
The place where I agree with RMS is we shouldn't assume that computing through a network service is the best option automatically. Pure outsourced network computing to a single vendor is going to have freedom and autonomy issues no matter how well the vendor meets the principles set forth in our Franklin Street Statement.
RMS is right that vendors want to deliver their view of "Cloud Computing" as the only and inevitable method for application delivery. Our challenge is to design, develop and deploy systems that operate "in the Cloud" but are decentralized enough that single point of failure doesn't really exist. This is hard technically, but facing that technical difficulty is essential as we pursue the greater goal of freedom and user autonomy.
Thus, I agree with evan's points that there important value in this emerging form of computing, but RMS has a point that we should not flock to it until we can address the freedom implications.
The term Cloud Computing gives many of us the impression that the data and software will exist virtually, but in reality it has a ground location housed in large server farms dotted around the globe. These farms are susceptible to Murphy's Law the same as any other piece of equipment and the policy whims of the company that owns them. As computer technology and applications take another leap forward, our understanding of how this will affect our daily lives has not been defined or fully understood. Security and anti-terrorism concerns will continue to evolve and may force these server farms to give up their secrets. The use of proprietary software further inhibits the portability, security and stops data interaction, building walls around network communities, stifling growth. It is the ability to use open source software to share data and information in new and unexpected ways that will be the enabler that will drive the evolution of cloud computing and social media interaction
@jahu: You and I mean different things by "cloud computing". I absolutely think of Web applications like Facebook, Google Docs, or box.net when I think of "the cloud". You seem to be thinking of a more focused definition: application hosting on proprietary platforms, like Google App Engine. Since Stallman references "web applications" in the quote above, I think he's speaking more to my definition than to yours. However, I agree that the Guardian article is too short to get a full understanding of RMS's usual well-rounded thinking on a complex matter.
@aculich: thanks for the link to Nimbus! That looks like a great project.
@evan: I had a private talk with RMS about the impact of moving data and computation out of your control(led hardware). Within that he distinguished well between Web Applications, Software as as Service and Cloud Computing. Because of that I assumed that the Guardian is citing RMS out of context.
Also, if you do not want "Cloud" to be the new "2.0", than you should use a solid definition of the term cloud computing - it already is used as a meaningless buzzword by salesman. The definition must not force some kind of low-level computational resource (like CPU cycles). But it has to insist on resources available to "unmanned" computers for universal use. This includes ongoing automatic use of a service without a person seeing it - cloud computing essentially is about abstracting certain resources! I have not seen ad driven web applications able to make profit without a human intentionally using them.
(btw. the examples you mentioned come close to my understanding of cloud computing, whereas gmail for example does not)