Maybe I'm missing something, but I don't see the point of proprietary network services (or cloud computing, or Software as a Service, if you prefer). Not when you have Free software as an alternative (Free, in this case, being analogous to open source or GNU/Linux).
In fact, proprietary network services strike me chiefly as a way to offer the incidental features of Free software without the provider giving up control. But in the last year, I'm glad to say, this dodge has started to become less tenable, as Free software has started to focus on network services.
Oh, I understand why developers might have enjoyed the idea, back a couple of years ago when network services were new. I may not be a developer myself, but I understand how the challenge of the delivery model might add interest to your work, at least before it became commonplace.
Nor would I suggest that network services should never have been developed. Diversity never hurts, and if a technique is plausible, someone is going to develop it. I accept that.
What I have trouble understanding is why proprietary network services are attractive to clients when Free software offers the same advantages and more. You want software that doesn't cost you anything? Free software offers that. You want to have the same software available, regardless of the computer you're using, and not have to worry about whether your licenses allow you to install on additional computers? Free software can do that. Reduced support costs? Hook up to the free software community, and you'll probably find a mailing list that does that.
How about centralized bug-fixes and updates? Through the repositories and package management systems of Free operating systems, Free software offers those, too.
If anything, locally-installed Free software has substantial advantages over network services. Some network services, such as Zoho and Google Apps, either have or are developing offline modes you can use when your Internet connection is down, but many still do not, making them less useful than any local application.
The same is true with privacy concerns. Although encryption is starting to be offered by some network services, not only can you do far more to secure your data on a local network or workstation, but, with Free software, you can scrutinize the code and satisfy yourself that no back doors exist for intruders. You don't have to trust the provider, because you can take steps for yourself.
Just as important, many network services have fewer features than their local counterparts, particularly those for office productivity. No doubt part of the reason is that local applications are more mature, but another seems to be that Web apps jettison features in the interests of faster transmission. As a result, network services can be especially frustrating if you're a power user, since many of the features you rely on for speed and efficiency simply aren't available in them. Frankly, given a choice between ajaxWrite and OpenOffice.org or ajaxSketch and the GIMP, who in their right minds would choose the Web app?
Surely nobody with serious work to do. Such choices would be like insisting on working in a text editor or a paint program when more mature applications are available.
The only reason I can see for clients preferring proprietary Web apps (aside from the fact that they're trendy) is that software as a service is less of a stretch for the average managerial mind than Free software. Even today, many find the idea of Free software a challenge to standard business practices, because it requires rethinking software procurement, supplier relationships, and, at times, existing business models. By choosing network services, a convention-bound company can often get the use of cost-free software (just as they could with Free software), but without having to worry about any of the mind-stretching aspects that go along with it.
Besides, outsourcing services is something that modern businesses do all the time.
But the ones who really benefit from network services are the suppliers. Unlike traditional software providers, their support costs are lower because most of the maintenance is centralized. Even more significantly, they can protect their so-called intellectual property without adopting a Free license. Furthermore, they can do so while offering -- at least to casual or light users -- what many outsiders consider the dominant feature of Free software: Availability at no charge.
A cynic might call network services as we know them a classic response to a changing market. Concerned about possible competition from Free software, network service companies have changed standard practices in the software market just enough to mount a credible challenge -- and, until recently, the ploy has worked.
Free software enters network services
Tim O'Reilly warned two years ago that network services were challenging Free software licensing. His comments were widely misunderstood in the media, but in the last year or so, enough people have come to similar conclusions that the Free software community is starting to expand into network services, settling any uncertainty by declaring most of them proprietary, and discussing what network services delivered under a Free license might look like.
One of the first signs of Free software's growing awareness of network services was the release in November 2007 of the third version of the Affero General Public License (AGPL). Named for an earlier, related license, the AGPL is a modification of the GNU GPL that, in essence, declares that providers of network services are distributors of software, and subject to the same requirements as any other distributor who users the GPL. According to Palamida's figures, the license has been used by 100 software projects as of May, 2008. Theresa Bui, vice president of marketing for Palamida, expects at least another 50 to start using it by August -- and probably more.
Still others are starting to flesh out what network services under a Free license might look like, such as Fabrizio Capobianco, the CEO of Funambol and one of the first adopters of the AGPL; Marco Barulli of the Clipperz project, who has repeatedly blogged on the topic; and the newly announced autonom.us, which might be called a think-tank for free software legal experts and activists.
The Free software position on network services is still evolving, but the emerging consensus seems to be that, unlike with locally installed software, the ability to copy, modify, and distribute software is not enough to guarantee users' freedoms. Instead, steps are necessary to ensure that users of network services can control their data and privacy. For example, Clipperz has what it calls a zero-knowledge policy, which uses encryption to hide both its users' identities and data.
Personally, I'm still not convinced that network services offer any advantages for users over local software. However, that will probably change as network services become more sophisticated -- possibly through Google Gears -- and, at any rate, my opinion seems to be a minority one.
For better or worse, network services are here to stay. And, that being so, I am glad to see the Free software community treating them as the proprietary efforts they are and offering an alternative. Otherwise, O'Reilly's prediction that Free software might become irrelevant in the face of network services seems all too likely.