In yesterday's Q&A about the Honest Public License, Fabrizio Capobianco was of the opinion that in the future, most software (open source included) will be run as services.  His word was "unstoppable," which I admit raised eyebrows on my end.  But it rests on some pretty solid observations about how open source works.

You can open most any page of InformationWeek and see the term "SaaS" splashed somewhere; it's not being used frivolously.  And the nature of most any open source project makes it a candidate to be transformed into a service, either by its original authors or by someone else down the line.  So "unstoppable" isn't a bad adjective: in theory, people will do it, because open source exists to be transformed.

"Unstoppable," OK -- but what about "inevitable"?  Is every open source project out there a candidate to be transformed into a service?  Probably not, because there are many projects that simply don't lend themselves to being rendered as a service -- a Web browser, for instance (like the open-source one I'm typing this blog post on), especially since Web browsers are how most people get access to services in the first place.  As long as bandwidth and latency remain finite -- and any World of Warcraft player will be able to tell you that they're both quite limited -- that will continue to be the case.

A key distinction that needs to be drawn here is between transforming something that exists and simply building with it.  One possible gross misinterpretation of what people want with the HPL would be that anything I build on a Web site with, say, PHP, would have to be open source.  Wrong, of course, but we've seen how fast fear and misinformation about open source can spread.  The point isn't to put all code in the public eye, but to make sure that people who are transforming existing projects and perhaps using that as a commercial venture aren't freeloading.

In my original post about the open source ASP loophole, I felt the best response to someone creating a closed Web service from something open was to create something similar, if not better, and open it up.  I still think that's a better alternative than changing the licensing for one's existing product (or future editions of same) in what could be described as a retaliatory fashion.  But now that I think about it, if you're just now creating something new that lends itself to being reworked as a service, there's nothing that says you're in the wrong for adopting, pre-emptively, a protective measure against something that may well indeed be "unstoppable."