Apache Software License, Hippo, and BlueNog
When I first got interested in open source software there was a lot of talk about the restrictions and liberties of various licenses and the risk that free-riders posed to the system. I have to admit that I never found these topics very interesting and usually referred the conversation to my colleague Stephen Walli (who is way more qualified in this area than I am — lawyers even listen to him!). For the most part, these (as well as the whole indemnification and SCO hysteria) have turned into non-issues, particularly for my clients who are users of the software and will probably never read a license anyway. Things tend to work themselves out.
But every once in a while, something interesting in the topic of licenses does pop up. You may remember I wrote a post describing how Bluenog took Hippo CMS, slapped their logo on it and sold it as commercial software. Well, they are still at it and they have even gone further as to remove any acknowledgement that they are repackaging someone else's software. The Apache Software License, which Hippo CMS uses, is very permissive and only requires that redistributions of the software contain a notice file giving credit to the original developers. Bluenog isn't even doing this. And, as you would probably expect they are not contributing back to Hippo either.
Bluenog is clearly in violation of Hippo's licensing terms so it may not matter what license Hippo is distributed under, but it did get me thinking about licenses again. The Apache Software License has been used very successfully for infrastructural components like the famous Apache HTTP Server and all those great Java frameworks and components. The key benefit there is achieving broad adoption. The terms are so generous that there is virtually no downside to including an Apache licensed component in your software. Adoption is a good thing for frameworks and components because lots of users help find bugs and help the project move forward. Even if a very small percentage of developers contribute back, the scale of the user base translates into a lot of support. This low barrier to adoption is particularly good for reference implementations of standards. Tomcat, Slide, and JackRabbit were all critical to the success of the standards they promoted.
As good as the ASL is for components and frameworks, I question its efficacy for business applications. Business applications, like Hippo, compete in a different market than infrastructure. They are going after a smaller (higher touch) install base and they are more actively competing against other products. Business applications need to innovate and differentiate from their competitors while infrastructure wants to be stable and standard. The potential for free-riders to undermine your investment to be unique is too great. This is why most other CMS on the market are licensed under the GPL or a similar license.
From a consumer perspective, it feels like Bluenog customers are getting ripped off. They are buying a software application that should be free. Customers are essentially paying Bluenog to ask questions on the Hippo mailing list that Hippo and the community are answering for free. It feels like Bluenog's refusal to acknowledge Hippo is an attempt to protect this arbitrage. Had customers worked directly with Hippo, they would not only save money, they would also know that Hippo has an entirely new product: Hippo CMS 7 that is a ground up rewrite from the 6.x series that Bluenog forked. I do think that this issue will eventually be worked out. Bluenog will probably not be able to continue practicing business in this manner: even if lawyers don't get involved. But, as you can probably tell, this drama does rankle my developer and open source sensibilities.