WebMink
Simon Phipps's personal commentary

Your commentary appears to contain a number of inaccurate characterizations of Steven's eWeek article. First, it is not simply a rehash of his earlier piece. The new article contains specific quotes from Mr. Schwartz that were not included in the first one, and a point-by-point response from a person, Eric Raymond, who was also not part of the original article.

Second, you misrepresent what ESR actually says when you say he actually agrees with Jonathan on all but the first point. ESR is clearly quoted in the article as disagreeing with Schwartz on the first two of four points, acknowledging that the next two are arguable, and strongly disagreeing with Schwartz's conclusion.

Third, you imply that ESR's "punchy" tone in the article is the result only of subjective editing, without offering any basis for such a conclusion. As a third-party observer with a clear bias for one partisan viewpoint, you really have no legitimacy in substituting your feelings in place of ESR's.

Fourth, you conclude your diatribe by making the same mistakes that Schwartz makes in his comments. You confuse the goals and philosophies of open source with those of corporations who want to gain some advantage through open source. Open source is about freedom, the freedom to create and disseminate ideas, and choice. It is developer-centric because it is the developers who are ultimately responsible for innovations and advancements, and freedom is a key enabler for progress. Choice means that anyone, even companies seeking to make a profit, are free to use the fruits of their labor as long as they respect the freedom of ideas principle. Questions of deployment or methods of doing business are not the direct responsibility of the open source community, but are properly addressed individually by companies according to their own circumstances. The FOSS community does not make it a priority to see open source software "mature to the mass-market", despite your mistaken impressions.
 

Your anonymous views are interesting, if rather formal and legalistic, because your attitude ('if you can't compile push off because you don't matter - only developers are entitled to freedom') displays the problem deliciously. Ignore the protection of end-user freedom and some corporation will come along and take it away, without breaking any laws or licenses.

And I was giving ESR the benefit of the doubt - I'm assuming his ad hominem towards Schwartz was not the main thrust of his comments.

You seem to have done an unusually close analysis of the two articles, I just read them. I know there's new stuff in the article, it's just not based on any new activity by Sun or Schwartz and I think it's deceitful to imply it is - it certainly led Pamela on Groklaw to think it was a new interview.

Oh, please leave a name & URL next time, I'd hate to see the anonymous trolling that happens over on Groklaw happen here.
 

I'm afraid you've now misrepresented my views with your paraphrased quote "if you can't compile push off because you don't matter - only developers are entitled to freedom". Open source promotes end-user freedom by giving them nearly unlimited choices. If users cannot compile code for themselves, they are still free to find binaries already built by someone else (either for free or for a fee), hire someone to do the job for them, or contract with a vendor as part of a larger service agreement. This arrangement serves both developers and end-users well, leaving everyone free to pursue their own goals without limiting those of anyone else. No corporation can take away end-user freedom in the open source context, because end users will always have the final say about who they will take their business to. It should also be pointed out that no corporation, proprietary or open source, can ever be counted on to guarantee end-user freedom in the long run. Only the combined community of open source developers, vendors, and end users can do that by vigorously defending the core open source principles, not those thrust upon them by people with other agendas.

OT: I think anonymous posting is useful when opponents find it necessary to attack the messenger when they cannot effectively refute the message. It's all too easy to dismiss statements by ESR as those of a free software fanatic, or those of PJ on Groklaw as a Sun hater. On that count, anonymity leaves a blank slate that forces the debate back to the contents of the message. There are good anonymous posts as well as bad ones on any forum that allows them. On Groklaw at least, I believe anonymous posts have added more than they have detracted from the quality of the site. Keep in mind that your own post to Groklaw pointing to this blog was done anonymously. So for now, I'll proudly continue to be anonymous.
 

I'm saddened that you're not willing to heed my request for an introduction. You're keen on diversionary precision, I can see. My Groklaw posting was from "Anonymous" because I do not have an account there and don't intend to create one right now. It was, however, customarily signed and had a link to my personal web site. It was thus hardly anonymous - if it was you wouldn't be making these comments here. So give me a break! You can keep on trying to 'educate' me as a poor dumb corporate employee of an enemy power if you want, but that's not very interesting and I'll not play for much longer.

I characterised your view that way because that's what I'm trying to draw your attention to and discuss. I realise you disagree with my focus but then I disagree with your view that it's impossible to shape or control software that's OSD licensed. I totally agree with the comment that "It should also be pointed out that no corporation, proprietary or open source, can ever be counted on to guarantee end-user freedom in the long run" so I place no absolute faith in any of those - a broad category.

However, I think that the freedom of both developers and users will be maximised by the combination of participative standards implemented by open source communities in ways that are demonstrably interoperable/portable/compatible. I personally think this combination needs considering and exploring by the open source community I'm part of (or are Sun employees, responsible for huge swathes of the open source software people value, automatically barred?)

You talk loftily of "defending the core open source principles, not those thrust upon them by people with other agendas" but as James Gosling commented recently in connection with Java, believing in the absence of agendas is naive. The core principle of open source is freedom, not any of the steps defined in 1998 as necessary at that time to achieve the principle. Sooner or later we're going to have to change to promote freedom, because someone will come along and use the licenses and legalism as a weapon to ensnare - that's the way of the borg. I'm not pushing Sun's or anyone's agenda when I try to draw attention to this - I think it needs urgent discussion.

The whole reason Sun is still reluctant to create its own open source Java implementation is because it can't see how to maintain Java's compatibility-derived freedom if standards and compatibility aren't valued. If our community could have a reasoned and non-partisan discussion about that topic we would get much further than we ever will by throwing rocks at community members for heresy.
 

OK, let's get past the generalities and talk about some specific examples to test your viewpoint. When talking about binary compatibility with Linux, you really have to specify which of several levels of compatibility you mean. Linux as an operating systems is a platform for hosting applications, developing applications, and administering resources (hardware, CPU time, network connections, data files, etc.), and there are compatibility issues for each of these activities. As far as hosting and developing applications goes, there are existing standards in place (Posix, LSB, etc.) which allow applications to be easily ported from one version of Linux to another with very little effort. This high level of compatibility is already present today, proven by the wide availability of business-class applications such as Apache and OpenOffice on virtually all distributions of Linux. The transparency of operating platforms available for such applications also guarantees that end-users are prevented from having their data assets locked into a platform from a particular vendor, ala Access and SQL Server from Microsoft.

It is in the third use category that the argument for binary compatibility in Linux is far less persuasive. An operating system kernel interacts closely with the hardware it controls, and performance is derived from a complex interaction of hardware, software, and external events. Linus himself has rejected mandating a formal ABI (application binary interface) for device driver modules which are written to plug into the kernel, because such formal interfaces inherently limit what can be done in future versions of the kernel due to issues of backwards compatibility. This is unacceptable in a product which must constantly adapt to support new classes of hardware while simultaneously meeting increased demands for performance and stability. Compatibility is best assured in this case not by binary interface standards, but by open source approaches which allow for rapid adoption and propagation of new developments throughout the developer community. Lack of an ABI also discourages the proliferation of binary-only device drivers from vendors, which pose the threat of devaluing the openness of Linux development.

In contrast to the case of Linux, we can look at the case of Sun's Java, which you brought up. Java is a combination of a programming language with a run-time environment which allows for broad cross-platform operations. Sun has published specifications for Java which in theory allow for multiple implementations, but to date the only viable products available are all derived from Sun's reference implementation. In this sense, the "open standards" approach has failed to free Java from being a one-vendor, one size fits all offering.

Sun has resisted calls for making Java open source, opening up its development to outside organizations who would have more incentive to contribute to a project which did not require them to relinquish all rights to their contributions. Unfortunately for Sun, they are directly competing with another proprietary product (.NET from Microsoft) which also purports to be a universal framework for future application development and hosting (as long as all platforms run some form of Windows). In a battle between two strictly proprietary frameworks, Sun will likely lose out to their competition because of the large advantage in market penetration that Microsoft enjoys with their Windows platform. A combination of open standards for Java along with an open source Java development model might be exactly what Java needs to more successfully compete with its primary competition, much as Linux has been able to achieve with a similar approach in operating systems.

To summarize this rather long post, open standards and open source address different aspects of the software development process. Open standards provide some assurances for software compatibility, but often trade off the possibility of incorporating many rapid innovations. They enable development of multiple implementations and foster competition in software products only when associated intellectual property claims do not retard such efforts. Open source naturally encourages growth through community involvement because people are encouraged to build on the freely donated work of others instead of having to reimplement everything from the ground up. It is important to note that open standards and open source are neither mutually exclusive nor mutually dependent. Each approach has its legitimate place in the world.
 

OOT: The only reason I even brought up the issue of anonymous posting, and clearly labeled it as off-topic, was the weight you seemed to place on its importance in your reply. You happened to mention my anonymity on two separate occasions, so I felt some obligation to reply. I'll hold to my original statement that I'd rather attention be placed on message content rather than on any preconceptions about the messenger.
 

So you want to keep me at a disadvantage. I don't like that, having disclosed my identity. I try hard to stay open and honest and am a firm believer in community. That's why I'm asking. You're making the big issue by wanting to keep the upper hand...
 

I think I'll have to write this as a new posting - the sucky comment system here is too restrictive on what HTML it allows and I've written too much for it to be readable without formatting. Let's continue discussion in the comments for the new posting.