Sun’s Jonathan Schwartz has posted an interesting blog entry that makes the case that open is more an issue of portability than one of available source code (see http://blogs.sun.com/roller/page/jonathan/20040808#rewriting_history_and_vocabulary). I’ve read this one through several times, and it feels as though Jonathan is trying to redefine the assumed meaning of “open” to make his case, deliberately trying to blur the distinction between “standard” and “open”.
.NET is neither a standard nor is it open. Apache is open, the standard part is harder to quantify there (but the sheer size of its installed base may indeed make it so). IE is not open, but for better or worse it is a standard (and it becoming less so right now). PHP is neither. Windows is not open, but it is definitely a standard (or lots of them, ugh!). That is my viewpoint, and you may agree or disagree with my list. And that is just the point, the terms “open” and “standard” are not one and the same, and users expect them to mean something (that is both vague, and ever changing). And, using the common understandings of these terms, J2EE is a standard, but it is not at all open.
After all, if Java were truly open, would Sun have the right to sue Microsoft over changes Microsoft made to their own Java implementation? If Java were truly open, would organizations have to pay Sun fees for the right to claim Java compatibility? And if Java were truly open, would Sun have ownership over Java logos and their use?
Sorry Jonathan, Java is a standard, I’ll give you that. But you can’t answer the “open” critics by redefining the term.

8 thoughts

  1. <em>"But you can’t answer the ‘open’ critics by redefining the term."</em>
    … but that’s how most of those arguments over labels are supposed to go, isn’t it…? 😉
    For those "isNot/isToo a standard" debates, one angle that has seemed useful and accepted is to distinguish between de jure standards (where some group with some authority describes desired behavior) and de facto standards (things which just plain work in the world today). People can then argue whether de jure beats de facto, but it does seem to move the conversation forward, at least a little bit….;-)
    For debates over the meaning of the word "open", one thing that has helped me is to distinguish where the code runs… in Apache you own your server, so you can directly affect its behavior, but something like the Macromedia Flash Player is installed on millions of machines you can’t personally configure, and that’s where predictability and reliabiliity beat the ability to customize. "Can I change the code?" takes on different values in different situations, true…?
    Jonathan started "Only a customer can define the word ‘open.’ That’s my view." He then goes on to define it (I think… lots of "let’s look at this, and here’s what someone said, and now let’s change the subject.") I guess we’re all customers of something, huh…? 😉
    cu,
    jd

  2. I guess similar claim can be made about many other platforms and programming languages. In my opinion "open" should be defined, when in software context, as software to which source code is provided free of charge and users are able to modify the software to their needs. Summarizing, anything that is licensed under one of (too many to count now) open software licenses is "open".

  3. Schartz might just be right on target with his entry… a mistake most people make is take "open source" and "free" (or their respective movements) to be one and the same… which is what leads to the "we don’t want to be lumped in with them" saying you see everywhere this topic comes up… this page [http://www.gnu.org/philosophy/free-software-for-freedom.html] on the gnu website points to the fact that all open source really says is “You can look at the source code.”

  4. Hello,
    Very interesting Site but it´s too late to start writing now (3:00 PM). I´ll come later.
    Geri
    htto://www.caino-test.com

  5. ".NET is neither standard nor is it open."
    Both the .NET CLR and the C# programming language have been submitted to the ECMA for standarization, and the Mono project (supported bo Novell) is doing an open source implementation of both for Linux.
    "After all, if Java were truly open, would Sun have the right to sue Microsoft over changes made to their own Java implementation?"
    Sun’s lawsuit was over Microsoft’s use of the word "Java" to desribe their implementation, not over the implementation itself.

  6. While we are on the topic of the abuse of the word "open" I’d like to see Macromedia retract what it calls is the "open" flash file format specification. There are many missing parts in the document (RTMP is completely top-secret) and the developer, according to the license, is not "open" to do with it as he or she pleases (only permitted to make software that outputs SWF).
    Don’t get me wrong, I love Flash, but it would be great to see Macormedia open up the license a little more so 3rd party players and authorware could be created for alternative platforms.

  7. Hi, "nobody nowhere"…. 😉
    Those words like "open" are always, uh, open to debate, so I try to avoid the labels myself unless I first define how they’re being used. Different people will just go for the gusto, however… fact-of-life, from what I see, welcome to my planet. 8)
    The free SWF file format documentation is more for creating valid SWF than for rendering valid SWF… the emphasis here is on the predictable Macromedia Flash Player doing the work, so the docs do define what works in that engine. (The Macromedia Flash Player source code is available to strategic partners, but these are usually embedded device manufacturers and other non-computer use, rather than for alternate players which could fragment the authoring base.)
    For the video part, there’s some non-Macromedia licensing in there… Sorenson did a lot of the work on the video component, and from what I understand Macromedia doesn’t have the right to describe how they make their files, sorry. (RTMP is a superset of the video spec here.)

Leave a Reply