Dialogue on definition of 'open' continues

Dialogue on definition of 'open' continues

Summary: IBM's senior program manager of Web services Standards Tom Glover (also president, and Chairman of the Board for the Web Services-Interoperability (WS-I) Organization) has responded to my feedback on defining open.  Why define open?

TOPICS: Open Source

IBM's senior program manager of Web services Standards Tom Glover (also president, and Chairman of the Board for the Web Services-Interoperability (WS-I) Organization) has responded to my feedback on defining open.  Why define open?  Well, defining it is worthless unless people understand all of the potential benefits (eg: self-control over your own technology destiny vs. passing that control to vendors).   Then, once the those benefits are well understood, having a definition makes it easier for technology buyers to recognize the degree to which something that claims to be open (or support open) really is open and which of the potential benefits accrue to them should they elect to use that "something" (a product, service, vendor, etc.).   I discuss the degrees of openness  -- a concept first explored graphically by IBM's Bob Sutor -- in my follow up to the flap between Apache vs. OASIS.  As a side note, the Apache/OASIS WS-Security conflict lives at ground zero of the reconciliation of the current incompatibilities between the open source and open standards movements.  Something has to give way and Sun could be nudging hard enough on the open source side of the conflict that we could soon see the equivalent of the Berlin Wall being torn down (stay tuned for my blog on this). 

In his blog, Glover gives the discourse some structure, distilling my discussion of "open" down to four bullet points:

  • Is open absolute or relative? I say it's relative but I had a hard time pinning down Glover's position based on what he wrote.  My interpretation of what he wrote is that once you cross over into the open range (with degrees of openness providing the measurement of granularity within), then open is an absolute state.  You're either in the open range, or your not. I called Glover to get a better fix and he said that I essentially got it right adding that in the world that he and Bob Sutor are thinking of, "in theory you can select a set of criteria to be used to score the openness of something, you can score things against those criteria, and you can decide on a score that's good enough to call something open."  However, one problem with that system is the question of who decides what that borderline score is?  Chances are that there would be some disagreement over this (there already is and we don't even have a systemin place).  So, given that,  I think we're better off sticking to relative because today, not everyone agrees on the point at which "closed" ends, and "open" begins. 
  • It's time to decouple the words "open" and "standard."  Glover agrees that other things are open besides standards and that there are standards that aren't necessarily open.  And here's something Glover also told me he agrees with: to the extent that it makes sense to come up with a range of granularity for open, it might be a good idea to do the same for "standard" (de facto vs. de jure) and then, perhaps, overlay them (sounds like a Gartner Magic Quadrant, only way way way better).
  • Compliance: to measure, or not to measure -- history has proven that, left unchecked (and I won't point any fingers here), vendors will claim compliance in situations where others might disagree.  So, the question here is, in addition to helping buyers recognize the degrees of open and the degrees of standard, once they pick their dots (specifications that, to one degree or another are open, standard, or both) on the aforementioned chart,  can you help them to recognize the degree to which certain solutions support that specification.  This is actually a third diminension to the chart.   Sun for example, keeps a very tight grip on compliance with the various Java specifications.  Technology buyers appear to appreciate that because of how much easier  it is for them to recognize and then acquire certifiably compliant solutions.  Very few specifications however have the equivalent of Sun's tightly supervised compliance regimen.  Not even the World Wide Web Consortium has something like this (although I have suggested it to them).
  • Adoption of open "stuff" gives buyes more control-- Glover asserts that buyers of technology -- ones whose infrastructures fall comfortably in to the open range -- are not the only ones who benefit from their "open discipline."  Glover says that in a win-win type of way, sellers of compliant solutions benefit too.  Maybe those benefits come full circle back to the buyers who end up with more choices.  But, at the end of the day, you wouldn't need many fingers to count how many solution providers don't have some sort of proprietary extensions to their supposedly compliant solutions that, when invoked, can make it difficult for customers to switch, should you ever want to.  So, as an advocate for buyers (ZDNet's readers), my focus on this point will always be about how buyers can get the most out of their IT and that doing so starts with making sure they don't relinquish control of any part of it to vendors because of reliance, dependence, or addiction (a range if you ask me) to proprietary technologies or formats.  But I completely understand the point Glover makes which is that a healthy ecosystem can emerge -- one in which everyone wins.
[Update 7/19/2005: See the response to this blog from IBM's Tom Glover]

Topic: Open Source

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Dialog on definition of "porn" continues.

    Good luck with either one...
    • I know it when I see it

      That's what it will all come down to.

      Carl Rapson
      • Yup, "open" is in the eye of the beholder.

        Be honest, it's a silly argument that can't be won either way.
  • Open or Owned?

    Ground zero is right. Here we go again. Three questions concern me: 1) Is Open absolute or relative? 2) Where's the threshold separating Open from Owned and Controlled? 3) Is there a solution?

    So let's get to it.

    Is Open absolute or relative?
    It's absolute in that there is a definitive baseline all Open Standards or Open Source licenses must meet to qualify as ?Open?. The threshold that separates Open from Closed (or Proprietary, Owned and Controlled) is definitive.

    However, above that baseline ?Open? is relative.

    Above the Open threshold, there is a range of ?Open? licenses. The open range can generally be categorized as running form BSD Open at one extreme, to GNU Open at the other. What separates the extremes is not a factor of openness, but rather how that openness is governed. The shorthand for open governess is that BSD Open is ?open for now?. And GNU Open is ?open forever?.

    Another way of referencing this governance that separates the extremes is to discuss an open license in terms of the four freedoms guaranteed in the GNU Open - GPL Bill of Rights. BSD Open promises one freedom ? open access to the source code. GNU Open delivers on all four. Something in between, like the Apache license, guarantees only two of the four freedoms (or rights).

    Some state this range of open governance as the difference between placing intellectual property into the public domain (BSD Open), and, nailing it there (GNU Open). The range of gradations that separate open licenses is really about the guarantee of rights rather than the mere statement of rights.

    Of course there is a great deal of argument within Open Source communities about the different licenses. If each license was to state transparently and unambiguously which of the four freedoms they guarantee, and which of the four they don't, we wouldn't be having most of these discussions. +1 unequivocal transparency.

    The Open Threshold:
    I disagree with you here David. We all know where the threshold point is - where ?closed? ends and ?open? begins. What's difficult about the point of open is that predators will always be coming up with deceitfully clever ways to fool the Open Internet marketplace into mistaking encumbered proposals for open.

    An encumbered proposal, whether it's for a protocol, a method, or collaborative source code, is one that is governed by a license where restrictions are based on some range of ownership, control, and permission assertions.

    An unencumbered license would be one that might have ownership claims (like copyright), but these claims have been restricted or ?lessened? exactly to enable open access, collaborative participation, and distribution.

    Let's push this difference further. The encumbered proposal restricts in some way open access, collaborative participation, and distribution. And it does this by asserting control or permissions based on ownership.

    When the unencumbered proposal asserts ownership claims, it's exactly for the purposes of restricting efforts to compromise open access, collaborative participation, and distribution.

    In both the encumbered and the unencumbered instances we must ask for what purpose are the assertions of ownership being made. Do these assertions seek to guarantee some or all of the four freedoms? Or do they seek to claim exclusive or permission based ownership of these rights?

    For much of the debate concerning the Open Internet, IPR (Intellectual Property Rights) issues of ownership involve claims over how digitally expressed ideas and digital implementations related to those ideas can be used. If ever there was an intersection of great import, it is this moment when we come to grips with encumbered claims of ownership, and the future of the Open Internet.

    If this discussion was about the Windows platform, and not the Open Internet, we would be yawping about whether or not control and permission restrictions through assertions of ownership make business sense for Microsoft. We would not be discussing Redmond's right to make such assertions, or make claims of ownership.

    This discussion however is all about core protocols and methods regarding the future of the Open Internet - the owned by none and used by all Open Internet. If you start with a premise other than the Open Internet, there's no anchor, no baseline measure of whether or not ownership claims should even be considered. If this about the Widows platform or some application to application agreement between Windows vendors, then it's all none of my business. It, on the other hand, it's about the Open Internet, then the efforts to claim ownership rights becomes of the utmost importance to all of us.

    We might not have reached a level of global awareness such that the oceans and atmosphere are considered part of the global public domain. But we have reached that point with the Open Internet.

    The Open Internet forces a classic confrontation concerning ?standards?. Because the Internet reached critical mass without becoming encumbered or having it's openness otherwise compromised, we have the advent of a new issue ? establishing the difference between Open Standards and Controlled or Owned Standards.

    The issue is further complicated in that the Open Internet brings with it the surging phenomenon of Open Source Communities. But here's the thing, neither the Open Internet nor the Open Source Communities sustained by the Open Internet can live with any consensus for connectivity and collaborative computational exchange other than that of Open Standards. Anything less is unacceptable.

    Open Standards exist. If they didn't we wouldn't have an Open Internet. Of course, Owned Standards exist too, but these efforts continue to lose the one battle that really counts. The battle for ?participation?. Which is exactly why predators want to disguise their restrictive ownership claims in the wrap of Open Standards. Yeah, the predators would love to decouple ?open? from ?standards? because then no one would be able to untangle the Gordian knot of what's open and what's not. We need an agreement process based on open transparency, not a standards structure favoring slight of hand deception.

    Transparency also demands compliance. But i don't see the general need to go the onerous ?testing? route if the licensing models governing use, collaboration, and distribution are unequivocally declarative. Like the GPL.

    If Open Standards are GNU Open guaranteed, as in ?open forever?, compliance is largely guaranteed by marketplace interaction. For instance, if vendor PT Barnumware claims to provide products that comply with a GPL'd OpenDocument XML specification, the governing license would guarantee a transparency certain to readily disclose the true measure of compliance. Within the four freedoms guaranteed by the GPL is the right to access and study the underlying source code. A right that enables computational consumers and interested Open Source Communities to measure compliance for themselves. Works for me.

    OASIS honchos Patrick Gannon and Andy Updegrove don't seem to understand that there can not be a compromise if the issue is that of the Open Internet and the Open Internet future based on Open XML technologies, methods, and implementations. Andy asks that Open Source stop making demands, and try instead to reach some measure of consensus and compromise with the proprietary vendors who drive traditional standards consortia. But how does Open Source reach consensus agreement with a bullet aimed at their heart? How does Open Source reach compromise when the finger on the trigger is not only someone who can't possibly be trusted, ever, but is also someone entirely intent on the kill? Get real. Open Source suicide is not up for discussion. Hence the hard line we draw between Open Standards and Owned Standards.

    The Open Internet is a world wide web of agreements concerning how digital information is formated, exchanged, and collaboratively worked. It's also a series of global agreements on how things connect and participate in this emerging digital civilization. Andy Updegrove is right, we do need collaborative mechanisms where global participants can connect to arbitrate, resolve, and perfect consensus on these ?agreements?. Andy is also right that the existing standards consortia are not up to the task of perfecting Open Standards. They are entirely lacking in the open participation and open access representation that makes Open Source collaboration so extraordinary. Nor can the standards consortia resolve the most fundamental conflicts between the command and control ownership interests of their proprietary vendor memberships, and, the interests of Open Source Communities and a global public who believe and demand that Internet remain ?Open?.

    And that's why were here at ground zero.

    Is there a solution?
    After the heated debate and skillfully orchestrated face off between influential and well informed decision makers, prompted to state their case in the ZD forums, surprisingly I think so. I think the answer lies with the W3C since they are the one standards group able to stand their ground and provide uncompromising, unencumbered Open Standards. They don't yet know how to guarantee and future proof their Open Standards, but their hesitation to address this need is certain to change.

    On the other hand other contenders like OASIS have compromised the public trust, shredding their principles to accommodate forum shopping predators. I don't see how they recover from that breach. Saddly the IETF dodged the heat surrounding the Apache ? Microsoft - MARiD tussle. Why would anyone put themselves in a dependence on the IETF given their linguine spine'd resolve?

    Sir Timothy has yet to figure out how to embrace the incredibly successful Open Source collaboration model. But when he does, the W3C is going to blow the lid off of participation in the Open Standards process. Then maybe we can kiss all this hand wringing good-bye.

    Leaving one last measure for Sir Tim to attend to. That is to GPL XML, and guarantee an open future for the semantic web.


    OpenOffice.org volunteer serving on the OASIS OpenDocument XML TC