< Back to Blogs
The OSP and You by Richard Wilder on July 25, 2008 06:15AM

I am the Associate General Counsel for Intellectual Property Policy at Microsoft, having joined the company about 9 months ago. My role is to work with a variety of constituencies inside the company and outside to help shape the approach we take to intellectual property. I am new to the company and cannot take credit for it, but am very pleased that in recent years, Microsoft has made progress in participating with open source communities. A part of that has been the implementation of the Open Specification Promise (OSP), which was launched in 2006. We think it is a simple and clear way to assure that the broadest audience of developers and customers working with commercial or open source software can implement specifications. We constantly listen to feedback from community representatives and respond to that feedback – through Q&A’s on the OSP page and directly to the community. Andy Oliver made some positive comments in this regard as recently as yesterday. When asked for clarification of the OSP with respect to the activities of Apache POI, we responded. The concerns were about implementations of specifications covered by the OSP that may be less than fully compliant – in particular due to implementation bugs. Such a situation is not explicitly covered by the OSP since it is meant to apply to a wide range of technologies and development models and it is simply not possible to address all specific situations in which it would apply. We addressed this situation in the following manner – and I apologize if the explanation is a bit technical, but I will try to avoid too much legal jargon.

The OSP says that it covers "any implementation to the extent it conforms to a Covered Specification" which addresses the heart of the conformance issue that was raised." To the extent it conforms" means that we do not require an implementation to be perfect; this can be because of implementation bugs or an intentional choice because the requirements of the particular implementation do not actually require full conformance. Under the OSP, implementations can be less than fully compliant. For example, a given implementation that takes a spreadsheet document, extracts information from it, and stores that information in a relational database might not comply with every required part of the spreadsheet document format but such an implementation would still be covered by the OSP.  By way of comparison, other promises in the industry may require complete conformance for the promise to apply, and those normally require full compliance as a condition (see IBM's Interoperability Specifications Pledge).  Some others make no statement about the subject at all, leaving it an open question as to whether full compliance is required.  At Microsoft, we felt that unless we indicated that the OSP is more flexible, some might conservatively assume that complete compliance is required, so we included the “extent it conforms” language in the OSP.  We chose to state explicitly that partially conformant implementations are covered, to the extent they are conformant in their individual implementation aspects.

As a result of this clarification, developers can have peace of mind that the specifications covered by the OSP, are, in fact, openly available without ambiguity. This is the kind of conversation and cooperation that marks our intentions with the open source community, and I look forward to continuing this dialogue into the future.

-Richard

Comments RSS
  1. I’m writing this from Portland, Oregon where one of the world’s largest Open Source conferences is taking

    posted at 12:21PM 07/25/2008
  2. Sam Ramji, Senior Director of Platform Strategy, just announced in his blog (from the OSCON conference

    posted at 12:31PM 07/25/2008
  3. Brendan said:

    You imply there has been a change to either the osp or its interpretation but don't identify what the change is.  If there has been a change please show what it is or link to the change.  Markup would be nice

    posted at 10:23PM 07/25/2008
  4. Sam Ramji said:

    There has been no change.  This note from Richard is an official clarification of the meaning and intent.  Consider it an official commentary from Microsoft on the OSP.

    posted at 12:20AM 07/26/2008
  5. We are hosting a confab in Seattle this week with fellow evangelists from around the world. We kicked

    posted at 05:37PM 07/26/2008
  6. André said:

    The OSP raises some doubts if it is legally enforcable on an international scale. Amy Marasco and others failed to respond to the concerns. Probably they believe the world is just the US. The Q&A you linked is just more of the same old marketing gibber as the real questions are not asked.

    Generally speaking legal documents are not related to "how you feel" about things. The question is if the OSP fulfills its stated objectives and works. Instead of thinking in a developer way and receiving community input to help to improve it the text is just being defended. This undermined trust in the OSP and affirmed the suspicion of a hidden agenda.

    I understand that patent indemnification is a legal subject in breeding stage. No one really knows how to do it right. Private law is internationally very much different and who says he understands all strings attached is a lier or a bad lawyer. But when Marasco on the one hand lobbies for hardcore RAND provisions at ISO and ITU, and Microsoft throws everything it can to undermine document interoperability initiatives at the governmental level (just think of the moderate EIFv1 and how the EU has been bullied) the OSP does not come out of very credible hands. Instead you feel there is a party that wants to screw you up. Yes, here the feeling comes into play, on the trust level.

    It is a communication problem of your lawyers, for sure.

    posted at 06:47AM 08/21/2008
  7. if you really want to try something radical, how about reconsidering the whole concept of software patents?  Fact is they really don't make sense.

    Microsoft has personally been at the receiving end of some hugely costly patent infringement lawsuits (activeX etc.).  So I really can't blame you for being very aggressive in seeking patents of your own.  But come on, 3000 patent applications per year!!  absurd!

    The world would be a much better place if everyone would agree to do away with software patents altogether and switch to copyright protection which is far more appropriate for written material such as computer programs.

    Some Reasons why software patents aren't valid:

    Did you know that software was not patentable for many years, and that all of the major developments that we depend upon, the entire internet infrastructure was created without patents?  If everything had been proprietary there would be no intercommunication instead we would be stuck in endless fragmentary world of ipx vs arcnet vs token ring vs bitnet,  ad nauseam with nobody able to talk to anybody else.  But because people were for a brief time in history, willing and able to focus on the greater good of humanity, they came together in unselfish ways and created the universal standards that we all now benefit from.

    Did you know that there are now so many different patents that it is impossible to write any program no matter how trivial without being in violation of something somewhere?  Programmers are walking blindfolded through a minefield.  And sooner or later it will blow up in your face, it is just a question of when.

    Have you considered the fundamental difference between a mechanical device which continues essentially unchanged for decades, versus software with a typical lifespan of about 5 years which is protected by a patent with a span of 15 years....  Does that make any sense?  Protecting a 5 to 8 year product with a 15 year contract?  The patent system was created in an era of slow changes and long product lifespans.  It is not a good fit to apply it to something like software which is a very different world from the one in which patents were conceived.

    Have you considered the first principle of a patent, the purpose of which was to get companies to disclose their products which would otherwise remain a trade secret...  the deal is that by providing *complete* information on how to build a particular device the company would be granted a time limited monopoly on selling that device, after which it would revert to the public domain for the benefit of all.  Based on this principle, it should be required that the complete source code of your program be published in order that it becomes possible for others to study it in sufficient detail that they can reproduce the functionality.  So where is all of that source code?  Without the source code the patent system is essentially granting protection to a trade secret....  this is entirely contrary to the original spirit and intent of the patent system.

    Have you considered the wholesale abandonment of the principle requirement that to be patentable something has to be non-obvious to a practitioner sufficiently skilled in the field.  Why has this principle been abandoned, why do we have such totally absurd patents on things like a mouse click?

    Have you considered that the patent office can not possibly be expected to cope with the onslaught of tens of thousands of software patents per year.  Nobody could possibly track all of that.  Thus they tend to rubber stamp everything and let the courts duke it out at huge cost and risk and with arbitrary results.  

    Do you really want to spend the rest of you life looking over your shoulder waiting for the next $500 million blindside by someone who never did more that put a vague overly broad concept on a piece of paper and send it off.  Then they wait until their unsuspecting prey does all the research and development and marketing to actually create the product, and they suddenly swoop in for the kill like they did to Blackberry,  how can you possibly run a business with such a high risk and uncertainty?  And how completely unjust that some joker who never invested an once of energy should be allowed to extort money from you in that scenario.

    For all of those reasons and more, we need to seriously consider dumping this whole insane concept of software patents, it is a detriment to the entire human race.  

    The Patent system was intended to benefit humanity, and to encourage investment into product research, but when applied to software in it's present form it is having exactly the opposite effect.  Since it is becoming nearly impossible to create a software product that is not going to be in violation of somebodies vague and overly broad but patented idea somewhere, then there is not much point in going to all the trouble of creating a new software product is there?  The patent system has become a major disincentive to creating new software products.

    Seems like the only thing we can do right now is to wait 15 years until all these silly patents expire, and then we can get back to work.  Or we can rethink this insanity.

    posted at 06:03AM 09/18/2008
  8. !Appenoodorads! said:

    Where is your   ?

    Elect  ?

    Offers brand name  ?

    Don't beat me!

    posted at 06:49AM 11/03/2008
  9. Stanislav Danielov Georgiev said:

    Iskam da naprava karta

    posted at 03:47AM 11/12/2008
  10. ot nego se prizovava drug ork

    posted at 03:50AM 11/12/2008
Post a Comment
*
*