Slashdot Mirror
Microsoft Redefines "Open Standards"
Glyn Moody writes "Microsoft is at it again: trying to redefine what 'open' means. This time it wants open standards to be 'balanced' — for them to include patent-encumbered technologies under RAND (reasonable and non-discriminatory) terms. Which just happens to be incompatible with free software licensed under the GNU GPL."
From the gpl-isn't-really-open-either-ya-know dept.
What is up with that? The majority of people that go around saying this about the GPL complain that you can't include the GPL in proprietary software or other unfortunate obscure issues. The GPL is designed to keep software licensed under it Free (or open if you prefer). Sure sometimes that causes unfortunate problems with other Free Software licenses, but while there are those that would like to take away the freedom that users and developers get with the GPL, it's a cost I'm happy with.
No, but being patent encumbered does unless the patent holder declares the patent is free for anyone to implement under any terms they wish (ie they use the patent totally defensively and agree never to initiate any legal action against anyone over it).
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
Corporate entities are not at all alike Human entities, and therefore its very likely that the definitions of "reasonable" that are used by both are quite incompatible.
Giving MS track-record that chance is probably near a 100% ...
Open source only means that the source is available to the users of the product.
Nope, OSI defines open source software as software that:
A. Free Redistribution
B. Includes Source
C. Allow Derived Works
And a lot of other stuff. See http://www.opensource.org/docs/definition.php for more info.
Taxation is legalized theft, no more, no less.
Sometime back in the late '80s, Digital Review (or a similar industry newletter) ran an article in which Bill Gates was quoted as saying something to the effect that Microsoft's operating system was an "open system" because you could buy a computer from a large number of vendors that it would run on. (So long as you were talking about computers based on Intel chips, I suppose he could could sort of get away with saying that, as self-serving as it was.) Claiming that whatever that Microsoft does is in any way "open" is sort of old hat with those guys.
CUR ALLOC 20195.....5804M
It depends on what you want to be open.
If you want the source of all derived works to be available to all, and encourage more community development, then you want GPL.
If you want the source of the original work to be available to all, but allow the option of closed source for derived works (give more options to the authors of /direct/ derivatives, allow it to fit into more business and distribution models), MIT and BSD are "more open".
So, "it", is defined by what your primary goals are. I tend to prefer modified BSD/MIT style licenses myself, but the GPL certainly has a place for a lot of development models.
Self proclaimed typo king, and inventor of the bear destroying coffee table (patent not pending).
The OSI are not a recognised standard body or industry authority - they are little more than recognised banner wavers and supporters of open source but they carry essentially no weight. Their definition is all very well and good, but its not *the* definition.
I have always felt, and continue to do so, that "open source" merely indicates that the source code for a product is available.
I don't think you really mean that...
I'd kill to have the source to some of my vendor purchased apps so I could fix some long standing bugs and send the patches back to them.
... because patching third-party software yourself already implies
1 - the freedom to modify said source, even if it is only for private use
2 - the freedom to recompile the source, modified with your patches
3 - the freedom to publish your changes, even if you only publish them to the vendor
Creating/distributing a derived work, or redistributing the code is not a priority in that case. Access to the code is.
There is a very fine line between a fix and a derived work. And I suspect that freedom 2) above would classify as a "derived work" already.
Some GPL software is patent encumbered. IBM, for example, donated some of their patents for Open Source projects.
So it's patented, but probably unencumbered, then.
Hint: "encumbered" means restricted or blocked or limited. If the patent license is consistent with the FOSS license requirements (for example the GPL requires no restrictions on right to distribute modified versions, etc.), then the fact that some part of it is patented does not mean it's encumbered from the FOSS point of view.
Proprietary software is usually copyright-encumbered - your license may not allow copying it, and may not even give access to the source code. Many FOSS licenses also make restrictions - when you modify, you may not remove the names of previous contributors, for instance. Does this mean we should refer to BSD or GPL code as being "copyright-encumbered"?
Those who can make you believe absurdities can make you commit atrocities. - Voltaire
... and thus the circle continues. MS Troll claims some ignorance on the part of the poster (akin to claiming any and all discussions end when someone loses by invoking Godwin's law) to try to stifle competitive viewpoints by calling said poster a conspirator.
In fact, Microsoft Evangelism is documented in many locations by Microsoft employees themselves. If you don't know of this, then you are simply being ignorant (or are being "shilled" yourself.)
Thank you very much for proving my case better than I would have done it myself :
The end user becomes more free by having to pay someone to write a reimplementation of a CDDL algorithm to use with some GPL'd code?
At least, a sponsored GPL reimplementation of this code would become the common good of humanity. The mere fact that it would be needed just proves that CDDL is not concerned by the end user rights.
The user becomes more free by not being able to give their friend a copy of the binary without remembering to include a written offer for the source code, even though their friend (if they actually wanted the source) could still get it from the upstream source?
Providing a friend with a binary only module is a bad gift indeed. What if he further needs to port it ? What if he changes from processor ? Should he be deprived of your gift ? Your friend in that case is the end user, and you should treat him as well as you've been treated yourself before, because he's the one the GPL intends to protect now.
[...burps...] GPLv2 [more burps] GPLv2 [even more burps] The GPLv3 [and on and on]
I'm sure you know the difference between specifications and implementations, and pointing out defects of a specific version of a products merely show bugs that are therefore corrected upon identification. It does nothing to prove the underlying scheme right or wrong.
And, yes, IAL, and I read the GPL from top to bottom, every version of it.
I'm not sure what IAL means, I Am Legend maybe? Or did you mean to say that you are a lawyer, in which case I am not surprised by your skill at doublethink, just disappointed by it. Interesting that you don't specify which of the (mutually-incompatible) versions of the GPL you prefer.
Yes, I'm a lawyer. My personal choice is for the current version with provision you can relicense under any later revision (GPL v3+). Your sacarstic style (for a missing 'A', which is just a typo) just hints that you need to revert to ad hominem arguments when you clearly lack basis for your claim.
Correct. 1KB has been 1024 bytes since at least the '60s. Suddenly deciding that it's 1000 does nothing other than create confusion. The K prefix on SI units means 1000, but byte is not an SI unit. Redefining the prefixes to be powers of ten is not even self-consistent, unless you redefine a byte to be ten bits (technically, it can be; an octet is 8 bits, a byte is an architecture-specific size, but most machines built in the last 30 years have had 8-bit bytes, so byte and octet are used interchangeably). By using power of ten units for prefixes and power of two units for bytes and nibbles, you make things needlessly complicated, and you break 50 years of existing software that uses the old meanings.
I am TheRaven on Soylent News
You don't have to give away the source code with every copy to comply with the GPL. You have to make the source available. This could be as simple as putting up a tarball of the source code on your web site and including a note with the binary notifying users that they can download the source code from your site if they wish to do so.
We all know what to do, but we don't know how to get re-elected once we have done it
So am I, which is why I can tell you that you're full of it. Go search the USENET archives, for example: you won't even find any significant mention of the term "open standard" prior to the introduction of "open source" in 1998. The term simply wasn't in common use. After that, many companies have been trying to misrepresent both their software and their standards as "open" in order to mislead customers into thinking that their products are something that they are not.
That's completely, utterly false. 'The Open Group' standards for DCE, UNIX and X dates back to the 1980's. The OMG had open standards in the early 90's for distributed objects. The ANSI, ISO, and IEEE go much further back (POSIX dates back to 1988).
Open source reference implementations are useful to supplement standards, but they're two different things, with two different outcomes. Open source without an open (potentially standard) interoperability architecture is unlikely to generate interoperable & competing implementations. Sure, you can always fork, but that leads to a cacaphony of slightly differing and incompatible options that geeks might love but most customers despise.
On the other hand, open standards without an open source reference implementation may cause problems with the standard's proper adoption, as there's no example for implementors to use. But going too far on the open source side is also a risk to standards adoption -- if an open source RI is copylefted, that dissuades adoption in its own way. Whereas a more permissive license, say MIT, Apache, CC-Attribution, etc. would better incentivize adoption.
-Stu
http://en.wikipedia.org/wiki/Category:Microsoft_evangelists
http://msdn.microsoft.com/en-us/bb905078.aspx
http://www.nearsoft.com/blog/MS-test.html