LWN: Comments on "A few notes on the OpenSolaris release"

The right to give up benchmarking privileges

giraffedata — Sun, 19 Jun 2005 00:10:42 +0000

A promise not to divulge information is very common in US business, and entirely enforceable under US law. I can't even think of a kind of promise of secrecy that isn't enforceable.

I like it that way. My promise of secrecy is valuable, and I demand the right to trade it.

Incidentally, at least one no-benchmark contract term I saw wasn't a restriction on divulging the benchmark results -- it was on actually performing the benchmark. That it's hard for the other party to know the benchmarker breached the contract unless the benchmarker says he did really doesn't make it a free speech issue.

A few notes on the OpenSolaris release

rickmoen — Sat, 18 Jun 2005 01:32:44 +0000

uwaucs wrote:

I mean no criticism of Sun Microsystems in saying this: (It's better to be more cautious about BXA regulations than less.) There really aren't issues with US export regulations and OpenSSH, and haven't been since October 2000, thanks mostly to Daniel J. Bernstein's litigation and some good pressure on the administrative law front from Red Hat, Inc.

For more on that matter, please see "Crypto Export" on http://linuxmafia.com/kb/Licensing_and_Law.

Rick Moen
rick@linuxmafia.com

A few notes on the OpenSolaris release

danieldk — Fri, 17 Jun 2005 23:41:49 +0000

How much is US export policy on crypto still a problem? As far as I know it is legal to export crypto software when a.) the source is available, and b.) you drop an e-mail notice at the US Bureau of Industry and Security.

http://www.bxa.doc.gov/Encryption/PubAvailEncSourceCodeNo...

But I am not a lawyer, so I may be understating the problem.

A few notes on the OpenSolaris release

uwaucs — Thu, 16 Jun 2005 16:18:28 +0000

http://blogs.sun.com/nico/20050610 SSH wasn't included in the relase because it was forgotten about until too late. There are also issues with crypto code due to US export regulations (linked from above). If you look at the roadmap http://www.opensolaris.org/os/about/roadmap/ you'll see they plan to release more components under the CDDL as time goes by.

I have to disagree with the community comment in the main article - there is already a thriving community of blogs http://planetsolaris.org/ http://www.opensolaris.org/os/blogs/ , many by internal sun developers, but also some people external to Sun who got early releases.

A few notes on the OpenSolaris release

dougm — Thu, 16 Jun 2005 15:22:49 +0000

od, rdist, telnet and sed are original Unix utilites--Sun's versions may very well be based on that code for compatibility reasons (older Solaris versions had the SysV versions of these programs as the default install), and they may not be able to redistribute source for them yet.

ssh is strange, but its BSD license permits binary redistribution. Not sure if they're withholding source for (bogus) security reasons or what...

The current version of patch is under the GPL, but I think older versions (originally written by Larry Wall) are under the Artistic license as well, so it's possible Sun forked one of those for its distribution.

A few notes on the OpenSolaris release

ekj — Thu, 16 Jun 2005 12:12:26 +0000

License conditions like that are, if not outlawed, then atleast void in a number of jurisdictions.

Saying you may not /publish/ benchmarks is a restriction on speech. Anyone can do benchmarks, they're just not allowed to tell anyone about their results.

Atleast in Norway those are void. Restrictions on speech in contracts are only allowable in certain limited cases which in general require a "honorable reason".

Wishing that only benchmarks "approved" by you can be published is not "honorable". Wishing that an employee that work for you and learn about secret plans for a new product refrain from telling your competitors probably is.

A few notes on the OpenSolaris release

tzafrir — Thu, 16 Jun 2005 06:17:23 +0000

However with the assignment to the FSF you can be assured that your code will only be used in free software. Here Sun can use it in its own product that will compete with yours. And they can actually "fork" your code for usage in their own totally propritary products.

A few notes on the OpenSolaris release

jamesh — Thu, 16 Jun 2005 06:14:08 +0000

There are a few differences between the Sun agreement and the FSF agreement though:

The FSF agreement is a copyright assignment rather than a joint copyright agreement (although the FSF licenses the code back to you under non-restrictive terms).
The FSF agreement includes a promise that they will release the code only under a copyleft license. If they break the promise, the assignment contract is void.

There are enough differences that you can't say that it is the same tradeoff. Whether this will affect the community that builds around the code is another matter.

A few notes on the OpenSolaris release

eru — Thu, 16 Jun 2005 05:55:27 +0000

The Sun compilers come with a rather more restrictive license; users [...] may not publish benchmark results,

License conditions like that should simply be outlawed. Traditionally, when you publish something, you must be prepared for public criticism of your work. That is how science, and culture in general has progressed. Software should not be exempt on any pretext.

Interesting, according to one developer, much of the pain of porting to gcc was caused by gcc's insistence on putting constant data into read-only memory. OpenSolaris, it seems, did a lot of writing to "constant" strings.

Say what? In this day and age, any C code that writes into string literals is considered childishly incorrect. Downgrades significantly my assumptions about Solaris code quality.

A few notes on the OpenSolaris release

robla — Thu, 16 Jun 2005 05:39:43 +0000

LWN wrote: It doesn't seem to have occurred to Sun that [a joint copyright assignment] might reduce the size of the "vibrant developer community" that it wishes to create.

...or, perhaps it was a tradeoff. Perhaps it was the exact same tradeoff that the Free Software Foundation also makes with software it develops. It's a pretty standard practice for very good reasons that a good lawyer would be able to explain to you.

Methinks you should give Sun's lawyers and business people a little benefit of the doubt when it comes to thinking about these things. They may not have made the decision that you would have made, but I can pretty much guarantee they've thought about it.

Rob