At the heart of every networking device is an operating system that enables traffic flow. In the case of networking vendor Juniper, that operating system for the past ten years has been JUNOS, a network operating system with its roots in the open source FreeBSD operating system.
Juniper has updated JUNOS every 90 days since 1998.
Every 90 days? How exact is this? On the second? I say we should jump!
“It’s funny. Almost everyone I talk to wants to know how we can continue to ship feature-rich releases every 90 days”
I could do that too by copying FreeBSD and setting my alarm carefully.
I think open source is cool too, but why must 100% of all software be open source?
Why can’t people just accept that some times, and maybe even most of them time open source is the best solution, but it is not the only solution to every problem.
Depends on how you define opensource and conforming to the licence. One only needs to look at the flack Apple got for the fact that they didn’t bend over backwards to provide assistance to KDE developers to get the changes merged from webkit into khtml.
For some, the mere presence of source code is enough to satisfy their definition of ‘open source’ whilst others claim that for something to be truly open source, it has to include the active participation of the said company in the development in a community like atmosphere.
As for 90 day development cycle, it would be an easy thing to do; just make sure that the updates you provide are either trivial or well tested. Give that they have only one target platform, and they can routinely test the OS, the release cycle of 90 days isn’t all that difficult.
There was a common saying in the Soviet Union:
“Anything that isn’t forbidden is compulsory. The trick is figuring out which is which.”
Just for the record (I know, this is a tired old topic), there’s a widely accepted and quite unambiguous definition of “open source software”, if by “widely accepted” I may mean the one found in Wikipedia:
http://en.wikipedia.org/wiki/Open-source_software
Of course, I won’t deny that everyone is free to invent and use other definitions, but they should warn about their non-mainstream definition if they use it in articles, blog entries and the like. Doing otherwise would be misleading and bad stile.
HERETIC! Thou shall not utter such blasphemy!
I STRIKE THEE DOWN!
Well, not me, but you will get modded down rather fast, I’m sure
you must be given options before you can choose … a few people must live in the extremes on both sides of the argument to give you the opportunity to decide how far you want to stand between the arguments ..
the world isnt black and white but some people must present it that way to clearly show the issues ..you can blur it to your convenience but its nice to know where the limits are ..
Edited 2008-07-18 21:09 UTC
Because…
The solution to the problem is the code itself. It being open or not doesn’t have any intrinsic properties that preclude it from solving any problem that is solvable with a Turing Machine.
They’re perfectly allowed to use FreeBSD in this way.
…how much the source code for this OS is modded from vanilla FreeBSD. If it’s not particularly modded, then the company can be forgiven for not giving back at all. On the other hand, if they’ve made substantial improvements to the codebase, it really is unconscionable that they would not be giving back. Yes, its allowed under the BSD license, but its not very nice. The fact that the BSD license allows this at all is one of the number one reasons why it stinks as a license.
*Alert* Stallman follower detected in quadrant B.
What you fail to understand is, people are mostly decent at some basic level… Although they made a proprietary product based on the BSD code base, they probably *do* report back to the maintainers.
A lot of commercial users of BSD products are more then willing to contribute back in some way.. but they’re not forced to do so at all.
Even if they don’t contribute back, they still have the include the licence notices in product documentation.
Edited 2008-07-19 16:21 UTC
I have my doubts as to whether Juniper has ever contributed anything at all to upstream FreeBSD. Certainly I can find no evidence of any such contributions after conducting a Google search. The very fact that they’re allowed to do this is, to me, a shame. If you benefit from FOSS, it just strikes me as greedy and wrong that you can get away with giving nothing back…
Well they did…
http://www.freebsd.org/news/newsflash.html
It took about 5 sec on google to find that information.
Hmm… They’ve been using FreeBSD code in their OS for a decade and all they’ve ever given back is a barebones port to niche platform. A port which they had completed YEARS before they made it available.
Not exactly a stellar track record. Yes the BSD license ALLOWS this, but that’s just proof of how broken and naive the BSD license is. The BSD license comes with a silent and unenforceable implication that “sharing is nice” The GPL is more like a kindergarten teacher that will put you in “time out” if you aren’t nice and don’t share. Given the level of greed in the tech world, the GPL approach is much more realistic vis a vis maintaining a truly free project.
i agree completely, i know and have worked for companies who have chosen BSD licensed stuff *PURELY* because they won’t have to contribute any of their precious IP back, they didn’t care about the merits of the code/OS.
mind you, i guess thats better than using GPL stuff and breaking the license….
Many developer chose to use it because they like it, who are you to criticize their choice?
For them, the BSD license is neither broken nor naive, you really sound obnoxious when you apply your own personal preference to everybody..
I wont bite on your gpl vs bsd bait. It’s just plain boring by now.
I don’t know if they gave more and neither do you by the looks of it. For all we know they could have donated a shitload of money.
They did.
<a href=’http://www.freebsdfoundation.org/donate/sponsors.shtml“>