View Single Post
Community Council | Posts: 4,920 | Thanked: 12,867 times | Joined on May 2012 @ Southerrn Finland
#23
Originally Posted by gerbick View Post
I think we're looking at it two different ways. You've decided to embed yourself in the minutiae of FOSS whereas it is important for it to be embedded everywhere and thus that propagates FOSS and its meritocracy however whenever a corporation, say LinkSys does that, they do so without adding any fixes, concerns or other finds into the mainstream and create a fork that benefits only themselves. Making money out in the open is quite hard - meaning that it's great to have an army of developers solve your problems yet many bigger corporations rarely give back.
Giving back is an essential part of it, and I do understand how it is very difficult for commercial enterprises to work that way. It takes a lot of work to convince the organization to behave "the right way" and to upstream corrections and modifications and unless there is a clear reason for doing it the management will see it just as an useless expense to be rid of.

There is the added burden of license management; the products need to be layered correctly in order for keeping the in-house code separate from FOSS code; this is something that requires considerable amount of work so the smallest companies might find it too difficult to do.

We have solved this by adhering to few simple rules;
  • The OS is a separate build that works-by-itself; Everything on the OS level is licensed under GPL/BSD/Apache et al, we will not accept any proprietary-licensed code in OS.
  • The OS provides only dynamically linkable libraries to applications and we check the licensing of those libraries very carefully.
  • We meticulously offer all enhancements and fixes as patches to upstream. Our policy is to always take the latest code from upstream and only apply our own patches if our fixes are not accepted upstream. (This really makes our life a lot easier too, as we don't have to support a wide set of inhouse patches)
  • All company-confidental code is built into applications that run on top of the OS. We evaluate all linking to OS libraries and use od OS interfaces to make sure no cross-contamination happens


Originally Posted by gerbick View Post
But for how long? Many great projects with a lot of hope that started off promising simply die to lack of updates. I'll rather respect somebody that continues to support their endeavors or at least try to remain visible and drum up energy. Folks like Shuttleworth, Torvalds and even to a lesser extent Somasegar, all rallied behind projects that required many others to support their endeavors but they did so with a passion that made adding to the project feel like you were adding to the better not to the worse parts of the FOSS meritocracy.
I have a somewhat darwinistic view on this; if the project is viable, then there will be an user base and active developers for it. If not, well then perhaps it deserves to stagnate.
(I am not saying "die" here but "stagante", since I don't believe it is possible for a FOSS project to really die, it is there for all to see and for someone to pick up if ever found worthy.

To recap; if it doesn't take into fire maybe it really was not the "great project with a lot of hope" you mentioned?


Originally Posted by gerbick View Post
Allow me to simplify my statements and be brutally blunt: I meant these so-called experts that bury themselves into the FOSS environment and rarely add much more than snide remarks, contrarian beliefs and impose their point of view unto every single ****ing discussion and rarely add anything of worth, ever.
I certainly hope you don't view my contributions this way!


Originally Posted by gerbick View Post
That is the second part of why I disagree with the current iteration of FOSS meritocracy. It actually works, but how do you enforce corporation to give back? They take and take.

And if they take, they should pay (respect or money).
I see this is a problem somewhere, but I feel that those corporations that work the way you describe are wasting their own resources; In the end your own maintanance load grows to unmanageable size pretty quick if you are not synchronizing with upstream.
It would help that they would realize it is to their own benefit to work the FOSS way.
I do have hopes that more and more companies will start working towards that; internally the people who do the FOSS integration in the companies do realize this, it is only that they'd need to make the management also understand it and align the workflow correctly.
 

The Following 4 Users Say Thank You to juiceme For This Useful Post: