Linus on kernel management style
Posted Oct 7, 2004 2:14 UTC (Thu) by mmarq
Parent article: Linus on kernel management style
" The name of the game is to _avoid_ having to make a decision. In
particular, if somebody tells you "choose (a) or (b), we really need you
to decide on this", you're in trouble as a manager. "
I dont really envy Linus position... but he as proved also to be a good manager trough all this time, not only a good programmer.
The point * is_not_details *, but a "Declaration of Mission", and IMHO, i really dont know who started it, but i think that was what made Linux to make * that decisive jump into the IT mainstream *,... not fancy features that Solaris had better, and in same small cases still has, but the *Declaration of IT World Domination Purpose*, was what turned the fire into a explosion... look how feeble Open Source Software Everywhere 'sounds' compared with World Domination!
Quality of code was and still is very importante, its necessary but is not sufficient... so more importante than that, was Linux as a software project to have this very strong *Mission Declaration*: 'be the best', 'Nº1', that captivated and turned on fire so many imaginations arround the world, including ,IMO, the big corporations IBM, HP, SUN, INTEL, etc,... that where in a dire need to renew there missions, when the IT markets started to show signs of estagnation...
... perhaps against Linus own likes,... but from that moment foward Linux was dressed with political clothes allover... and not only can't escape it anymore... it has to fulfill it's destiny or die... eventualy.
Sounds dramatic ?... it is !... choices have to be made. And those choises IMO, dont concern more any possible very important particular patch, if it goes_in_or_not, than the purpose of *why*(its not a roadmap) that software is made...
Linux dosent have to stop being a top server system kernel to became a top Desktop system kernel !,... and for that what is missing is hardware support leveling MS Windows... so in the many changes that will be made eventually, and the many choices that have to be made too,... perhaps not exactly a) or b), but perhaps a version of A) plus a version of B)!,... i'm suggesting to complicate Linus life a hell lote more... and he's gonna feel like killing me for sure !(just kidding).
Anyway, is or isn't Linux right now perhaps an even bigger project than MS Windows(kernel) ever was, and perhaps on pair with Longhorn(kernel) ?
Many suggestions lay in the air,... including more maintenance help,... or dividing the main code tree into specialized ones when hitting 'code freeze'... because IMHO what goes on right now its not enterely suitable, because the wild changes period can be so dramatic, that when the Kernel reachs a supposed stable number(2.4, 2.6, 3.0?), after code freeze, it still is managed like a developing beta, and stays on like that, with VM or other intrusive changes in the middle, until it reaches a much higher sub-stable number, when then a *REAL* stable number is indeed achieved.
If a *real* stable 3.0(example) could endure for the average 2, 21/2 years that goes between the development series, and i mean a *real* stable Kernel, with perhaps only a 3.0.5(example) in between if justifiable, then "we" would have much more drivers, be it in loadable kernel modules, in DKMS, or else...
Then a new development serie could start right away after a stable Kernel, and development is where this exposition here;
makes all the sense,... and not in *stable*, where it colides somehow with what is needed, specialy when it has to take in consideration users that, even if they know, are not inclined to change code, or even to compile the kernel.
In conclusion, there would be stable kernels, not stable *series*, because that is for the development cicle.
There is a golden opportunity with the integration of telephony(VoIP, h323), with the integration of Digital TV (HDTV) and after that with Mass Media distribution. Microsoft, http://www.osnews.com/story.php?news_id=8448
is prepared to move on and try new lock-ins (DRM),... if only the Open Source World could generate more broader acceptance among hardware manufactures and *all* (BSD, OpenDarwin) kernel developers,... everybody has much more to gain than to lose..
Althought not being a Linux kernel developer, i belive, deep inside "we" all secretly somehow aspire to the dream of a form of Open Source World Domination,... was with that feeling that i dared to make suggestions.
to post comments)