sjh - mountain biking running linux vegan geek spice - mtb / vegan / running / linux / canberra / cycling / etc

Steven Hanley hackergotchi picture Steven
Hanley
About

email: sjh@svana.org

web: https://svana.org/sjh
twitter: https://twitter.com/sjhmtb
instagram: https://instagram.com/sjhmtb

Other online diaries:

Aaron Broughton,
Andrew Pollock,
Anthony Towns,
Chris Yeoh,
Martijn van Oosterhout,
Michael Davies,
Michael Still,
Tony Breeds,

Links:

Linux Weekly News,
XKCD,
Girl Genius,
Planet Linux Australia,
Bilbys,
CORC,

Canberra Weather: forecast, radar.

Subscribe: rss, rss2.0, atom

←September→
Mon Tue Wed Thu Fri Sat Sun
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30      

←2004→
Months
Jan Feb Mar
Apr May Jun
Jul Aug Sep
Oct Nov Dec

Categories:

Archive by month:

Wed, 29 Sep 2004

More on the Linux v Sun discussion - 11:25
This morning I notice in Miguel de Icaza's activity log he makes mention of the Sun and Linux kernel discussion I have talked about Previously. Miguel suggests Greg KH is missing the point. I am not so sure, Greg was not as Miguel suggested arguing with the "everything we do is fine, there is no need to improve" viewpoint. Greg is a well balanced guy and looking at the crap he has dealt with on LKML and other places over the years he definitely seems to understand and respect other viewpoints and will change when a technically correct and superior change is displayed.

Miguel commented about Greg rejecting the Sun guys API stability arguments. I don't know that he rejected them so much as pointed out that that the API is stable in the kernel <-> userspace interface and has been for many years. Kind of like GTK or Mono or something having published API's and having internal structures. There is not much software if any that needs to use internal structures and such with those libraries. In the kernel though if someone has out of tree kernel code it has to keep up with the kernel internal structures. Andrew Morton has talked about this issue at OLS this year as have various other people, code that gets into the kernel will be maintained.

Of course the trick then is getting your code into the kernel, to do this you really need to grok Linux kernel culture and work with it. Mikal pointed out there seem to be exceptions where Linus or others appear arbitrary. Such as FUSE which Mikal suggests wont get into the kernel as Linus thinks it is too close to a Microkernel model. Personally I would hope there are good technical reasons FUSE has not been accepted rather than simply saying all file systems should be implemented entirely in kernel space (after all do we really want GMailFS in kernel space?) Of course Linus is only human (unlike Alan (more Linus quotes)) and has been known to allow code into the kernel in a strange manner in the past. Such as when Dave Miller got the bottom halves stuff in a few years ago. (anyone got a link to something about this I wonder?)

[/comp/linux] link

home, email, rss, rss2.0, atom