Re: NetBSD Wishlist

[Thomas Adam <thomas%xteddy.org@localhost>]

Hi --

On Mon, Nov 22, 2010 at 04:41:26PM +0100, MTS Consult wrote:
> Hello,
> 
> Here my wishlist for NetBSD 6.0 to have more fun using it (and really
> using it and not switching to other systems):
> 
> 
> + please drop CVS,RCS,SVN . They are not correct, slow, take too much
> CPU and just pain in the ass to use
>    DragonFly BSD has shown how awesome a distributed cvs is: they use
> Linus' GiT.
>    So why not developing a Distributed Tracking System the NetBSD way?
> A NetBSD repository manager: Ngit.
>    http://www.youtube.com/watch?v=4XpnKHJAok8
> http://www.youtube.com/watch?v=8dhZ9BXQgc4

Heh.  Once more round the mulberry bush we go...

http://mail-index.netbsd.org/tech-repository/2009/04/29/msg000107.html

And all subsequent threads.  It's long.

I do personally intend to resurrect this particular issue, albeit for myself
-- and to stop hashing theoretical technical merits.  Apparently actions
speak louder than words here, so I shall get to this when I have the proper
amount of time to dedicate to it.  For now, I have other NetBSD commitments
to see to.

I personally think that Fossil is completely the wrong choice, but we'll
see.

> + Out of the Box Kernel enabled Full Virtualisation expandable by
> different userland modules on different platforms - Why do we have
> XEN?

What else would you have?  Xen is a good choice, no?  KVM is too tied to
Linux to be of any use external to it.  Are you referring to something else
here though when you say virtualisation?  And note that, what *is* "Full
Virtualisation" anyway, if it's not what we have already in terms of Xen?

> + Support for more Filesystems. Whats going on in LINUX is crazy but

What does LINUX here stand for?  :)

> also awesome. New FS like JFFS2, RFS, BTRFS rock! they should be
> considerd to be supported right from the beginning. NetBSD
> intercompatibility with different MS and LINUX FS is annoying. NetBSD

I agree it should be an *aim* of NetBSD to do this -- and to be fair it does
support a fair few file systems, but do you know how difficult this actually
is?  Just taking BTFS alone would take months to get right.

Plus you also have to look at it from a cost-effective point of view as
well.  We should be thinking about supporting those file systems which will
gain the most use.

> should be able to talk natively to all other Storages from different
> Systems. Thats suits NetBSD better as it runs on all systems but just
> could talk to itself and other BSDs... Funny that with special
> software even Windows and LINUX could write to slices but NetBSD still
> is behind. Also I like to see HAMMERFS integrated natively and
> certainly ZFS. UFS and FFS are - uhm... old in every aspect and
> fashion. Maybe a Virtual Filesystem Layer (VFS) in the Kernel is
> helping there.

HAMMERFS is interesting, yes.  Its revision history implementation looks
interesting.  But again, see above. 

> + NetBSD LIVE CD/USB with each release - also would be nice to have
> NetBSD Server and NetBSD Desktop showcase CD with each release! I dont
> want to build each time the stuff I need. Even as Admin sometimes I
> just want to use it. LINUX CDs are so handy and I am mad about the
> fact that I cant just quickly download a NetBSD Server Edition and
> "WORK" on a problem instead of first putting hours into creating a
> system. NetBSD is time consuming like hell.

Have you seen Jibbed?  http://www.jibbed.org/

> + integrating a small X-Window into the System.
>    Something like Project "Wayland" from Canonical that has a
> minimalistic mem footprint expandable by userland modules.

Wayland is just a compositor -- no, no, you can't just throw away X11 like
that.  Don't be a bandwagon-junkie here.  This is XGL all over again.  :)

> Now flame

I prefer constructiveness.  :)

-- Thomas Adam

-- 
"Deep in my heart I wish I was wrong.  But deep in my heart I know I am
not." -- Morrissey ("Girl Least Likely To" -- off of Viva Hate.)