[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

focus for Embedded Debian


It's only a month to to Extremadura. I'd like stir some
flames^Wdiscussion on where we should focus most. If we don't 
have clear targets, we'll just end up hacking on random things. 
While random hacking isn't necessarily a bad thing, I'd like to be 
more ambitious.

Currently, emDebian is incomplete. Well maybe not that incomplete,
pieces exist, but they do not work as a whole. When people ask about
embedded debian, we give incoherant answers: "How does one
crosscompile for emDebian?" "Use scratchbox or dpkg-cross" -  
"How do I create a rootimage?" "Use emdebian scripts from 
stage.maemo.org or use use SLIND". And so on.

What I think an embedded linux developer typically needs from
their distro:

* a selected minimal set of packages
  - glibc/uclibc + busybox/gnu userland
  - with prebuilt rootfs's for popular architectures
* easy way to crosscompile their own applications and add them
  to their rootfs
  - where "own" application might be something like apache2
  - with whatever happens to be the most hot toolchain at that
* various easy ways to test and debug the system

With abudance of available hardware, they likely have their own or hw
manufacturers kernel, so the it's propable not in the scope of the
base embedded distro. 

It they want Debian, they probably also want:

* freedom
* apt-get
* Debians quality
* Debian's brand name
  - for these people we should also provide embedded ubuntu - that name
	has even more hype ;)

Anything important missing? Are these realistic things to target,
which parts of these are most incomplete in emdebian?


Reply to: