[maemo-developers] [maemo-developers] I'd like to help improve boot times...

[From: Jose Antonio Vacas Martinez]

I think that improving the booting time would be a big step in maemo.
I've been using palm pdas for a long time and I miss the 0 seconds booting
time.

Magnus, if  I can help testing ask me.

Cheers

On Sat, Mar 15, 2008 at 3:26 AM, Ryan Pavlik <abiryan at ryand.net> wrote:

> I think it's worthwhile work to look into, because it does take a while
> to boot up.  True, I don't boot mine often either, but mostly because
> I've heard (and experience bears out) that on the devices 1 boot = ~3-5
> days of standby/sleep.  My thinking is perhaps if boot is faster, a)
> less power is needed and b) sophisticated PM can be brought up perhaps
> sooner (this is completely out of my realm though), besides the obvious
> effect of faster boot.
>
> (Probably the fact the device seems to have 3 separate splash screens
> then boots to a not-yet-functional desktop exacerbates the feeling of a
> slow boot.)
>
> Good luck!  I might suggest trying the VMWare Maemo SDK - for me it was
> way easier to get working than a direct install of the Scratchbox
> environment, but of course YMMV.
>
> Ryan
>
> Magnus Deininger wrote:
> > Well, true that but it sure can be an annoyance if your battery runs
> > out and the first thing you notice is that the device is having you
> > wait a couple of minutes. And the way I see it, if your device boots
> > fast enough, you could just as well shut it down completely instead of
> > putting it into stand-by mode, thus essentially giving you unlimited
> > stand-by time.
> >
> > On the bright side, if if bootup times aren't too much of a concern
> > for most users, einit also provides for a centralised hardware event
> > handler and allows the actions to this to have very little overhead by
> > being able to write handlers in C or anything else that compiles to an
> > ELF object file. It also has service supervision and provides for
> > backup plans if something won't start, which tend to come in handy a
> > lot on embedded devices.
> >
> > -- Magnus
> >
> > Quoting Frank Banul <frank.banul at gmail.com>:
> >
> >
> >> Hi,
> >>
> >> No offence but I would guess most people don't boot their devices
> >> very often. It would be nice to have faster boot times but I rarely
> >> boot my device. Just an opinion. I hope it's helpful.
> >>
> >> Frank
> >>
> >> ----- Original message -----
> >> Hi there,
> >>
> >> I'm Magnus Deininger, aka jyujin, head of the eINIT project, a~nd I'd
> >> like help the maemo project by improving bootup times. One of my
> >> friends is telling me that maemo devices are booting 'unbearably slow'
> >> and that this would be quite the complaint with customers; he's going
> >> so far as to offer his Nokia 770 for the cause if i can make it boot
> >> faster.
> >>
> >> My pet project (eINIT, http://einit.org) is currently doing reasonably
> >> well at making computers boot faster, but we haven't tried to make it
> >> work on too many embedded devices just yet, mostly because we're a bit
> >> tight on devs and time... and people that would be interested in
> >> seeing it on their embedded device. We're still doing quite the
> >> progress in boot times tho, as could be seen by this little video
> >> (taken by that friend of mine who'd like to see his nokia 770 boot
> >> faster):
> >> http://video.google.co.uk/videoplay?docid=-2744120711983143659 .
> >> That's booting fairly fast, especially considering it boots right into
> >> compiz... I also made this bootchart: http://omploader.org/vZHdj --
> >> which is on my main desktop, without any real optimisations. Both
> >> boxes (the one in the video, and mine) used to boot in about 1 to 1:20
> >> minutes before, so that's /quite/ an improvement right there.
> >>
> >> Now seeing as my friend has some issues with his device, I'd like to
> >> help him, and the maemo project, but since I'm a bit tight on time,
> >> I'd first like to know if anyone else is actually interested in me
> >> porting eINIT to maemo, since I'd hate setting up a devel environment
> >> for a single device ;). Which brings me to why I'm posting on this ML
> >> in the first place: I'd like to know if people /are/ interested in
> >> this or not. Just get right back at me if there's any interest. I'll
> >> monitor this ML for the time being, and I'd also like to invite any
> >> interested developers to join our IRC channel, #einit on freenode.
> >>
> >> So, that concludes my offer I think.
> >>
> >> Greets, and happy hacking,
> >> Magnus Deininger
> >> http://einit.org/
> >> IRC: jyujin on freenode; #einit
> >>
> >> PS.: Oh, not to forget, eINIT is BSD-licenced, so there shouldn't be
> >> any weird licencing issues from a business POV either, if that would
> >> ever be an issue.
> >>
> >>
>
>
> --
> Ryan Pavlik
> www.cleardefinition.com
>
> #282  +  (442) -  [X]
> A programmer started to cuss
> Because getting to sleep was a fuss
> As he lay there in bed
> Looping 'round in his head
> was: while(!asleep()) sheep++;
>
> _______________________________________________
> maemo-developers mailing list
> maemo-developers at maemo.org
> https://lists.maemo.org/mailman/listinfo/maemo-developers
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.maemo.org/pipermail/maemo-developers/attachments/20080316/ca6eaaae/attachment.htm