Ask Sean Moss-Pultz about community matters

Carsten Haitzler (The Rasterman) raster at rasterman.com
Tue Dec 16 00:53:58 CET 2008


On Mon, 15 Dec 2008 18:30:09 +0100 Didier Raboud <didier at raboud.com> babbled:

> Timo Juhani Lindfors wrote:
> 
> > "Yorick Moko" <yorickmoko at gmail.com> writes:
> >> * It's still more reliable and easy to use a 20EUR cellphone than use
> >> the FR as a daily phone, more than six months after Openmoko said it
> >> was mass market ready. I love the FR, but should we admit that some
> >> things really went wrong? (for example working on a graphical
> >> installer app before telephony worked fine, or when everyone was
> >> craving for the qwerty button (I never heard even one user who said he
> >> didn't want it) and nobody listened? Will things change?
> > 
> > I realize you are asking Sean but I can't resist giving my own
> > opinion: I've been thinking along the same lines. However, the
> > workaround to GSM bug #1024 was really not easy to discover (it relied
> > on undocumented AT command that afaik was not even in the NDA'd source
> > code openmoko people have access to). Also, the wlan issues required
> > deep kernel knowledge. Asking people who do pretty UIs to stop before
> > these are resolved probably does not do much good since they could not
> > help with the lower level stuff anyway?
> 
> Wrong people? Change people!
> 
> I mean: for me, Openmoko made a mistake by hiring GUI people instead of
> kernel ones at the beginning.
> 
> My questions would then be:
> 
> * Do you (as Openmoko's CEO) consider that mistakes have been made in the
> hiring process (in terms of goals and so in terms of capabilities, not in
> terms of persons) ?
>         * If so, what have you done or what will you do to correct these
> mistakes?
>         * If not, where are the mistakes (if there are after all)?

it has nothing to do with hiring gui people. believe it or not.. the gui people
do and can help with the lower level. for example - if gui people weren't
involved the /sys/class/power_supply interface for showing battery and
charging info wouldn't work as no one would have used/tested it - why? the
older gui only used the deprecated apm api from the kernel whihc provides a
much more limited view of information. gui people build infra that is
gui-friendly/ready that talks to the lower levels and manipulates them.
without someone manipulating them in an expected use pattern a lot of stuff
would remain broken. nothing to do with the gui people. more to do with how they
were directed - i.e. told what to do. i estimate about 50-70% of the time was
spent with fiddly "make it a perfect rendition of our flash mockup" than making
stuff actually solid and work as a foundation (even if a little clunkily). and
before you say it - "eye candy" comes for free when using EFL. so spending time
on it is a bogus argument as the libraries already do this and were already
written for it - they are just being used.

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster at rasterman.com





More information about the community mailing list