GTA02 is now an expensive brick

[Carsten Haitzler (The Rasterman)]

On Thu, 8 Jan 2009 04:40:25 -0800 (PST) Gothnet <openmoko at nastylittlehorse.net>
babbled:
 
> By the way, the Q and A email/looking forward to 2009 announcement from Sean
> Moss-Pultz just confirmed a lot of what I'd thought about the OM management
> situation. Where we have things like:
> 
> "People, like Harald and Raster, are given immense freedom to work on what
> they feel is most important."
> 
> And then a story about how 3+ months were spent reimplementing a
> keyboard....
> Perhaps people like Harald and Raster need their priorities given a nudge by
> OM management once in a while? You know what, I don't give a **** about
> which keyboard is best for terminal use, when I can't reliably make and
> receive a phonecall/SMS.

maybe you should have read my reply. it wasn't 3 months. also note that i am not
a kernel coder so just jumping in and adding more cooks into what was already
in the hands of kernel devs woukld be silly given that there was a mountain of
userspace stuff to solve. nor did i have gsm firmware in hand at any time. there
were people dedicated to just these. 3 guys worked pretty exclusively on the
kernel. another guy was the gsm firmware guy - another team was doing gsmd.
what you got from sean was a very surface-level answer.

the truth was that it was the engineers saying "lets just stop changing
direction and make the damn thing work". when sean and wolfgang and will went
"oh lets use qtopia - cant we just port it?" i was going "how about we just use
the 2007.x base and finish it to working? replace 1 bit at a time over time -
eg replace the wm and launcher, improve keyboard, replace dialler etc. keep
gsmd and just work on it to get it functional - make power management work".

months were sunk into diversity which was pretty much meant to ignore gsm/gprs
except for using it to break the telcos network - diversity was all about wifi,
voip etc. - you have no idea how much time and how many engineers were sunk into
that that went nowhere. then a retrophone idea (which was meant to be the phone
bit and was basically an afterthought after diversity) - which had no tie-in
technically or really to real work (eg use 2007.x).

then there came "now qtopia!" then that had to be ported to x11 (holger time),
then integration of qtopia and normal x11 and window manager stuff needed doing
(as qtopia really had no concept of a wm and was assuming it was in qws and
without one, owning the screen). then came all the new ui designs that had to be
"pixel perfect" implementations of flash stuff for the launcher and the whole
desktop (not done by me) - even the font was specified and font sizes and
positioning of stuff - when i remember saying "how about we just work with what
already works- in the way it works and improve it over time".

i distinctly remember advising against qtopia as it'd be a time sink to get to
work (if trying to mix with x11) considerign a team was dedicated to gsmd and
so much work had gone into 2007.x. of course as a result 2008.x which was meant
to make gta02 "workable" used none of 2007.x thus everything was lost. mind you
qtopia brought holger on board and that was a huge asset. dropping x11 wasn't an
option as we'd tie ourselves to a little-used windowing system (in open
source circles) and a toolkit pretty much moving all development to c++ and qt
as well as forcing the GPL license on all developers (no freedom to use other
licenses like bsd, mit-x11, etc. etc.) unless they paid up to trolltech
(who now are owned by a direct competitor too - nokia) further narrowing the
potential user and developer base - so the x11 port was pretty much a "no
choice" if qtopia was on the cards.

the real history is one of enigneers trying to push off the ever changing
flavor of the month idea and trying to just get stuff to work. while kernel
guys wrestled, userspace was being flung every which way... and not by me - not
by harald. guess who was driving all the changes? you might want to dig up a
bit more complete history before jumping to conclusions. :)

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