Developer Only Discussion List

[Hamish Greig <hgreig AT bigpond.net.au>]

According to a news site poll our preference for 1.0 release was both a rock 
solid sorcery ( http://news.sourcemage.org/pollbooth.php?qid=1.0date&aid=-1 )
and a good installer/rescue cd, highest votes actually went to an iso with 
full sources, but considering versions can change on a daily basis, including 
sources in the iso, even for stable grimoire is a waste of time and space.  
( http://news.sourcemage.org/pollbooth.php?qid=isoSize&aid=-1 )
I have only been actively working on the ISO project for a month or so, and 
after browsing through bugzilla, old meeting logs, and even the old ISO teams 
roadmap it has become obvious there are a lot of features wanted/ requested 
for 1.0 that have not been implemented. Many bugs have been arbitrarily 
closed, others have been misfiled or poorly dealt with, and some others have 
been untouched for months.
I have no control over sorcery, but I can excercise control over the ISO.
So immediately below are several features I will implement for the 1.0 
release 
ISO, regardless of the upcoming deadline. Below that are possible additions, 
that although nice, can wait until later if necessary.
The one thing I do not plan on implementing myself in this time is hardware 
detection. Unet has been working on this, if his work is ready when mine is, 
then I shall merge it, test it and release an ISO we can all be proud of.

definite additions:
* choice of grub or lilo as bootloader
* individual spell caches instead of monolithic image.tar.bz2 making for 
better rescue performance and easier choice of software to install
* full software raid support 
* restart function in installer (to start over in case of mistakes without 
rebooting).
* addition of nfs and openssh software, optionally installed, for rapid 
deployment in a remote environment. (make it possible to install, reboot and 
have instant remote access, instead of rebooting and still having to setup 
remote access)
* also nfs will allow for netbooting/ net installs (using a floppy)
* menu selection of possible root devices within the initrd shell (for rescue 
mounting)
* a menu to select install or rescue mode , followed by menu driven 
partition mounting for rescue mode, so a system that won't boot (due to init 
or glibc problems etc) can be recovered easily without shelling out , 
manually running a lot of commands and untarring a 600MB image.tar.bz2 .

possible additions:
* gcc inclusion in iso_root so CROSS_INSTALL can be used to recover a system 
using latest grimoire version (instead of the cached version on the ISO ), 
using CROSS_INSTALL and INSTALL_ROOT=/mnt/root would let you run cast --fix 
from the ISO.(only if sorcery is fixed to not use absolute paths for entries 
in install and md5sum logs)
* FHS /srv and /media usage (voted against by us en masse, but is still 
becoming a standard so may be added)
* netboot floppy
* full pcmcia support (depends on feedback from some testers)


After this 1.0 ISO has been finished and released I will retire from ISO 
lead. 
I WILL fix any critical bugs if they exist, but my intention is to leave a 
suitable infrastructure so anyone can take over and continue the work.
I will also be retiring from grimoire duties. I may continue to use SMGL and 
file bugs, or I may not, I can't say at this point in time. 
The simple reason I am leaving is that working on SMGL is no longer fun for 
me.
Despite all the work I have done in the last year, I now find myself working 
harder than ever before, trying to achieve what we all wanted but no-one  
managed to organise or implement. This is frustrating as hell. I know there 
is no more I could have done, I spent hours working on grimoire bugs, I 
tried to start discussion of teams roadmap goals and their lack of 
achievement several times, I tried to communicate with other developers 
openly in an attempt to help them plan their work and give them support if 
they needed it. I attended all the meetings and made myself available on IRC 
and mail lists. I filed bugs when I found them, and posted comments to others 
when I had input relevant to that bug. At times my frustration even caused me 
to be rude to people, I don't apologise for this, I do however wish I had 
managed to get my dissatisfaction across another way. 
To put it bluntly I hate the lack of open communication within this 
"opensource" project and I am giving up.

I will only announce and release  1.0 ISO when I consider it it has been 
fully 
tested and is ready, regardless of external pressures.

Hamish

Attachment: pgpIAqHo1gvET.pgp
Description: signature