GTK netinst cd snapshot available
after yesterday's d-i meeting i'm more and more convinced that we should
prepare some kind of experimental graphical installer that even non
d-i experts can easily use.
This because of many resons:
1)As time passes by the graphical debian installer may seem more and
more vaporware even to people involved in d-i development.
If nothing happens the interest for this project inside the d-i
community will sink down and development will slow down until it stops.
2)We need experts in font packaging: i think no one will ever prepare,
let's say, a new set of fonts if he's not sure his work will be really used.
If we can demonstrate that the graphical d-i REALLY works, and not only
exists as a bunch of source files and udebs, we'll be able to attract
more developers who may help us.
To do this the proper way of course we need the missing udebs (fonts,
libraries etc..) that are being prepared right now.
But, as someone yesterday suggested, we cannot wait to have all the
perfect ingredients befor relasing something: even if dinner is not
ready yet let's start to prepare a little appetizer.
Since i'm not very pratical of udeb packaging i've prepared an hacky
inside its initrd (50 Megs) i've placed the d-i_dfb.tar.gz testcase
files (some non latin fonts are included too).
This should allow us to say to people in outer world "Hey, do you know a
graphical d-i is being developed? just download this .iso, burn it and
press enter at boot time to give it a try".
If we are lucky we may find developers, font designers etc.. that may
get interested in this project out and help us in doing what we cannot
do by ourselves.
I also think we should stay with DFB rather than temporarily switching
to X because to come back later to DFB, because
1)Switching cost (in terms of time) : developing, packaging, testing an
X environement that has to be used for some onths only would require
many energies to d-i developers.
On the contrary DFB environement is almost enterily packaged (we have
also a set of fonts, even if ugly)
2)We must demonstrate to Mike Emmel his efforts in making GDKDFB work
are worth it: we'll be then soon able to switch to GTK 2.8 on a DFB backend.
3)DFB should be better that X in long-term because of its smaller size.
what to do with this cd? spread the voice, give it to try to people you
may think that are able to help us and let's hope this speeds up the
have a nice weekend! < Here in Turin is a rainy day :( >