Le 23/04/2017 à 11:25, Linux for blind general discussion a écrit :
Your name isn't showing up for some reason. You make some good
arguments. Comments below.
Because the Mailing admin made it disappear to avoid massive spams users
received indi(idually after posting here.
On 4/21/2017 6:05 PM, Linux for blind general discussion wrote:
My only concern with this argument is that it seems like it takes
longer to
get fixes pushed up stream then it does to spin up a custom
distribution. It
also seems to me that a lot of what's needed in a custom distribution is
packaging and customization as opposed to programming.
From my experience, mostly what's needed is custom scripts and config
files. A lot of this could be done with custom packages. A thought I had
was asking for community submissions of scripts which make programs more
accessible. They could be put into either a tar archive or .deb package
which anyone could install on any system as config files aren't arch or
platform specific. I decided it would be too hard for one person to keep
up with and make sure new versions didn't break the old config files.
I really agree with this. To handle upgrades, we need non-regression
tests. And informing upst!eam of the existence of such "addons". But
yes, it's the good way of doing I think.
Initially, yes, it would take longer to get fixes pushed upstream.
That's why I suggest working with the upstream software developers and
not the distro developers. That way, a fix to ALSA or Firefox is picked
up by all major distros instead of one at a time. However, in some
cases, that isn't possible. I'm referring to installers. If a few
unknown people submit patches, the review process takes a long time and
the patches might not be accepted. If a nonprofit organization is formed
specifically for that purpose and not only submits patches but posts
them either on github or similar, not only do they get more review but
they are more likely to be taken seriously since there is an actual
organization behind them. I know there are exceptions, but I think most
people take Apache, Mozilla and Gnome much more seriously than a single,
random developer. When they know your reputation and you have submitted
enough patches, you can commit directly upstream, but as you say, that
takes a long time. Case in point, it can take years to be an official
Debian developer who can maintain packages without review.
Right. However as Debian has an accessibility team and is
accessibility-sensitive, patches will be reviewed quickly and applied.
The main instal;er maintainer knows fine accessibility.
I look at Vinux and Ubuntu as an example. It seems like Luke and the
Vinux
developers were able to get a lot more accomplished in a lot less time
working with Vinux then they were able to do by trying to push things
upstream in Ubuntu. Focusing on getting changes upstream, when there
could
have been a Vinux available, would have meant fewer blind Linux users and
some of those blind Linux users would have spent more of their own time
setting up and customizing Ubuntu and finding and installing accessible
applications. If you think vinux users are going to just quit and give up
when they run into a Linux issue, I'd think that problem would be even
more
prevalent if they didn't have a Vinux where a lot of the work was already
done for them.
Yes and no. First, don't forget that Vinux was itself based on Ubuntu.
It wasn't built from scratch. Ubuntu had serious accessibility problems
in the past. I wouldn't say it's perfect now, but it's a lot better.
Arch might be more current for the desktop, but it's a lot harder to
install. For a mainstream, graphical distro, Ubuntu MATE is very good.
When I tried Vinux, it had a broken installer and the system crashed. I
got a few comments that people gave up for those reasons. I understand
the major bugs are fixed now, but it still doesn't use what I would
consider a friendly, customizable installer. While I'm opposed to
blind-centric distros, I realize they have their place and some people
want them. I think what you say below makes very good sense. Create a
custom distro based on Fedora or whatever. When it's stable, work with
upstream to get those customizations included. What I would like to see
is a custom flavor of Ubuntu MATE. It would boot up talking and have the
features of Vinux but would be supported by the Ubuntu developers and
use the Ubuntu installer. Remember that just as 95% of Ubuntu is based
on Debian, at least 95% of Vinux is based on Ubuntu. Really, in terms of
code and custom packages, we're talking about a small number of bytes
compared to the overall distro. To me, it makes little sense to waste a
lot of time recycling 95% or more of what another distro has already
done just for a few little fixes.
Just to say that the Debian installer is accessible in braille, speech,
magnification, the graphical interface would be easily if work was done,
and when installing with speech or braille, the most accessible desktop
is automatically suggested for installing.
IMHO, I think a hybrid approach is the way to go. Build a custom
distribution, prove that it can work, build up your blind user base
and work
to get those changes upstream. I know it seems like this is spreading an
already thin resource even thinner, but I think it's the most likely
road to
success.
I hadn't thought of this in that way, but I think you have something.
The big problem with any niche distro, blind or not, is the small user
base and lack of support. If you can get even, say, a dozen developers
and dedicated support people and build up a solid user base, that speaks
louder to mainstream distros than a few blind guys playing around with a
distro one or two others put together. I remember in the early days of
Speakup that this was the case. No one took it seriously. There was no
official kernel support. When 2.4 came out, the patches for 2.2 kernels
broke. When 2.6 came out, the cycle repeated. I used custom boot
floppies to install Debian Potato because I wouldn't have Speakup
otherwise. Eventually, Gentoo, Debian and Slackware included it.
Finally, it made its way into staging and, unfortunately, still doesn't
have official kernel support. If a nonprofit could raise money even for
one paid developer, that would help immensely. Even FreeBSD includes
Orca because it's part of Gnome and officially supported.
I also think that custom distributions is just part of the Linux
ecosystem.
How many custom distributions are there out there to satisfy every
niche? I
don't think this should be any different for the blind Linux user.
There are about 300 as I recall. I doubt if you've heard of 250 of them.
Again, in principle, I agree. If I want to make a talking rescue CD or
someone else wants their pet distro, sure, go for it! If I implied that
I'm opposed to the freedom to do what you want, I apologize. I'm opposed
to blind-centric distros mostly due to the lack of resources and
community support. I wrote to a Sonar developer about having very low
volume. He said he didn't know how to fix it and that's as far as it
went. However, for those who want Linux From Scratch or whatever pet
distro they like, have fun! Just remember you're mostly on your own. I
find it interesting that more devices for the blind are built on Linux
or Android.
_______________________________________________
Blinux-list mailing list
Blinux-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/blinux-list