Tony Baechler here.
Kyle, first, from a different post, I have no feelings about Arch. I don't
know if they care about accessibility or not. I would think if they do, they
would make Talking Arch official, but there are probably technical issues.
Why they can't do like Slackware, Debian and Ubuntu, I don't know. Even a
total novice can press the letter "s" and Enter to start speech. As you say,
RH, Ubuntu and probably others have Alt-Super-S. There is no reason why Arch
can't do something similar. Better yet, adopt Talking Arch as an official
flavor and produce monthly images automatically along with the regular Arch,
eliminating the need for you to do it.
Second, you're right. I apologize. I was in a hurry. After rereading your
post, no, you didn't say RH hosted Orca. I very much doubt that RH made that
much effort to make their installer accessible, but I don't use it. They do
say that Fedora is accessible now, but I haven't looked at it in years. At
the time I went distro shopping last, Debian worked and Fedora didn't, so
that was that. That was around 2007. Gentoo didn't have a special talking
image either. It had Speakup built into the same kernel used by everyone
else. All you did is pass the speakup.synth parameter. Gentoo wasn't for the
novice and I wouldn't place it in the same category as Ubuntu. I have no
idea of the state of Gentoo anymore.
Finally, no, Speakup isn't my beloved by any means. I tried yasr and I
couldn't get it to work. Speakup with hardware speech is the only means of
hearing every boot message. Nothing in user space can do that. Can you put
Fenrir in the initrd? Oh, you can't? Well, no speech for you if your system
crashes or doesn't boot. I can't tell you how many times Speakup saved my
bacon when my system stopped at the initramfs prompt. Why can't you put
Fenrir in the initrd? Simply because you need a boatload of drivers for your
sound card, ESpeak, libespeak and who knows what else. That bloats the
initrd and it won't fit into memory. Unless, of course, Fenrir supports
serial synths. I've never heard of it, but I want to try it! I actually
think user space generally makes good sense for everyday use and on
desktops. For servers, you really need to know what's going on from boot to
shutdown.
As I understand it, it isn't the fault of Speakup that it took so long to
get to staging and is still there. The kernel developers were absolutely
opposed to including it and wouldn't help at all. I've read a message from
them saying as much. They were approached many times. Development died more
or less because it couldn't go any further. Without support from the kernel
upstream, it couldn't be part of the source and couldn't keep up with
developments. It's a miracle that it finally landed in staging. As you
rightly say, it took many years, far longer than it should have. I don't
blame Kirk and the others. Again, this is another great example of why we
need a nonprofit. I'm not really about the blind community getting together
and demanding change in most cases, but nothing is going to get done
otherwise. Speakup is a prime example. Without developer interest, it can
never meet kernel upstream standards.
As an aside, the reason why I only focus on getting the blind to try Linux
as opposed to everyone is for the above reasons. The more blind people we
can get, even if they aren't power users or coders, the louder voice we
have. As others have pointed out, stuff gets done with Windows companies.
Apple has VoiceOver. Not much gets done with most Linux upstreams unless
they really go out of their way to accomodate us poor blind folks. It
shouldn't be that way! Most of them have never heard of us and have no idea
what we need, how to help or what to do.
On 4/24/2017 5:40 AM, Linux for blind general discussion wrote:
Tony,
I said absolutely nothing of Red Hat hosting Orca. I said they ship it with
the distribution, which they are not at all obligated to do, as proven by
the fact that Linux Mint didn't come with Orca in the live environment for a
very long time. As for Speakup, it has never been fully ready for prime time
up to now, and there are very good reasons why it is still stuck in the
staging tree. If you want to talk about too little too late, then I would
talk of Speakup, which is only recently getting its act together enough to
hopefully make it out of staging and into the stable kernel tree, maybe in
the next couple of years if we're lucky. Meanwhile, we have a very nice
package called Fenrir, which has taken the screen reader completely out of
the kernel, putting it fully in userspace where it belongs. Perhaps this
will address the issue of speech from a text only environment much better
than Speakup ever could, as it can not only work on kernels without staging
enabled, but it will also eventually be far more portable to things like
FreeBSD, which has never had even a proof of concept kernel-based screen
reader, and has up to now required ssh in order to get it to do anything for
those of us who need speech output.
Regarding installer accessibility, I have used quite a few installers, and
Red Hat was one of the first major vendors to ship an installer that while
not accessible by direct methods e.g. via speech on the machine where the OS
was to be installed, did come with a method of gaining access to the
installation terminal via telnet, and also had kickstart files that could be
used in place of the on-screen system. Of course Speakup had to be used via
Speakup Modified, and before that, the kernel had to be patched, but I
wouldn't call that not caring by any stretch. Once the graphical environment
started becoming usable, Red Hat, now called Fedora, was already shipping
Orca in its repositories, and they were one of the first to include the
quite new at the time Espeak, which was far more responsive than Festival,
and all the other distros soon followed. I'm not sure where in the world you
have come to the conclusion that Red Hat simply doesn't care about
accessibility. Is it because your beloved Speakup, which is stuck in the
staging tree for more than 3 years now still isn't enabled in the Fedora
kernel? Sorry, but it's way past time to look elsewhere for text mode screen
reading to something that isn't locked into a kernel. No other screen reader
is bound to a kernel, and there are excellent reasons that go far deeper
than accessibility for disabling staging in a vendor kernel. Rather than
complaining that a distro vendor doesn't enable a potentially insecure
and/or unstable part of its kernel so that we can have a screen reader in
text mode, those who use text mode on a regular basis and need a screen
reader for it need to either learn how to muck about in the Linux kernel
itself so that the screen reader can get out of staging and into the kernel
proper, or better yet, contribute to Fenrir development, where everything
goes on in userspace and the screen reader only relies on interfaces to
stable and well-tested parts of the kernel that are never disabled in any
distro or vendor kernel. If Red Hat decides not to accept a Fenrir package,
then and only then can we begin to arrive at the conclusion that maybe
perhaps they don't give a care for accessibility.
~Kyle
_______________________________________________
Blinux-list mailing list
Blinux-list@redhat.com
https://www.redhat.com/mailman/listinfo/blinux-list
--
James 5:16 Confess therefore your sins one to another, and pray one for
another, that ye may be healed. The supplication of a righteous man
availeth much in its working. (ASV)
_______________________________________________
Blinux-list mailing list
Blinux-list@redhat.com
https://www.redhat.com/mailman/listinfo/blinux-list
