speakup serial support

Wed Feb 10 14:57:43 EST 2010

Hi Trevor,

You didn't imply that we were not concerned about this issue; you did
not come off in a bad way at all.  I just wanted to attempt to explain
what the issue is that we are running into.

We are definitely concerned about this, and it is being worked on.  It
is going to take some changes in the kernel itself to make it happen,
and we are waiting to see what happens in that area.  Like I said, I am
optimistic about it, so hopefully we will have some good news at some
point.

William

On Wed, Feb 10, 2010 at 02:44:20PM -0500, Trevor Astrope wrote:
> Thanks William. Now I understand what the issue is with the tty devices. 
> This is the part that illuded me before. It is a chicken and egg scenario 
> . I get it now.
> 
> And I didn't mean to imply that the speakup developers weren't concerned 
> about this issue and I apologize if it came off this way. I am sure 
> everyone is committed to the continuation of speakup as a kernel-level 
> screen reader and that these barriers will be overcome in time.
> 
> In the meantime it is just frustrating to not be able to access speakup on 
> my new desktop with a pci express serial card when I was so optimistic 
> that I would be able to get it to work.
> 
> Thanks,
> 
> Trevor
> 
> On Wed, 10 Feb 2010, William Hubbs wrote:
> 
> > Hi all,
> >
> > I agree with Trevor's statements below.  I have found in my area that
> > desktops with serial ports are extremely rare, if you are lucky enough
> > to find one at all.
> >
> > What we need is a way, as has been discussed, to use /dev/tty* devices
> > to talk to serial ports.  Once we can do that, speakup will be in a much
> > better position than it is right now.
> >
> > Samuel, please jump in and correct me if I am not clear about the
> > situation, but here is what I understand.
> >
> > The problem is that there is not currently a way to access a device,
> > even from inside the kernel, before /dev is mounted, so you would lose
> > the early boot messages if we switch over to that method.  But, we have
> > sent emails upstream regarding this issue, and the response we got was
> > very positive, so I am optimistic about this being fixed so that we can
> > eventually move away from directly accessing the i/o ports.
> >
> >
> > William
> >
> > On Wed, Feb 10, 2010 at 02:09:21PM -0500, Trevor Astrope wrote:
> >> John, yes I do manage servers and I do agree that having speakup in the
> >> kernel is immensely important if you have a serial port and synth.
> >>
> >> Perhaps it is different where you are, but where I live, desktops with
> >> serial ports are extremely rare. The only ones I managed to find were some
> >> low end Acers. I also found some business machines with serial ports, but
> >> they are twice the cost for about half the performance as a consumer
> >> desktop machine and we don't buy them where I work.
> >>
> >> Like Kelly mentioned earlier today, I also do not install speakup in the
> >> kernels of the servers I manage, but in the machine I use to manage the
> >> servers.
> >>
> >> So, I respectfully disagree with you about the availability of serial
> >> ports in modern desktop machines and I stand by my statement that speakup
> >> as a kernel-level speech system will become less relevant over time unless
> >> it can support external serial ports and usb serial ports. In my opinion,
> >> this is where speakup development should be focused, as more and more
> >> people will face this issue as they upgrade their machines. But I am not a
> >> speakup developer, so I have no influence on the direction it takes. I can
> >> only offer my opinion, which I have stated several times on this list and
> >> I can only hope that speakup developers agree with it and take up the
> >> challenge.
> >>
> >> In the meantime, I do have a job to do and I will need to decide whether I
> >> continue using speakup with software synth, which will make my job more
> >> difficult or use something else like orca or a mac, which still won't
> >> solve the problem of having access to early kernel messages, but may give
> >> me more flexability going forward.
> >>
> >> On Wed, 10 Feb 2010, John G. Heim wrote:
> >>
> >>> Well, perhaps its a minor point but plenty of modern computers have serial
> >>> ports. I've never seen a server that didn't have a serial port. In fact,
> >>> except for laptops, I have yet to see a computer that doesn't have a serial
> >>> port. That includes the 200 or so desktop units we have where I work. Even
> >>> the machine I built myself has a serial port.
> >>>
> >>> It certainly is a huge over statement to say that having speakup in the
> >>> kernel  has no advantage. If you manage servers like I do, having speakup in
> >>> the kernel is just about the most important thing there is for a screen
> >>> reader. I don't really care that much about what happens after the machine is
> >>> booted. About the only time I need a run time screen reader is if something
> >>> is wrong with networking. But mostly, I can admin these machines remotely
> >>> after they boot.
> >>>
> >>> ----- Original Message ----- From: "Trevor Astrope" <astrope at tabbweb.com>
> >>> To: "Speakup is a screen review system for Linux." <speakup at braille.uwo.ca>
> >>> Sent: Tuesday, February 09, 2010 3:09 PM
> >>> Subject: Re: Main advantages of SBL over Speakup
> >>>
> >>>
> >>> Samuel, do you mean there is no kernel convention for accessing serial
> >>> ports or there is no speakup support for accessing serial ports according
> >>> to kernel conventions?
> >>>
> >>> It would be really great if speakup could use ttyS# devices, so speakup
> >>> would work with modern motherboards that do not have built-in serial
> >>> ports. The way I see it is speakup can only use software speech on modern
> >>> computers, so unless it can access external serial ports or usb serial
> >>> ports, there really is no advantage to speakup being in the kernel so far
> >>> as I can tell...
> >>>
> >>> On Tue, 9 Feb 2010, Samuel Thibault wrote:
> >>>
> >>>> Bill Cox, le Tue 09 Feb 2010 14:23:25 -0500, a ?crit :
> >>>>> I hear that it doesn't follow kernel
> >>>>> programming conventions, for example in how it interfaces to the COM
> >>>>> ports.
> >>>>
> >>>> Yes, because no such thing exists (yet).
> >>>>
> >>>> Samuel
> >>>> _______________________________________________
> >>>> Speakup mailing list
> >>>> Speakup at braille.uwo.ca
> >>>> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> >>>>
> >>>>
> >>>
> >>>
> >>> --------------------------------------------------------------------------------
> >>>
> >>>
> >>>> _______________________________________________
> >>>> Speakup mailing list
> >>>> Speakup at braille.uwo.ca
> >>>> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> >>>>
> >>>
> >>> _______________________________________________
> >>> Speakup mailing list
> >>> Speakup at braille.uwo.ca
> >>> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> >>>
> >>>
> >
> >> _______________________________________________
> >> Speakup mailing list
> >> Speakup at braille.uwo.ca
> >> http://speech.braille.uwo.ca/mailman/listinfo/speakup
> >
> > _______________________________________________
> > Speakup mailing list
> > Speakup at braille.uwo.ca
> > http://speech.braille.uwo.ca/mailman/listinfo/speakup
> >
> _______________________________________________
> Speakup mailing list
> Speakup at braille.uwo.ca
> http://speech.braille.uwo.ca/mailman/listinfo/speakup