Re: Switching Algebra and CIrcuit Design for the SIght Impaired (Newbie)?

Hi Jason,
And thanks for replying. I initialy thought I would get no replies at all.

Jasen Betts wrote:
> Some circuit simulation software like LT-Spice (free and quite good)
I just gave WinSpice a try but cannot make much out of it frankly speaking.
I have no need for simulation though it may be necessary in some other
courses. In fact, I think having no simulation might be the best option in
an exam, <grin>.

> if, being legally blind, you can get your head around a chessboard full
> of pieces
I think it depends A lot on how complex the software is and how motivated I
am. I'm able to use early versions of Reaktor for synthesizer design at a
musician-friendly level, even though the interface is a bit on the complex
side. Custom controls, bitmapped text, poor contrast and no keyboard
interface could make any app relatively hard to use, though,

> to read and understand raw spice netlists, (a text-based language
That could work. I generally prefer textual formats, though they are often
linear and achieving clear hierarchies would require some software that let
one concentrate on the structure at different levels. A Windows tree control
as in CHM help files comes to mind as the first example. You can close the
branches you don't need. A similar effect is hard to achieve in most text
editors.

Another requirement on the course I'm doing seme to be Karnaugh maps. Again
VIsio cannot do them and I would prefer some more specialized piece of
software. I'm currently using ASCII tables but it is a bit clumsy.

> multiple feedback paths are involved the notation fails.
That's a good point. FOr doing graphs perhaps Graphwiz could be used to a
certain extent:

http://www.graphviz.org/

It does allow you to concentrate on the essentials and leave laying out the
elements visually to the computer.

> state transition diagrams or tables can explain more complex systems
Tables could be represented in ASCII and maybe I'll find some notation for
state transitions, too.

> They used (and still do use) text characcters that approximate the circuit
> sumbols.
Ah guessed that. The same tac is used in the port symbols where the stuff
inside the rects could be written in ASCII. Such as: & for an and port and
>=1 for or, for instance.

> --->|-- is a diode --||-- a capacior and --[100]-- a 100 ohm
> resistor etc..
> expect these circuit diagrams are
> totally unintelligible when rendered by audio reader software.
Yes, they are, for the most part. The notation is usable on a braille
display or if using full-screen magnification but pretty slow still. For
speech the main problem is that it doesn't understand the syntax used. The
screen reader would need to be linked to some intelligent parser or lexer
for a sensible audio representation.

As a case study, here's how one of your earlier lines is like using Dolphin
Supernova.

First, full punctuation.

> --->|-- is a diode --||-- a capacitor and --[100]-- a 100 ohm

Is renderred in speech as:

greater than dash dash dash greater than vertical line dash dash is a
diode dash dash vertical line vertical line dash dash a capacitor and dash
dash left bracket one hundred right bracket dash dash a one hundred ohm

Sorry if the layout is a bit off, I was lazy enough to use a find and
replace operation here.

I haven't structured the prompt in any way because it is read out linearly
from left to right without any major pauses in it. Some readers offer a
marginal improvement by counting repetitiv symbols but it still doesn't make
things much clearer. I also lowercased the text to highlight the fact that
case changes are not usually spoken though they can constitute a word
delimiter in pronounciation.

As an equally unclear but less straining option, the same thing again this
time with no punctuation:

is a diode a capacitor and one hundred 'ay' one hundred ohm

Naturally, I'm mostly using no punctuation in something like e-mail.

If you'd like to know more, I could point you to a thread in another group
dealing with programming with speech:

http://tinyurl.com/bkks7

For more sight info, I've got a page about that, too:

http://www.student.oulu.fi/~vtatila/sight.html

> when communicating on the telephone the only method is the netlist
> label the nodes in the circuit and then describe how the nodes are
> connected by the components,
So this comes down to serializing and reconstructing a graphical
representation to text, I think .

> A' looks like A-prime to me I think ~A could be a better symbol
Yeah, some languages even use ~= for unequality. Not to be confusd with the
operator =~.

> I'd go with "&" or "." for "and", "|" or "+" for "or", and "~"
My original goal was to stay true to the Boolean algebra notation. Some
e-texts also use the dot, plus and the single quoet char respectively. Yet
none of these are speech friendly (quote, ampersand, period).

If operators made of word characters are allowed then:

a and (not(b) or c)

would be pretty straight forward. Another way is writing the operators as
though they were functions:

and(a, or(not(b), c))

Or for a Lisp-like flavor:

(and a (or (not b) c))

Heck, post-fix is also a possibility <smile>:

b not c or a and

I'm only kidding with this last one.

> C (C++. java, etc) programmers would probably prefer ! && || and ^.
That's even worse, ampersand ampersand etc... Boolean is pretty bitwise by
nature, isn't it.

--
With kind regards Veli-Pekka Tätilä (vtatila@xxxxxxxxxxxxxxxxxxxx)
Accessibility, game music, synthesizers and programming:
http://www.student.oulu.fi/~vtatila/


.

Loading