[Blindmath] Working math homework and exams

Sat Nov 19 19:15:02 UTC 2011

Hello,
Its been interesting reading your comments on what you have done and what you find difficult. I am sure that having views from people like you will help those developing software to understand what needs attention and how to solve the problems.

To answer some of your questions, not necessarily in order.

To answer your first question, the Braille display tends to follow the system focus/cursor. As you type the cursor has moved so the Braille display moves back to the cursor. Can this be solved, may be but it probably depends on screen reader. I know that NVDA has options for whether the Braille display should follow the system cursor or the review cursor and that there is options for whether the review cursor should then also follow system focus when it changes. So probably with NVDA there might be a way to configure it the way you need. As for other screen readers, I don't know, I have the feeling that orca on Linux does not have suitable options to get Braille to perform how you want.

I'll jump to your last question, if there is a way of producing nemeth from MathType. I thought the tiger software suite for the tiger printer was able to produce Nemeth from a word document containing MathType equations or you could use dots plus if embossing it on a tiger printer. Alternatively you could export to MathML and then use liblouisUTDML (http://liblouisutdml.googlecode.com), I don't know if liblouisutdml being a command line application would be a problem for you.

Now back to your third question, identifying different dot patterns in Braille which have the same shape but different position. Unfortunately this probably will really only improve in time as you do more. Well I say improve, that was more about reading efficiently, there may be small things which you can do if uncertain but they may slow things down while checking. Normally I mainly rely on characters around to help with positioning, but when something is isolated on its own this doesn't help. On paper remember there may be a line above or below which may help you estimate position but beyond that there isn't much more. On a Braille display you can use other things. On my alva display you can feel a slight join between each cell, particularly if you gently run a finger nail across. Also there are the routing keys just above each cell, this can help judge whether the dots are to the left or right of the cell. Also if really stuck I guess you could take the cursor to it and so then judge things from the dots used by the cursor (I have my cursor set as the extra two dots of the display cell, dots 7 and 8).

Michael Whapples 
On 19 Nov 2011, at 16:37, Ben Humphreys wrote:

> Louis,
> 
> Thanks for the recommendations but I have several questions:
> 
> 1.  When using a Braille display,you're going to have to necessarily separate the braille cursor line from the active cursor line you're working on so as to refer to previous work while typing new work.  The whiz wheels on the FS displays seem good for this but I've found the braille cursor has a nasty habit of jumping back to the active cursor line without being asked to.
> 
> 2.  Having lost my eyesight at 40, learning Braille sufficient to read efficiently, let alone accurately enough to do math is no easy endeavor.  While I really appreciate that I enjoyed eyesight for the first half of my life, I envey students who learned braille as a kid when their brains were spongy and getting proficient was relatively effortless.
> 
> 3.  How does one differentiate 2 from b from ' from ^ accurately?  On my display, at least, they are all 2 vertical dots. in various permutations.
> 
> 4.  As for Nemeth, I'm not sure what the benefit is here, other than yet another step to get from Mathtype to a format capable of reading.  I could see if the original content was in Nemeth but how much original university-level math is in Nemeth format?
> 
> Aren't you going to need a human to perform that extra conversion step for you?  Is there a Mathtype to Nemeth converter?
> 
> One of these days, if I ever experience a Mathplayer that actually works, I may develop a whole new appreciation for Math ml.  Unfortunately, the issues necessary to get this working for the uninitiated such as myself have been too numerous to overcome.
> 
> Great discusion though!
> 
> Thanks
> 
> Ben
> 
> At 10:25 AM 11/19/2011, you wrote:
>> Hello Ben,
>> 
>> Three comments.
>> 
>> 1.  A Braille display would greatly lesson the memory issue.
>> 2. You could open a second file for scratch purposes.  Alt-tabbing between
>> two files is faster than going up and down a single file.
>> 3. Nemeth code is invaluable.
>> 
>> Braille will greatly simplify your scientific efforts.
>> 
>> 
>> 
>> Regards
>> Louis Maher
>> 713-444-7838
>> ljmaher at swbell.net
>> http://www.nfbtx.org/localchapters/houston
>> 
>> 
>> -----Original Message-----
>> From: blindmath-bounces at nfbnet.org [mailto:blindmath-bounces at nfbnet.org] On
>> Behalf Of Ben Humphreys
>> Sent: Saturday, November 19, 2011 9:00 AM
>> To: blindmath at nfbnet.org
>> Cc: Doris Pichardo
>> Subject: [Blindmath] Working math homework and exams
>> 
>> Hi everyone on the Blind Math mailing list
>> 
>> Earlier in this term, we had a brief discussion on how to make use of a text
>> editor to take notes and do math homework.  At the time, the discussion
>> focused on how to represent math constructs like exponents, division,
>> special symbols etc.
>> 
>> I am now towards the end of my first calculus class, and my first math class
>> as a totally blind student.
>> 
>> Brief results: So far, have a B- in the class whereas I believe I would be
>> much closer to an A, homework and exams take 2-3 times longer to do than
>> other students.  Use of Tiger embosser invaluable for visualizing graphs but
>> a lot of extra prep work necessary to get math material and graphss in a
>> form suitable for embossing or reading and doing homework.
>> 
>> My instructor uses several formats for material.  Most often, she creates
>> material in Microsoft Word using Mathtype.  I obtain the Microsoft Word file
>> from her in electronic format, use Mathtype's convert to Latex feature, then
>> heavily process with a Perl script to remove all the extraneous junk and put
>> it in a straightforward format that I can read in Notepad.  Before I wrote
>> the Perl script, I required a human to remove all the extraneous Latex and
>> the human found it faster to type from scratch than to fixup the Latex.
>> 
>> Other times, my instructor scans in problems found on the web or out of
>> other textbooks.  These have to be typed in by a human so I can read them in
>> a text editor.
>> 
>> Finally, she solves class material and homework in handwriting and places
>> the scanned images of that online for the benefit of all the students.  The
>> format is scanned PDF. So I need a tutor to compare my homework with the
>> correct solutions.
>> 
>> Our "official" textbook is available from Recording for the Blind (Learning
>> Ally) but it's not only cryptic as most math texts are, but the readers are
>> aweful. While I appreciate the volunteers who record these textbooks, it's
>> an exceedingly hard thing to render a math textbook in audio format.  I get
>> stressed out just listening to those poor people trying to describe a crazy
>> equation or graph in words.
>> 
>> So that's the logistics of being a first-time blind calc student in a
>> university setting with no "inline" instruction such as might be provided
>> back in secondary school.
>> 
>> 
>> Back to the original story...
>> 
>> I planned to do my hhomework and exams in a text editor.  I choose this
>> approach for several reasons:
>> 
>> 1.  I'm a good typist and excellent JAWS screen reader user
>> 
>> 2.  Handwriting is out of the question
>> 
>> 3.  I am just learning Braille and so I figured a text editor + JAWS would
>> be much more reliable.  If you've ever compared the 2, b, apostrophe, and ^
>> characters on a Braille display, you can easily see how a beginner would
>> completely blow an equation like
>> y'=2b^2   because 5 of the characters are all two vertical dots in
>> various configurations.  Recipe for disaster...
>> 
>> 4.  I've heard in the past that folks used to use a Perkins brailler.  This
>> had the advantage that you could type your work in braille, and refer back
>> to previous work on the page relatively easily.
>> 
>> Unfortunately, this sounded like a heavy, noisy, and impractical solution
>> for me since I don't know braille that well.  And I was unsure how one makes
>> corrections (erasures or cross-outs) using such a method.
>> 
>> So I choose Notepad + JAWS as my solution.
>> 
>> I choose early on to use my own simplified expressions in lieu of
>> complicating my life with Nemeth or for heaven's sake Latexin a text editor.
>> 
>> So x-squared became x^2
>> a/b is a simple fraction
>> a+1//b was a shorthand to (a+1)/b and easier to understand
>> lim x->infinity-symbol (sideways 8)  became simply lim x~infinity
>> 
>> Then I used JAWS dictionary to redefine things like "^2" to be "squared", y'
>> to read "y prime", ~ for "goes to", and cos x to read cosine x.
>> 
>> My simple hybrid representations worked relatively well.
>> 
>> Now for the complications:
>> 
>> 1.  When equations got long, it was really easy to get "dizzy"
>> reading through them with a text editor.  Example:
>> 4x^5 - (2x^2)(3x) + 45 = 16x^(4/3)
>> 
>> 
>> 2.  As you work an equation with more than one or two terms, it was
>> really easy to make a mistake:    Example:
>> 
>> 2x^3 - 3x^2 = -5
>> could easily become
>> 3x^3 + 2x^2 = 5
>> 
>> Because the human brain can only remember about 7 things at once, and if you
>> look at that equation, there are about 9 individual coefficients, exponents,
>> and signs to remember.  It's enough to drive you mad.  So you use copy and
>> paste as you work the problem and then change little pieces on each
>> iteration.
>> 
>> 3.  The sighted students have the benefit of a scratchpad in the form of an
>> area off to the side of the page or on the back where "temporary"
>> calculations can be performed.  Then the result can be brought back into the
>> original problem.
>> 
>> I found I had to do my scratch work inline, which distracted me from the
>> original problem and forced me to scroll through long lists of calculations
>> to get to previous steps. this too was enough to make you dizzy and confused
>> just moving through all the math with a screen reader in your ear.
>> 
>>  So I started labeling my steps with comments like you might do in a
>> programming language.  Example:
>> 
>> # Original equation
>> 4x^2- 400x = 0
>> 
>> # add 100x to both sides and cancel
>> 4x^2= 400x
>> 
>> # Divide both sides by 4
>> x^2 = 100
>> 
>> # Solve for x
>> x=10
>> 
>> 4.  The substitution problem - and this was the big one.  When you get a
>> problem with lots of variables, like in a word problem, you have to write
>> down all your variables, do some manipulation, and then substitute
>> everything back in.  Example:
>> 
>> A box has a base whose length is 10 and width is 5.  The height of the box
>> is 2.  Calculate the surface area of the box.
>> 
>> So you write down
>> 
>> l=10
>> w=5
>> h=2
>> 
>> s = 2lw + 2lh + 2wh
>> 
>> Now you have to ssubstitute in the values, which means you havfe to move
>> your cursor back up, memorize one or more variables, and then bring your
>> cursor down, place on the variable to be substituted, type the value, delete
>> the original variable, and repeat without blowing anything.
>> 
>> A sighted student does this completely intuitively because he or she
>> rewrites the surface area formula, substituting variables on the fly by
>> referring to them visually.
>> 
>> When you Add more fariables, fractions, exponents, and signs, the tendency
>> to blow it goes way up.
>> 
>> After taking a 2 hour exam that actually took me 7 hours, of going back and
>> forth dizzily within long problems, I figured there had to be a better way.
>> 
>> Enter the the Computer Algebra System.
>> 
>> I figured a Computer Algebra System could solve many of these problems for
>> me.  It could store and substitute variables, solve for x without using that
>> horribly error-prone quadradic equation, never blow an exponent or +/- sign,
>> produce graphs suitable for a tactile embosser, and give me the ability to
>> double check my answers, to say nothing of being a very capable talking
>> "calculator."
>> 
>> I started off with Maple and found the workbook and the other Java-based
>> user interface marginally to completely inaccessible.
>> 
>> Then I discovered the command-line version called cmaple.  that was much
>> more accessible.  I used the "interface(prettyprint=0)" command to force the
>> exponents into the same line as the equation instead of Maple printing them
>> above the equation.  And at that point, I had a pretty good solution.
>> 
>> Unfortunately, Maple suffers from several problems which make its use by a
>> blind student problematic:
>> 
>> 1.  The two standard user-interfaces, workbook and Java-based, , are
>> marginally to totally inaccessible.  Use the command-line version for best
>> results.
>> 
>> 2.  The program is costly even at the student price of around $100 and
>> requires a fair bit of fussing around to procure as a student.
>> 
>> 3.  The installation program is a pig and requires sighted assistance and
>> many non-keyboard mouse clicks in certain places to complete.
>> 
>> 4.  And here's the worst part -- you are granted a single license when you
>> purchase so operating on your home desktop, your laptop, and your school
>> workstation is going to be a problem.
>> 
>> Out with Maple, in with Maxima:
>> 
>> Once I discovered the potential of a Computer Algebra System (CAS), I was
>> hooked.  I then discovered Maxima and Axiom, two open-source programs simlar
>> to Maple but without the cost, licensing, or installation issues.
>> 
>> I've been using Maxima ever since.
>> 
>> Next steps:
>> 
>> I envision integrating the input and output of Maxima with my text editor so
>> I can do my homework in one seamless environment, capable of placemarkers,
>> cut and paste, variable substitution, calulations, etc
>> 
>> I will be switching from Notepad to Edsharp as my editor of choice since
>> Edsharp is so much more capable and extensible.  Creating the glue between
>> Edsharp and Maxima will be my project for the winter break.
>> 
>> A Potential Downside for CAS Use by Students:
>> 
>> A CAS is so capable, it introduces not only the time-saving features
>> described above, but the ability to solve some problems without doing the
>> real math.  Fortunately, most instructors would spot this by noticing you
>> haven't shown your work.  In addition, I've found the hardest work in math
>> is the problem setup and interpretation and the CAS can't do that for you.
>> Still, there may be some reluctance on the part of instructors to allow you
>> such a powerful calculator / programming language.
>> 
>> Next semester...
>> 
>> 1.  An integrated editor / CAS
>> 2.  Calculus II
>> 3.  And in the future, maybe even Physics :~
>> 
>> I hope you all might find my experiments and hybrid solutions useful.  I'm
>> pretty sure Amanda and Dr. Baldwin will for sure.
>> 
>> Ben
>> 
>> 
>> 
>> 
>> _______________________________________________
>> Blindmath mailing list
>> Blindmath at nfbnet.org
>> http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
>> To unsubscribe, change your list options or get your account info for
>> Blindmath:
>> http://nfbnet.org/mailman/options/blindmath_nfbnet.org/ljmaher%40swbell.net
>> 
>> 
>> _______________________________________________
>> Blindmath mailing list
>> Blindmath at nfbnet.org
>> http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
>> To unsubscribe, change your list options or get your account info for Blindmath:
>> http://nfbnet.org/mailman/options/blindmath_nfbnet.org/brh%40opticinspiration.org
> 
> 
> _______________________________________________
> Blindmath mailing list
> Blindmath at nfbnet.org
> http://nfbnet.org/mailman/listinfo/blindmath_nfbnet.org
> To unsubscribe, change your list options or get your account info for Blindmath:
> http://nfbnet.org/mailman/options/blindmath_nfbnet.org/mwhapples%40aim.com