Discussion:
[tug-summer-of-code] Some Idea
Zachary Hoffman
2009-02-28 04:50:12 UTC
Permalink
Dear TeX Users Group,
Hey, I'm an undergraduate working towards mathematics/piano performance
double degrees and I'm interested in developing for TUG in Google's SOC
2009. I have an idea that you could add to your list of potential
proposals, though I plan to use it when applying this March.
In the past, Texas Instruments has released four very heavily used
graphing calculator models: the TI-83
Plus<http://en.wikipedia.org/wiki/TI-83_series>,
the TI-84 Plus, and silver editions of both, all of which use the Z80
processor to run slight variations of the same operating system. Because of
its large user base (millions), the 83+ has been a desirable platform for
all kinds of applications, especially math-related ones.
While PDF files outputted by some implementations of TeX allow many
people to witness its beauty, this filetype goes completely unsupported on
graphing calculators. If given the opportunity, I would write an
application to render source manuscripts on the TI-83+ and similar
calculators. I have yet to work out all of the specifics, but I have
determined that the project is very feasible despite time and hardware
constraints. I'd love to discuss this more between now and March 17 and I
look forward to seeing TUG on the list of mentoring organizations this
spring.

Best wishes,

Zachary Hoffman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://tug.org/pipermail/summer-of-code/attachments/20090227/cf273dd8/attachment.html
Jonathan Fine
2009-02-28 08:04:39 UTC
Permalink
Hello Zachary
Post by Zachary Hoffman
In the past, Texas Instruments has released four very heavily used
graphing calculator models: the TI-83 Plus
<http://en.wikipedia.org/wiki/TI-83_series>, the TI-84 Plus, and silver
editions of both, all of which use the Z80 processor to run slight
variations of the same operating system. Because of its large user base
(millions), the 83+ has been a desirable platform for all kinds of
applications, especially math-related ones.
Very interesting. The wikipedia page gives the technical specs:
CPU: Zilog Z80 - 6 to 15Mhz
ROM: 24Kb
Flash ROM: 164Kb - 2Mb
RAM: 32 - 128Kb
Display: 16x8 text, 96x64 pixels (character is 12x8 cell)

This is challenging environment for TeX. I started TeX (in 1987) on
CPU: 80286 - 10Mhz
RAM: 1Mb, later extended to 2Mb
Hard Disk: 40Mb
Post by Zachary Hoffman
While PDF files outputted by some implementations of TeX allow many
people to witness its beauty, this filetype goes completely unsupported
on graphing calculators. If given the opportunity, I would write an
application to render source manuscripts on the TI-83+ and similar
calculators. I have yet to work out all of the specifics, but I have
determined that the project is very feasible despite time and hardware
constraints.
I doubt that the calculator has enough memory to run a port of TeX, but
would be delighted to be proved wrong.

Please would you provide some links to some open source programs for
these calculators, that might be similar to what in mind?
Post by Zachary Hoffman
I'd love to discuss this more between now and March 17 and
I look forward to seeing TUG on the list of mentoring organizations this
spring.
Yes. I also hope we're accepted, and look forward to discussing with
you further.
--
Jonathan (mentor for MathTran JavaScript last year)
Martin Schröder
2009-03-04 19:56:27 UTC
Permalink
Post by Jonathan Fine
CPU: Zilog Z80 - 6 to 15Mhz
ROM: 24Kb
Flash ROM: 164Kb - 2Mb
RAM: 32 ?- 128Kb
Display: 16x8 text, 96x64 pixels (character is 12x8 cell)
[...]
Post by Jonathan Fine
I doubt that the calculator has enough memory to run a port of TeX, but
would be delighted to be proved wrong.
Do we know the specs of the first system DEK used for TEX82?

Best
Martin

Zachary Hoffman
2009-03-01 09:37:06 UTC
Permalink
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://tug.org/pipermail/summer-of-code/attachments/20090301/e17884cd/attachment.html
Zachary Hoffman
2009-03-02 14:15:05 UTC
Permalink
Hi, Jonathan

Thanks for the reply. Actually, how did you reply?
Post by Jonathan Fine
Please would you provide some links to some open source programs for
these calculators, that might be similar to what in mind?
Certainly. Graph?, a grapher of two-variable surfaces and
differential equations, is an application of great utility that brings
handheld calculators closer in functionality to more complex computer
algebra systems.
Graph?: http://www.ticalc.org/archives/files/fileinfo/247/24741.html
source: http://sourceforge.net/projects/graph3/

A Scheme interpreter for the 83+ was written in GSOC 2007. It shows
that parsing source files created on more powerful systems can and has
been done on this platform.
Summer of Code page:
http://code.google.com/soc/2007/detached/appinfo.html?csaid=9BDCF5C6DDBB195E
binary: http://group.revsoft.org/scheme.8xk
source: http://group.revsoft.org/scheme.zip


DAWG (Dictionary and Word Games) was developed by Detached Solutions,
which also made Graph?. Although it is freeware and not open source,
it is a good example of something with a fairly large database.
DAWG: http://www.ticalc.org/archives/files/fileinfo/258/25844.html
Detached Solutions: http://detachedsolutions.com/aboutus/
Post by Jonathan Fine
I doubt that the calculator has enough memory to run a port of TeX...
If it used a single bitmap font for the final rendering and only
supported styles for ASCII characters, then the size of glyph-related
data should be around 160 kilobytes.

The DVI file format is unoptimized for (complete) machine readability
by nature and takes away TeX-specific aspects of a document, which is
unnecessary here. Another option for document format is some
compressed form of source manuscripts, where commands are identified
by 16-bit numbers instead of their English representations.

Many macros could that are of no use on a calculator, such as those
that stray from monochrome rendering or embed active objects, need not
be included.

Displaying TeX typesets on a 96x64 LCD means that less of a document
is displayed at once, which would allow the device to render portions
of a document in less time while consuming fewer resources. Users
could still read the text with ease with a simple autoscrolling (along
text) function and could navigate the document in a view where the
width of the document is fitted to that of the screen.

All in all, how does this proposal sound? Thank you for giving it
consideration.

---
Zachary
Continue reading on narkive:
Loading...