Mathematical typewriter emulation
This page is under construction.
What is MTE?
- a system for using computers to emulate a mathematical
typewriter having about two keyboards worth of symbols
- easy to learn
- quick for producing typewriter-quality mathematics
- appropriate for informal mathematical text: e.g. tests
- (probably) formally realizable as an SGML implementation
based on a large collection of entities corresponding to
characters and symbols and a few elements for primitive
- But MTE should not be viewed as a structured markup language
for mathematical text. The significance of formal realizability
under SGML is that it would lend itself thereby to SGML-based
automatic processing. For example, there would appear to be a
good prospect for automatic translation of MTE to HTML-3.
I have collected some pointers to resources on
It is NOT:
- a substitute for TeX
- a narrow dialect of TeX
- automatically translatable to or from TeX
- have application to a specialized dialect of HTML hypertext.
What would be required for this? A specialized dialect viewer with
- The ability to display many characters in many type styles.
- The ability to create subscripts and superscripts.
This would be essentially different than using embedded graphic
images of special character displays on HTML pages in that the entire
item served would admit filtration through special client-side programs
such as readers for those with vision impairment.
It would involve substantially less network traffic and easier
client-side processing than fully typeset formats such as TeX/DVI and
MTE stands for "mathematical typewriter emulation". It is
capable of doing what a mathematical typewriter (circa 1965) could do,
no more no less. Materials can be prepared using MTE in roughly the
amount of time required for creation on a mathematical
typewriter (but less the time spent correcting errors on paper).
MTE files need to be filtered prior to printing. The filter needs
to perform character-to-string substitutions (see my utility
CONV) for conversion of individual "high ASCII"
characters and certain "control characters" and to change a few
ANSI escape sequences to the corresponding printer commands.
The demonstration items below were produced by converting screen
images in which each character is represented by an 8 X 8 bitmap
to GIF format. The GIF images are slightly grainy. The appearance of
printed images (using matching fixed width fonts) depends on
the quality and resolution of the printer fonts.
The FIRST DEMO consists of about 25 lines of random mathematical
statements. What is offered is:
- a graphic image
of a micro-computer screen. The GIF image is slightly grainy due to
enlargement. It was rendered on the micro-computer using an MTE viewer
- interprets four hex codes as sub/super-script commands.
- functions in character blending mode.
- interprets ASCII codes in the range 160-255 as characters in the
installed screen font.
- respects a few ANSI escape sequences (e.g. boldface, underline)
- the raw MTE file,
- which may be safely though not intelligibly viewed with
Gnu Emacs. One wants an editor that
- recognizes "high ASCII" codes as characters in a specified font.
- allows configuration to render a few control codes (in the range
128-159) for sub/super-scripts with special symbols.
- provides a macro facility.
- enables the user to spawn an MTE viewer on the current buffer at
the current location.
The SECOND DEMO is a review sheet
for part of third semester calculus. Nothing useful will be obtained
unless auto-loading of images (GIF) is turned on. The total size of
the three images is about 18 kilobytes.
MTE is useful only within a community where there is agreement on
the set of available symbols and on the correspondence between the
set of symbols and the set of computer codes that represent them.
This is an ASCII symbol chart (rows 0, 1, 8 and 9 have been left
empty). In general, the symbol one sees in a given row and column
depends very heavily on the configuration of one's terminal. However,
where the Roman alphabet may be assumed, usually the assignments in
rows 2 through 7 follow the ASCII standard.
It is obvious that the ordering of symbols in this chart for high
codes is idiosyncratic and is not being proposed as a standard.
- Graphic image of chart
- made from the MTE viewer's rendering of a
- MTE file
- that contains one instance of each character.
Road closed beyond this point. Enter at your own risk.
The key software item is the MTE reader called "efr" that I wrote
for the Amiga desktop platform in 1989.
It would be great if someone would volunteer to port this
reader to X11. It should not be that difficult for someone familiar
with X11 calls.
Here is a primitive temporary link to a directory
containing zoo archives of the source
and the Amiga user package
along with text files listing the zoo archive contents.