The following are suggestions how to replace math books with hyperlinked modules, in general, and particularly suggestions how to support it in a future version of TeXmacs software. TeXmacs is an advanced math writing software, and the easiest way to implement my ideas are to modify this existing software, rather than writing a completely new software from scratch.
We want presenting math as hyperlinked modules published on the Web. In principle, it is possible to do with TeXmacs, as that software supports hyperlinks and exporting into HTML+MathML (or HTML with images) format. However, the current version of TeXmacs does not provide all features to do it effectively.
First I will describe what I want (irrespective of any particular software).
I want to split a math book (a collection of related math knowledge) into modules.
Every module should contain one theorem or one definition. I may also contain a proof or several proofs of the same statement.
Also, for a given theorem, its module should contain:
- (hyperlinked) list of theorems and definitions this theorem depends on
- (hyperlinked) list of its generalizations
- (hyperlinked) list of other theorems it generalizes
- (hyperlinked) list of counterexamples of non-possible generalizations (example: we cannot strengthen “x=y implies x/z=y/z for nonzero z” for zero z)
- (hyperlinked) list of its consequences (“consequence” probably should not be defined formally as implication, because otherwise every true theorem is a consequence of every other true theorem, but understood in a human way)
We also need the possibility to transform a set of modules into a traditional linearly ordered PDF (or HTML) book. We should have the option to exclude some of features when forming a book, as described in the following paragraph:
When converting into book form, we shall have the option to exclude all proofs. Not being too pedantic, we may allow some users to believe us without reading proofs.
Problems with existing TeXmacs version: We can implement this as modules being separate TeXmacs .tm files. Hyperlinks are already supported by TeXmacs. But we have no TeXmacs feature to transform a set of modules into a linear book. For this, we may make a “content” TeXmacs module, which would contain an ordered list of modules, to make a book of this order. We need the feature to export (say in HTML format, into a folder on a disk) all modules from the “content” file, without manually exporting the files one-by-one.
Most of the following can be done with TeXmacs macroses, without modifying its core. However, who knows, we may probably need changes in TeXmacs itself.
We also need to add to TeXmacs a mechanism to check for invalid forward references, for example when “list of theorems and definitions this theorem depends on” is circular.
We need hyperlinks ordering the list of modules linearly, for the case if somebody wants to read it completely. These links should be based on the above mentioned “content” file.
We also need to generate an index and a list of notations from all modules in a “book”
Also it may probably be sometimes useful to put several modules into one file. This issue should be further investigated.
Just a few minutes ago I conceived a definition of generalized Fréchet filters with definition for every poset on which filters are considered (however, I have not yet calculated the class of posets for which generalized Fréchet filter is defined; it should be easy but I am busy with other business).
Generalized Fréchet filter on a poset is a filter such that
See my book for a definition of .
I’ve done a little discovery today: Proximities are the same reflexive, symmetric, transitive funcoids.
For now I leave to prove this as an exercise for a reader. But later I am going to include this theorem into the book I am writing.
My article was accepted for publication in European Journal of Pure and Applied Mathematics, but it didn’t compile with their LaTeX templates. After waiting a reasonable time until they would tackle the problem, I have withdrawn my article and sent it to another journal.
I would search for the bug in their LaTeX template myself, but I was a great LaTeX expert in the past, I have forgotten much of LaTeX tricks, as now I use TeXmacs to write my manuscript, I convert them to simple LaTeX only to send it to a publisher. So I have chosen not to attempt to resolve the problem with LaTeX but just to send it to an other journal.
I’ve introduced another version of cross-composition of funcoids. This forms a category with star-morphisms. It is conjectured that this category is quasi-invertible, because I have failed to prove it.
This should be included in the next version of my book.