8. Something about writing informal specs.
I'm curious to know what you would put in such a section.
I definitely find it difficult to write correct informal spec of a true routine. Either it is
imprecise as in this page:
or it is rather difficult to know how to use it in practice. For instance as in:
But it is maybe due to the conciseness of the documentation. (And these ones are rather formal.)
Definitely I think that the use of what they now call "ghost variables" or "model variables"
And there is still a lack of accurate and powerful tools. Notably to write easily true formal,
mathematical computer scince object in a language like TeX
(they have the property of being longer to write than the mathematical ones and it is a pain
to write them of change them).
But maybe you want to speak of the algorithm itself and the involved data structures.
Not its implementation.