[tex-k] bugs & suggestions in The METAFONTbook

胡亚捷 (Hu Yajie) 2500418497 at qq.com
Sat Jul 4 13:57:26 CEST 2020

Page C266, lines 19--20
"You can say either `incr x' or `incr (x)', within an expression; but `incr x'
by itself is not a valid statement." Well, neither is `incr(x)'.

Page C288, lines -13 and -6
Maybe the `Boolean' should not be capitalized, like the rest of the book?
(If so, also decapitalize the index entry.)

Page C290, two code fragments
The first code fragment ends with a period, but the second ends with a
semicolon, even though they're similar (two def...enddefs) and are both
followed by new sentences. I wonder if Knuth has any pattern in mind...

Page C295, line 2
`3.37' is apparently an approximate value, but the text doesn't say that.

Page C298, line 25
Best to put a space between the `makepen' and the `(', as in the rest
of the book.

Page C328, lines 18--26
Boldface the `special's?

Volume C, Index
    Change `295' to `298'.

    Add `337' (where it is defined).

`<internal quantity>'
    Move the `265' to the `internal quantities' entry, since the page
    does not use angle brackets when mentioning the concept.

    Delete the `275'.

    Add page 318, which mentions that <label> may appear in ligtable,
    charlist or extensible commands, and they are mutually exclusive.

`labels on {\it proofmode} output'
    There is no identifier called `proofmode', as far as I know. (You
    may also want to change page 228, line -3.)

    Change `315--317' to `315--321' (the whole Appendix F).

Volume C, In general
Is it possible to change the "..." macro so that italic correction is
not applied when it is followed by a subscript? Compare, for example,
the "theta"_1 and \penpos{1'} in the program on page 29.

Since Knuth said in the 2008 tune-up note that TeX and MF do not guard
against malicious input (http://tug.org/TUGboat/tb29-2/tb92knut.pdf), it
may be better to *drop* my arithmetic overflow bug reports. By the way,
do you consider the different ordering between `tracingequations' (first
deduced first) and `showdependencies' (last deduced first) a bug worth

Incidentally, Knuth may be interested in this Chinese `meta-font':


Previous reports:

More information about the tex-k mailing list.