see also: How to adjust alignment within Math ($$)?
see also: discussion on xmmath
just i personally think the alternating right/left alignment, the need for && to keep same alignment, feels weird/dark/magic/chooseyouword.
here’s my worse idea, free to comment, though.
in short, play with spaces like in a "b" c"d"e:
& right&left & center &also center maybe bilateral or even intellegent^(TM)&
edge cases on the first/last one within a row:
right&left \ left & right
an extra leading/trailing & is mandatory for first/last column to align center. that’s about to build a table.
the convention of a &= b is a little bit hurt but a&= b just takes a deletion of the space. and it looks closer to the outcome.
of course, i agree that &s in cases() should work just like outside.
\s inside branches… that’s another storyi have some wild remark in the xmmath link, though, like, i’d like to have some higher-level-parenthesis so that
\s inside will behave like line-breaks inside an inline-block.
a bit off-topic about me exploring\and finding[lr(1 & 2 \ 3 & 4)]draws matrix in xmmath… what i wanna mention are stirling numbers (two kinds), although vec can do it, it isn’t semantically precise. i’d like to suggest a generalbinom(delim:"(",a,b)too.
if thinking even wilder… let (){}[] in math mode be essentially equivalent, that is, be available for function-calling when possible, ofc the “function” can detect on what parenthesis it’s being called, soover{} under[] cases{} mat[] binom() binom[] binom{} lr() lr[] lr{}would let us save a lot ofdelim:but what about#none(stop thinking
currently, cases(delim:#none) and lr(c o n t e n t) and & together nested just works fine (without control on alignment), but & inside binom seems not working at all (slap) yeah i admit it’s kinda meaningless..