IMHO every such collection of Bash tips should include information on POSIX shell compatibility of each construct. Wait, has the author written a whole book on Bash? Does it ever mention the existence of other shells?
P.S.: It does. But not much more than that.
Running a disk defragmenter and a GW-BASIC number crunching script in parallel under DESQview taught me lots of things, e.g. repairing data from what'd been the FAT16 HDD partition before.
AFAICR, the feature one had to pay for was ROM containing PowerPC code. It made Apple-compatible PCI cards much more expensive than x86-compatible. Especially video.
I would really like to use TeXmacs for preparing linguistic publications, which may include multilingual text. TeXmacs does not seem to support Unicode combining diacritics. E.g. when I paste `бе́ля` in a TexMacs document, the accent is rendered as a full-width character between `е` and `л`, no matter which font I choose. Typing `⌘' e` (Latin `e`, not Cyrillic as above) renders `é`, but it does the job only for precomposed Unicode characters known to TeXmacs. OpenType `gpos` table has (e.g. in Times New Roman) a `mark` feature, specifying a "base" mark-to-base anchor point on Cyrillic vowels and a "mark" mark-to-base anchor point on U+0301 (combining acute accent), so the renderer can position the mark correctly above the preceding character. However, TeXmacs reimplements most of font rendering instead of reusing Harfbuzz which implements countless script-specific shaping quirks using OpenType tables. IMHO it's a Sisyphus' job.
SAS DATA steps create new datasets (table) from one or more input datasets. As far as I understand the author, Bistro creates just a new column in an existing table.