354700+ entries in 0.229s
phf: the only redeeming quality in
this case,
that
there aren't better standards available, except for maybe Ada (i've not looked). you end up with
trade offs, or hard constraints
that make
the result non-portable. a positive aspect of cl standard is
that you get reasonably similar behavior on very wide range of hardware, but i agree
that
things could always be better
phf: mircea_popescu:
there are parts of standard
that are underspecified :)
BingoBoingo: <asciilifeform> and it is ~wrong~
to IMPLICITLY arithmetize on non-arithmetic objects. << What would
Terry Davis say about
this?
mircea_popescu: phf and if idiots make functions going ___function where in
the ordered list do i look for
them ?
phf: mircea_popescu: wait, partial ordering in
this case means
that alphabet is ordered. it's meaningless
things like "is _ greater
then a"
that are not defined
mircea_popescu: and no, i won't be looking for
the beatles records past
the why records.
mircea_popescu: the fucking point of having an alphabet is
that no, it couldn't move.
mircea_popescu: i can also
trivially define
the new location of washington dc.
mircea_popescu: and i should not be surprised, lest
the scribe gets her ass colored.
mircea_popescu: nevertheless, sorting lists of words is
the most basic, primary, entrly level
task for any scribe.
mircea_popescu: so if i sort a file containing a bunch of words, i do not know if i should look for ___stuff at
the end or beginning or scattered anywhere.
mircea_popescu: "Of
the standard characters,
those which are alphanumeric obey
the following partial ordering:"
mircea_popescu: asciilifeform so is
there a special function
to inquire which of
two comes before in
the alphabet ?
phf: hehe, where do people normally get
this shit from???
mircea_popescu: no, in
the sense
that
they are specifically labelled
thus with a label specifically made
to label numbers.
mircea_popescu: so apparently characters are introduced by ... an escape sequence (lol)... which happens
to be... NUMBER SIGN ?
mircea_popescu: "As an example,
table 13-1 shows
the character labels, graphic symbols, and character descriptions for all of
the characters in
the repertoire standard-char except for #\Space and #\Newline."
mircea_popescu gives up in frustration. if anyone recalls
the damned
thing say.
☟︎ phf: asciilifeform: i
think numerics is a
truly dodgy
tradeoff. it
took me until quite recently
to grok
that "character code == ascii" or "character code == unicode codepoint" are not
the only
two possible interpretations.
mircea_popescu: by jumping
the index, smashing stacks and doing all sorts of shenanigans.
mircea_popescu: the idea was
that it sacrificed outlier cases
to get a better performance curve geared
towards more common cases
mircea_popescu: the bit discussing
the cheating involved in
the asm
translation of search in c was added
to
the log about a year ago
mircea_popescu: no,
this is not how search works "anywhere" in
the universe.
mircea_popescu: dude if you imagine it is acceptable for me
to sit here looking at
two strings AND NOT KNOW exactly what
the machine will do if i
tell it
to search
mircea_popescu: "The implementation may choose
to search sequence-2 in any order;
there is no guarantee on
the number of
times
the
test is made."
mircea_popescu: may be false rather
than
true, if
the value of z is a character."
mircea_popescu: "In general, characters in Common Lisp are not
true objects; eq cannot be counted upon
to operate on
them reliably. In particular, it is possible
that
the expression
phf: include any number of extra characters, say code 120 #\lambda, and also decide how it wants
to present
them
to console/file (say, by writing unicode, or letting user choose
the encoding). so a maxima machine can have a set of mathematical symbols for
there and
then, but not with any sort of claim
to univesality.
phf: apropos i like how cl does it,
there's a set of characters
that are predefined
to be standard and portable, and it's basically an mp set: alphabet, digits,
top row, handful of controls like newline. each character has a numeric code and a verbose name #\a #\b #\space #\newline.
there's a handful of predicates
that let you query characters basic nature, upcasep/downcasep/whitespacep.
that's it. an implementation
though can choose
to
trinque: asciilifeform │ and (λ (x y) (....)) ~is~ superior
to (lambda... << with
this I have no problem at all.
mircea_popescu: eh, life managed
to extinguish itself on its very own well enough.
mircea_popescu: they did all sorts of stupid shit, but again,
the whole "a wizard did it,
to a vectorizer, logically" is enough
to sink
that ship.