Open s9105947 opened 2 years ago
To specify custom species, the species Type extension allows: "user are free to append a free text after a colon". Is really any text allowed? (What even is text?) IMO "free text" should be severely restricted, at least to "printable characters without semicolon (;)".
Correct. Derived from the base standard, we always mean pure ASCII when we write text.
Yes, generally we leave other:
unspecified for cases we have not yet thought about and that get standardized in later versions.
A semicolon is also fine in that case, maybe someone things of another compound or list and wants to use the same convention. When it gets standardized later one, we only drop the other:
, making file conversions easy. Thus,;
are fine here.
I would not pro-actively allow <a-type-that-we-already-define>;other:<someStuff>
yet, since I think we have no use case for this yet and it just complicates the syntax & conventions. (Unless you have a concrete case you need to achieve right now, of course.)
Thank you for the clarification on ASCII, I overlooked that ._.
I like the concept of reserving other:
for entirely custom types and forbidding it from lists
So if I understand you correctly, a speciesType
is:
;;
) are forbiddenother:
, followed by see belowCorrect, so far?
This would leave these questions:
other:
be permitted?other:
?I'd suggest: yes, no, class "print" of POSIX locale (IEEE 1003.1-2008, s. 7.3.1, l. 4187) 123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_abcdefghijklmnopqrstuvwxyz{|}~
(+backtick ` +space +horizontal tab)
To specify custom species, the species Type extension allows: "user are free to append a free text after a colon". Is really any text allowed? (What even is text?)
IMO "free text" should be severely restricted, at least to "printable characters without semicolon (;)". notable cases:
\\n
or\\a
By gut instinct I'd suggest to allow only "universally" safe and unambigous characters after "other:":