Worthwhileness of mixing new language-level keyword into classic arguments comma-separated list is at least questionable. It maybe make sense to consider following expanded form to _coexist_ with current shorthand: Shorthand (classical argument list): attr(bgcolor, color, black) Expanded (JSON-like associative array): attr( name: bgcolor, type: color, fallback: black ) In shorthand form, arguments are listed in algoristic order. In expanded form, order can be arbitrary which make this serve to be mnemonically friendly as your purpose is: attr( fallback: black, type: color, name: bgcolor ) 16.09.2011, 03:07, "fantasai" <fantasai.lists@inkedblade.net>: > Currently the syntax of attr() is > šššattr(<ident>, <type>, <fallback>) > e.g. > šššattr(bgcolor, color, black) > > I get rather confused with positional syntaxes, because I can never remember which > argument goes first, second, etc. So I've been pondering alternatives and came up > with > šššattr(<ident> as <type>, <fallback>) > e.g. > šššattr(bgcolor as color, black) > > which makes it less positional -- the type is more closely tied to the attribute > name than to the fallback, and then the comma behaves to separate alternatives > like it does in font-family. > > Thoughts? > > ~fantasaiReceived on Friday, 16 September 2011 00:29:36 UTC
This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:04 UTC