Re: Meaning of S-shaped bag delimiters

On 1/8/2026 7:09 AM, Ivan Panchenko via Unicode wrote:
> It is sometimes assumed that the S-shaped bag delimiters (U+27C5 and
> U+27C6) are for multisets (because multisets are also called “bags”).
> However, looking at the attestation shown in N2680, they appear to be
> used around a well-formed formula rather than for a multiset. This
> suggests that “bag” simply describes their shape, though the
> description “The second pair of proposed characters is a set of
> s-shaped delimiters used as bag delimiters” in the proposal does make
> it sound as though they were multiset delimiters.
>
> https://www.unicode.org/L2/L2003/03410-n2680-add-math-chars.pdf
>
> Given that the referenced document is no longer available, does anyone
> know what these characters mean?
>
"We", that is the "math" subgroup of the PAG did some digging and came 
up with both an archival location for the missing document as well as 
citations that confirm usage in the context of multiset.

        https://web.archive.org/web/20041109101630if_/http://www.ma.utexas.edu:80/users/jcorneli/h/Xi.ps

        the quoted text is section 35 page 14 of this version

And

        the \lbag and \rbag symbols were in the 1994 version of the fonts

        https://web.archive.org/web/20010110005900/http://www.tex.ac.uk/tex-archive/fonts/stmaryrd/stmaryrd.dtx

        |% \title{The St Mary's Road symbol font} % \author{Jeremy
        Gibbons \and Alan Jeffrey} % \date{Version 2, March 1994} % %
        \maketitle % % \section{Introduction} % % This is a brief guide
        to the St Mary's Road symbol font, a new symbol % font for \TeX\
        and \LaTeX. It is designed to live with the American %
        Mathematical Society's fonts, contained in {\tt amssymb.sty}. %
        % It provides a number of new symbols, including ones for
        derivation of % functional programming (such as $\varcurlyvee$,
        $\moo$ and % $\merge$), process algebra ($\llfloor$, $\oblong$
        and % $\lightning$), domain theory ($\bigsqcap$), linear logic %
        ($\binampersand$ and $\bindnasrepma$), multisets ($\Lbag x
        \Rbag$, % $\nplus$, and $\subsetpluseq$) and many more. |

as a conclusion we think there's nothing that needs to change.

A./

PS: I'm merely acting as reporter.