> Aditya Mahajan:
> The default should be \mathupright (as is the case in amstex
and latex, and probably also plaintex).

> Hans Hagen:
> next beta: when one of the stylealternatives text mode else mathmode
> (default nothing == math upright)

> Hans Hagen:
>> Because like in \mtext, the \text command is ESSENTIAL (for the function names to have the mathcodes, kerning and ligatures of the text mode instead of the ones of the math mode).
>that is not needed because if you use a style known as text style a text font is used

Alright, but why not ALWAYS use your text style (or my \text), why talking about \mathupright then? \mathupright should never be used to display a name in text style in my opinion. I don't get it, is there something that I missed with that \mathupright? (And I'm not managing to download the new beta of yesterday so I don't have access to the new code yet).


Appart from that story of \text or not \text, what do you think about the suggestion that I gave in the conclusion of my previous post? Do you plan to still use version 1 or to implement version 2 after the freeze for TeXLive 2015? My conclusion was:

CONCLUSION. — It all comes down to the choice of the best user interface for math functions and math texts (usually used as subscripts of variables). Their styles and colors is handled with \setupmathematics. There are two choices of user interface:
1. $\normalmathop{\mfunction{Arsinh}}\nolimits x\ \text{and}\ n_{\mtext{air}}$.
2. $\mfunction{Arsinh} x\ \text{and}\ n_{\mtext{air}}$.

First, the default functionstyle and textstyle in \setupmathematics should be \rm\tf (the text mode equivalent of \mathupright, as we are always in text mode thanks to the \text hardcoded in \mfunction and \mtext).

Then, in the 1st solution \normalmathop is used instead of \mathop because the latter is modified by ConTeXt to convert \rm to \mf, which makes \setupmathematics[functionstyle=\rm] without effect.

The 2nd solution is obviously way cleaner, that is why I think it should be preferred. The implementation is given in my '2nd DEFINITION of \mfunction' for replacing the existing definition in math-ini.mkiv:
———————————————————————
\unexpanded\def\mfunction{\dosingleempty\domfunction}
\def\domfunction[#1]#2%
  {\iffirstargument
     \doifelse{#1}{op}{\normalmathop{\text{\usemathematicsstyleandcolor\c!functionstyle\c!functioncolor#2}}}{\doifelse{#1}{limop}{\normalmathop{\text{\usemathematicsstyleandcolor\c!functionstyle\c!functioncolor#2}}}{\normalmathop{\text{\usemathematicsstyleandcolor\c!functionstyle\c!functioncolor#2}}\nolimits}}
   \else
     \normalmathop{\text{\usemathematicsstyleandcolor\c!functionstyle\c!functioncolor#2}}\nolimits
   \fi}
———————————————————————

We also saw that \mfunctionlabeltext is inconsistent with ConTeXt \labeltext because the former has a style attached to it, so \mfunctionlabeltext should be removed from math-ini.mkiv and \mfunction{\mathlabeltext } should be used instead when needed.

To go even further in userfriendlyness, we the new macro \definemathcommand should be added in math-ini.mkiv:
———————————————————————
\unexpanded\def\definemathfunction{\dodoubleempty\dodefinemathfunction}
\def\dodefinemathfunction[#1][#2]#3%
  {\ifsecondargument
     \definemathcommand[#1]{\mfunction[#2]{#3}}
   \else
     \definemathcommand[#1]{\mfunction{#3}}
   \fi}
———————————————————————

Finally, with this new \mfunction definition and the removal of the misleading \mfunctionlabeltext, the functions defined in math-def.mkiv should be rewrite as:
———————————————————————
\definemathcommand[cos]{\mfunction{\mathlabeltext{cos}}} % Or \definemathcommand[cos]{\mfunction[nolop]{\mathlabeltext{cos}}}, it is the same.
\definemathcommand[det]{\mfunction[limop]{\mathlabeltext{det}}}
\definemathcommand[diff]{\normalmathop{\text{\mathfunctionstyle d}}\mathopen{}}
%etc.
———————————————————————

or
———————————————————————
\definemathfunction[cos]{\mathlabeltext{cos}} % Or \definemathfunction[cos][nolop]{\mathlabeltext{cos}}, it is the same.
\definemathfunction[det][limop]{\mathlabeltext{det}}
\definemathcommand[diff]{\normalmathop{\text{\mathfunctionstyle d}}\mathopen{}} % It is a special function so it should still be treated with the more general \definemathcommand.
%etc.
———————————————————————

ConTeXt users are now provided, if Hans agrees to implement the code given in this conclusion, with three extremely useful macros:
— \mtext (that should be used for all math texts such as subscripts of variables):
  $n_{\mtext{air}}$;
— \mfunction (the equivalent of LaTeX \operatorname macro of the amsopn package):
  $\mfunction{Arsinh} x$
  $\mfunction[op]{Arsinh} x$ or equivalently $\mfunction[limop]{Arsinh} x$;
— \definemathfunction (the equivalent of LaTeX \DeclareMathOperator macro of the amsopn package):
  \definemathfunction[arsinh]{Arsinh} $\arsinh x$
  \definemathfunction[arsinh][op]{Arsinh} $\arsinh x$ or equivalently \definemathfunction[arsinh][limop]{Arsinh} $\arsinh x$.


Maggyero