Re: menclose: several values in the "notation" attribute

    For the moment, I treat "radical", "madruwb" & "longdiv" notations 
as special cases:
- In combination with other notations, only the "special" notation is 
displayed.
- When several "special" notations are used, an arbitrary priority order 
allows to determine which one to display.

    This behavior gives the same result as the test image, but since you 
said it's a bug I suppose I need to change that. Even if I agree 
overlapping things like radical+madruwb+circle is not really relevant, I 
think the MathML spec should says something about these combinations. 
The simplier is to indicate that notations always overlap.

Fred

PS: FYI, the current version of my patch allows Firefox to deal with all 
the MathML 2 notation values.
> I'm going through old email and noticed I hadn't replied to this.  
> Sorry for the delay.
>
> There is no reason that radical should be special.  It is a bug in the 
> test image (and in MathPlayer).  On the other hand, it is hard to 
> imagine a case where the radical and some other enclose effect want to 
> overlap -- usually one would be nested inside the other (and hence, 
> nested mencloses would be used).
>
>     Neil
>
>
> On Thu, Oct 30, 2008 at 8:56 AM, <fred.wang@free.fr 
> <mailto:fred.wang@free.fr>> wrote:
>
>     Thanks for all your anwsers. In what I did for Firefox, the
>     notations overlap so
>     I agree with this option. However, as Bruce indicated, some
>     combinations are not
>     clear: for instance in the MathML testsuite, there is a notation
>     "radical
>     circle"[1]. What are the rules you used to establish that only a
>     radical should
>     be drawn (incompatibility between notations, priority, order of
>     values...)?
>
>     [1]
>     http://www.w3.org/Math/testsuite/build/main/Presentation/GeneralLayout/menclose/menclose1-full.xhtml
>
>     > I'm very glad to hear you are improving Firefox's support.  The
>     answer to
>     > your question is that they do *not* nest.  Use nested mencloses
>     to achieve
>     > that.  The order is not important.
>     >
>     > Neil Soiffer
>     > Senior Scientist
>     > Design Science, Inc.
>     > www.dessci.com <http://www.dessci.com>
>     > ~ Makers of Equation Editor, MathType, MathPlayer and MathFlow ~
>     >
>     >
>     >
>     > On Thu, Oct 30, 2008 at 5:38 AM, Bruce Miller
>     <bruce.miller@nist.gov <mailto:bruce.miller@nist.gov>> wrote:
>     >
>     > >
>     > > fred.wang@free.fr <mailto:fred.wang@free.fr> wrote:
>     > >
>     > >> Hi all,
>     > >>
>     > >> I'm trying to add support for <menclose/> in Firefox and I
>     wonder what
>     > >> should be
>     > >> displayed when several values are given to the "notation"
>     attribute. What
>     > >> I
>     > >> think is that order and repetition of the values don't
>     matter: "box circle
>     > >> box"
>     > >> is the same as "box circle" or "circle box". Is it right?
>     > >>
>     > >
>     > > Interesting question
>     > > [... that the WG should sort out for MathML 3 :> ]
>     > >
>     > > My first inclination would be to think of them as
>     > > nesting (a box around a circle around the content,
>     > > versus a circle around a box around the content),
>     > > simply because it is hard to see (and harder to specify)
>     > > how many of the various pairs should overlap
>     > > (eg. how should "radical roundedbox" look?)
>     > >
>     > > OTOH, I suspect the original intention was that
>     > > they would overlap (in which case the order doesn't
>     > > matter), since that can create many more effects
>     > > than we would want to list as explicit values.
>     > >
>     > > Further, nesting can easily, and more clearly,
>     > > be done simply by nesting <menclose>'s.
>     > >
>     > > If this is the concensus of the WG, we at least
>     > > should clarify this.
>     > >
>     > > speaking for myself,
>     > > bruce
>     > >
>     > > --
>     > > bruce.miller@nist.gov <mailto:bruce.miller@nist.gov>
>     > > http://math.nist.gov/~BMiller/
>     <http://math.nist.gov/%7EBMiller/> <http://math.nist.gov/%7EBMiller/>
>     > >
>     > >
>     >
>
>
>

Received on Friday, 14 November 2008 10:12:41 UTC