From: Karlsson Kent - keka [[email protected]]
Sent: den 27 oktober 1999 13:29
To: Multiple Recipients of Unicore
Subject: RE: Math alphabets
 

Hi!

 (You should read the interpreted HTML version of this message, due to two tables below.  I know not all of you like that for various reason, but I've tried to up the point size at least, so that it is readable without a magnifier...)

> ME
> >>Style markup is _already_ used to select fonts, such as
> bold and italic
> >>fonts. Boldness and italicness can also be generated
> algorithmically on
> >>many WYSIWYG systems if special fonts are available. If you
> choose to add
> >>additional style markup designed to select fonts the
> logical conclusion is
> >>that such style markup (fraktur, double-strike, script)
> could be applied to
> >>any font, which would be dreadful and impossible.
>
> KK
> >Again, nobody has asked for that. And I'm PDS nobody has
> asked for that!
>
> It is the logical conclusion of saying that this stuff can't
> be handled by
> UCS plain text for character searching (different from formula layout
> perhaps)

I don't understand that sentence.  And I don't see how your conclusion is the consequence of anything put forward as an argument in this debate.

> >You might get some insight into what I mean by catching up
> on LaTeX2e.  Do
> >you think that \frac somehow algorithmically generates
> fractur letters from
> >an arbitrary font?
>
> No, I suppose I wouldn't. But are you suggesting that two
> DIFFERENT kinds
> of markup should be used to make <bold>a</bold> and
> <fraktur>a</fraktur>?

I'm not sure what you mean.  Of course there has to be SOME difference, otherwise how do we ever tell the difference, since a difference is intended.

(You'll need to look at the HTML version of this e-mail, to see the tables correctly.)

Let's review what's available in LaTeX2e and MathML 1:

 

LaTeX2e

MathML 1 ('presentation')

italic single letter identifier shortcut

i

N/A (use <mi>i</mi>)

italic identifier

\mathit{id}

<mi>id</mi>

upright identifier

\mathrm{id}

<mi fontstyle="normal">id</mi>

bold identifier

\mathbf{id}

<mi fontweight="bold">id</mi>

'calligraphic'/script identifier

\mathcal{id}

<mi>&iscr;&dscr;</mi>

doublestruck identifier

\Bbb{ID}

<mi>&Iopf;&Dopf;</mi>

fractur identifier

\frac{id}

<mi>&ifr;&dfr;</mi>

ordinary symbol

+

<mo>+</mo>

bold symbol

\boldsym{+}

<mo fontweight="bold">+</mo>

upright digits

0123

<mn>0123</mn>

italic digits

\mathit{0123}

<mn fontstyle="italic">0123</mn>

bold digits

\mathbf{0123}

<mn fontweight="bold">0123</mn>


Suggestion (round 0) for a hypothetical MathML 2 and for "my plain text math" (not intended to be complete, but sufficient for illustration here):

 

p.t.m.à la me

MathML 2 (à la me) ('presentation')

italic single letter identifier shortcut

Ii    (or i)

N/A (use <mi>i</mi>)

italic identifier

I{id}

<mi>id</mi>

upright identifier

R{id}

<mr>id</mr>

bold identifier

B{id}

<mb>id</mb> ((or <mi kind="bold">id</mi>))

'calligraphic'/'script' identifier

C{id}

<mc>id</mc>

doublestruck/'bbb' identifier

D{ID}

<md>ID</md>

fractur identifier

F{id}

<mf>id</mf> ((or <mi kind="fractur">id</mi>))

ordinary symbol

+

<mo>+</mo>

bold symbol

B+

<ms>+</ms> (or <mo kind="bold">+</mo>)

upright digits

0123

<mn>0123</mn>

italic digits

I{0123}

<mn kind="italic">0123</mn> (or <mni>0123</mni>)

bold digits

B{0123}

<mn kind="bold">0123</mn> (or <mnb>0123</mnb>)

Note that this distinguishing markup is right where the identifier is, not somewhere far out in the surroundings.

For p.t.m.à la me, I, R, B, C, D, F, {, and } are new "control" characters for p.t.m., while i and d are themselves (as are the I and D within {} above). { and } (as new control characters) would be used for p.t.m. grouping in general.  Note that p.t.m. needs a number of "control" characters anyway in order to convey super/subscripting, over, etc. etc.  See any book on LaTeX for inspiration on which math expression structure/display operators to make into "control characters".  Note that this involves a lot of syntactic constraints, e.g. {} balancing, what can follow I, that ^^ (superscript control character twice in a row) is an error (maybe recovered in some way), etc. etc.  The I, B, D, etc. need not be "combining" characters, anymore than ^ (p.t.m. superscript 'control' character) is 'combining' in the Unicode sense.

B.t.w. I DON'T see a 10-fold increase in verbosity for "MathML 2", relative to MathML 1, here. (Though that could easily be added.  ;-)

> >We appear to live in completely different worlds here.  I'm
> quite used to
> >using both eqn (in its days) and LaTeX, and I'm a bit
> puzzled as to your
> >arguments.  When you have caught up on LaTeX, and written a few math
> >formulas yourself, and maybe looked a bit at MathML too,
> maybe you could
> >come back and enlighten all of us.
>
> I couldn't write a math formula to save my own life. :-) I
> just make fonts
> and am thinking about architectural issues.

I’m also thinking about architectural issues, and I find the “math alphabets” proposal to be horribly ill-conceived and shortsighted.

> >Searching, e.g., for <mb>a</mb> is no more difficult than
> searching for
> ><mi>&bolda;</mi> (or its equivalent).  The same goes for the
> math markup
> >for-the-time-being-known-as "plain text math".
>
> I disagree. Quark XPress has a facility for searching styled
> text,

So does MS Word, for instance.

> but most
> software does not, and there are many different kinds of
> markup used even
> today. The internet sure doesn't support this; search engines
> IGNORE markup.

In order to find e.g. an “upright” math “cos” in MathML documents TODAY, the markup must be considered.  Why should ‘fractur’ identifiers be any different in that regard?

        Kind regards
        /kent k

 

> --
> Michael Everson * Everson Gunn Teoranta * http://www.indigo.ie/egt
> 15 Port Chaeimhghein Íochtarach; Baile Átha Cliath 2; Éire/Ireland
> Guthán: +353 1 478 2597 ** Facsa: +353 1 478 2597 (by arrangement)
> 27 Páirc an Fhéithlinn;  Baile an Bhóthair;  Co. Átha Cliath; Éire