Re: STIX fonts for clients

Hi Paul,

Mozilla Firefox is the only browser with good native MathML support.
Version 3 is going to be released soon and one have to solve in no
time the problem of delivering the STIX fonts. I am suggesting a
solution, which is simple and quick. Opera is moving also in this
direction, but it is a way behind, which gives a lot of time to find a
more complicated and general solution.
Cheers, Samy


On May 26, 5:52 pm, "Paul Topping" <[email protected]> wrote:
> Yes, I understand what you propose. However, this may create yet another font mess. It has long (~10 years) that the STIX fonts will become the math fonts that all will use. I think it would be far better for the FF/Mozilla community to write a proper font installer (it can't be that hard, we did it) and place pressure on the OS makers to clean house with respect to the difficulty of doing so. Perhaps they can even help with the project.
>
> Paul
>
> > -----Original Message-----
> > From: [email protected]
> > [mailto:[email protected]] On
> > Behalf Of [email protected]
> > Sent: Monday, May 26, 2008 5:16 AM
> > To: [email protected]
> > Subject: Re: STIX fonts for clients
>
> > On May 24, 7:09 pm, "Paul Topping" <[email protected]> wrote:
> > > Although we don't have any side in this argument, I feel the need to
> > > correct a misconception. While our MathPlayer plugin for Internet
> > > Explorer does install its own fonts, it installs them into Windows
> > > rather than keeping them private. In other words, its installer
> > includes
> > > a font installer. The same fonts are also installed with our MathType
> > > product on both Mac and Windows and are installed into the system and
> > > are available to all applications running on the machine.
>
> > > I can say from experience that installing fonts on both Mac and
> > Windows
> > > is not trivial. There are a number of issues on both platforms. I
> > can't
> > > say anything about Linux, etc. except to say that I've heard it is
> > > difficult there also. The OS vendors do not make it easy. They seem to
> > > brush a number of issues under the rug:
>
> > > - They allow fonts to be installed in multiple directories but leave
> > it
> > > to developers to figure out what that means. If fonts with the same
> > > names are in more than one, which takes precedence? If your installer
> > > wants to delete old versions of a font and replace it with a new one,
> > > how does it find all the versions?
>
> > > - There is not much attention to versions in fonts. As with any
> > > software, fonts are improved as time goes by. The various fonts
> > > installers do contains versions but some contain more than one, so
> > which
> > > takes precedence. The answer is usually none of them. Installers
> > > routinely ignore versions. OS documentation does not make it clear
> > what
> > > is supposed to happen.
>
> > > Just thought you might want to know.
>
> > > Paul Topping
> > > Design Science, Inc.www.dessci.com
>
> > > > -----Original Message-----
> > > > From: [email protected]
>
> > > [mailto:dev-tech-mathml-
>
> > > > [email protected]] On Behalf Of [email protected]
> > > > Sent: Saturday, May 24, 2008 4:33 AM
> > > > To: [email protected]
> > > > Subject: STIX fonts for clients
>
> > > > With the release of Mozilla's Firefox3-Candidate1, the release of
> > the
> > > > stable version is expected soon. I belong to the academic community
> > > > that seeks the use of the native Firefox MathML support. There is a
> > > > long debate between people with the same interest as mine and the
> > > > developers of the Firefox browser, about delivering or not the STIX
> > > > fonts to clients using MathML as one package together with the
> > > > browser's download. References for this debate can be found in:
> > > > 1. The bug report 295193 (
> > > >https://bugzilla.mozilla.org/show_bug.cgi?id=295193
> > > > ).
> > > > 2. The bug report 415223 (
> > > >https://bugzilla.mozilla.org/show_bug.cgi?id=415223
> > > > ).
> > > > 3. In the present list (
>
> >http://groups.google.com/group/mozilla.dev.tech.mathml/browse_thread/thr
>
> > > > ead/27000c43ddc95c12
> > > > ).
> > > > My argument is that there is a simple solution, which can satisfy
> > both
> > > > sides, and bring benefit to the spread of MathML.
>
> > > > I am trying to summarize the view for delivering the STIX fonts
> > > > together with the browser's download:
>
> > > > The academic clients are not developers. They need the computer as a
> > > > mean of helping them to meet the challenges of research and
> > teaching.
> > > > They need something simple that works, without following
> > instructions
> > > > that they don't understand, no matter how clear they are. A simple
> > > > example is the use of MathML by Firefox2. The Window's installer of
> > > > the fonts for MathML was straight-forward even for most of the
> > > > academic clients. For Mac the instructions needed the help of a
> > > > computer expert, and even he was warned about avoiding a bug that
> > was
> > > > explained elsewhere. For Linux it was catastrophic, even an expert
> > > > could not always help. On top a person with the root password should
> > > > also be called.  If Windows is the dominant platform, is it worth
> > > > bothering with the others?
>
> > > > What was the result? Only Window's clients (and they are not too
> > many
> > > > in the academic community) could benefit. Mac and Linux clients had
> > to
> > > > give up. On the other hand the main purpose of MathML is to be used
> > > > for mathematical expressions, and not too many people are using
> > them,
> > > > with the exception of some academic staff.
>
> > > > Now let's see what are the opposing arguments of the browser's
> > > > developers, and they make sense:
>
> > > > If the browser delivers also the necessary fonts, the downloading
> > time
> > > > of the browser will increase substantially, discouraging that way
> > > > people of using the Firefox browser. This argument is probably the
> > > > most easy to overcome, by presenting to the users the choice of
> > adding
> > > > the fonts during the downloading of the browser. Furthermore the
> > space
> > > > of all the STIX fonts (not zipped) take is less than 1.2 MB , which
> > is
> > > > less than 16% of the space needed for downloading the browser today
> > > > (7.5 MB). So far some of the browser's developers accept this
> > > > solution.
>
> > > > The next argument is much heavier, but in my opinion there is a
> > work-
> > > > around. The argument says that downloading the fonts is not enough,
> > > > one have to install them. The installation procedure differs from
> > > > platform to platform and it is a very heavy task to be done by a
> > > > browser. I agree, but why install the fonts in the system? Just put
> > > > them somewhere in a directory of the browser's software. They don't
> > > > have to be installed in the system. The browser should take care to
> > > > look for these fonts in the appropriate directory, when a <math ..>
> > > > element is encountered, instead of looking among the installed fonts
> > > > in the system (as for now). If a developer wants to use the STIX
> > fonts
> > > > for another than math purpose, he has to install them in the system
> > > > (in addition to the same - existing in the browser).
>
> > > > As an example of this work-around, the MathPlayer plug-in of Design
> > > > Science for Internet Explorer uses math fonts, which are included in
> > > > the plug-in itself. It does not require installing the fonts.
>
> > > > Hope that the just proposed solution can be easily implemented, and
> > > > we'll witness soon Firefox3 free of installing separately the STIX
> > > > fonts.
> > > > _______________________________________________
> > > > dev-tech-mathml mailing list
> > > > [email protected]
> > > >https://lists.mozilla.org/listinfo/dev-tech-mathml
>
> > Thanks Paul Toppings for the valuable information. Your experience
> > indeed backs up the claim of the Firefox browser developers of the
> > difficulty to arrange installers for different platforms.
>
> > My proposition just to put the STIX fonts in a directory of Firefox to
> > be used by the browser for MathML purpose only does not cause any
> > problem for precedence. That way the OS does not know of their
> > existence.
>
> > If a developer wants to use the same fonts for a different purpose,
> > he'll have just to install them in the system. An additional  1MB on a
> > hard disk of 100GB is negligible. It will be no interference between
> > the fonts in Firefox and the system, and no problem of precedence will
> > arise.
> > _______________________________________________
> > dev-tech-mathml mailing list
> > [email protected]
> >https://lists.mozilla.org/listinfo/dev-tech-mathml