[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Comments on 0.56

>  From a usability standpoint, wouldn't it sometimes be better to have a symbol
>  from cm (even if it isn't a perfect match) than no symbol at all ? 

> this is a very important issue, this is where usually implementors &
> users disagree. If a symbol is needed, it is needed, period. The fact
> that the designer did not consider it makes that usually you won't use
> that font set, but you sometimes don't have the choice. On the other
> hand you can't mix anything.  I hit the problem when trying to find a
> math font for Minion: fontinst made dotlessj unfakable, and a cm
> dotlessj would have been a nightmare next to a minion j.

There has been a very long discussion about dotlessj on tex-fonts 
last year, but without any real conclusion IIRC.

The questions remains what we can and should do.  For instance,

- if a lc. variant Greek letter doesn't exist in Euler or Symbol,
  should we substitute the corresponding non-variant Greek letter
  to fill the slot?

- if the default Latin and Greek alphabet is upright in Euler Roman,
  should we put the same glyphs in both sets of Greek rather than
  leaving one of them empty and filling them with several rows of 
  missing glyph markers?

I'd tend to say yes in both of these cases.  However, I would be
skeptical about default substitutions in the case of other symbols.

> Using postscript, you often can fake more quasi-alphabetic glyphs
> (i think of reflected letters, barred ones, etc.). What cannot be
> faked should not be implemented but should remain accessible as a
> backup default. 

By this logic, none of the unavailable symbols should be filled with 
a missing glyph marker, so that some replacement can be added later,
or do I misunderstand you?

Anyway, even if we map MS1/MathTime -> dummy, nothing is stopping 
you from substituting MS1/CM to get a specific symbol, if you decide 
that you want to live with the inconsistencies caused.  However,
I still maintain that it would be wrong to automatically substitute 
MS1/MathTime -> MS1/CM by default.

Cheers, Ulrik.