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

Re: comments on 0.59


> 
> Here are a few quick observations on mathfont/0_59/src0_59.tgz:
> 
> * etx/old/omx.etx: 
>   your version seems to be the original version from the fontinst 
>   distribution which differs from the version in my source tree.
>   In particular, it misses the bug fix for \rmoustache and the
>   renaming of glyph names from xxxbig/Big/bigg/Bigg to xxx1/3/5/7.
>
>   It is possible, though, that this file was never actually used 
>   anyway, except in an experimental version, in which I tried 
>   a partial VF mplementation of MXP with cmex+stmary directly.
>   (BTW, etx/fontspecifc/stmary.etx also seems to be missing.)

For some reason I didn't find omx.etx in my source tree anymore and
copied the original over. 
    
> * etx/new/mx1*.etx:
>   seems to have changed significantly compared to my version.
>   I'll have to check this more carefully, I guess.

Yes, this may be true and if so it is due to my inability to properly
apply your patch back in february.

> * mtx/defaults/mspmissing.mtx:
>   needs to be updated to reflect changes in etx/new/mspraw.etx.
>   The following lines
>     \missingglyph{smallsigmaintegral}
>     \missingglyph{smallslashintegral}
>   should be removed from mspmissing.mtx and added to ms2missing.mtx
> 
> * mtx/defaults/ms2missing.mtx:
>   For some reason I had changed the first row of the arrow kit
>   from \missingglyph to \controlgglyph, but this change seems 
>   to be missing in your version.  I can't recall the reasoning,
>   though, so it probably doesn't really matter.

I think the only glyph for which this is relevant is EXT and I thought
that I had left that one in ?

> * mtx/{cm,concrete/euler}/mc*hax.mtx:
>   There appear to be several changes compared to my version,
>   which I haven't checked in detail yet.  Did you miss some
>   of my patches or did you change anything yourself?
>
> * tex/{cm,concrete/euler}.tex:
>   Same question as above.  Diff'ing your version against the
>   original 0.58 has fewer changes than diff'ing with my version,
>   so I wonder if there's anything I didn't that got lost?
> 

Yes, I had feared that some version skew might have crept in and
that I didn't get all the changes. If in doubt, use your version.

> Other suggestions:
> 
> * derived/labsy[5-9].mf:
>   What about renaming them to lasyb[5-9].mf?  Otherwise MakeTeXPK
>   might produce a directory linotype/aboecklin as per fontname.

Fine with me. I think I arrived at that name by analogy with 
cmsy - cmbsy
cmex - cmbex

> 
> Some comments on de-virtualization:
> 
> * Although your first attempt at organizing the files is already
>   quite good, I think it would be more useful to try to organize
>   the files by type of symbols.  In particular, 
> 
>   - keep uc/lc Greeek letters in separate files
>   - keep all the Hebrew glyphs from cmsy + msbm together
>   - keep all the delimiters from cmr + cmsy + xma together
>   - split cmr-stuff into delimiters, punctuation, other
>

Yes, but this was hacked together quickly just to see how it would
work and to give Taco a good start on the de-virtualization. It certainly
need more reorganization. 

>     (and don't worry too much about crisp=0 in cmr and >0 in cmsy,
>     this only leads to stylistic problems such as square corners
>     in brackets but rounded in corners in floors/ceilings, which
>     we might avoid if we use the same parameters for all of them)

It is obvious that you know more about the parametrisation of cm 
than me.
 
>   Most importantly, rewrite all the files to use the technique:
> 
>     iff known slot_whatever: cmchar "Whatever";
>     beginchar (slot_whatever, ...) 
>     ... endchar;
>
>   This way, you could input files several times and skip over
>   glyphs with undefined slot assignment without having to repeat
>   code fragments and split out common parts in a separate files.
> 
>   For instance, you could input the same file for uc Greek letters 
>   in upright and italic, but use separate files for lc Greek.

Good idea. 

> I might a have closer look this weekend, if I can afford some time.
> Cheers, Ulrik.

Thanks. I you could afford the time to create a patch containing stuff
which I missed/misapplied would be helpful. I'm obviously not very good
at this. Sorry for the inconvenience.

Matthias

-- 
Matthias Clasen, 
Tel. 0761/203-5606
Email: clasen@mathematik.uni-freiburg.de
Institut fuer Mathematik, Albert-Ludwigs-Universitaet Freiburg