[comp.text.tex] cmr12 magstep 0 vs cmr10 magstep 1

nbeck@weber.ucsd.edu (Nathaniel Beck) (05/10/91)

I recently posted a question about cmr11 which Don Hosek was kind
enough to answer. But on thinking about fonts and scaling I became
even more confused and am hoping that some guru out there can make
matters clear again.

Emtex (and other pc tex's and latex's) comes with a series of fonts
or font libraries, with a series of standard (say 300dpi laserjet
fonts) at magstep 0, .5, 1, etc. Now my naive calculation tells me
that cmr10 magstep 1 should look pretty much like cmr12 magstep 0
and my lfonts.tex chooses cmr12 magstep 0 for its basic 12 point
type. 

Here are a series of questions.

1. Why should I want to keep cmr10 magstep 1 if I have cmr12 magstep
0?

2. Don Hosek has provided one answer to that question - he prefers
the scaled font for some purposes. But at that point it would make
more sense to think of cmr10 magstep 1 as cmrnew12 magstep 0 so it
could easily be used in the application preferred by Don Hosek

3. Since I am taking up several megs with magstep .5, etc fontlibs,
wouldn;t it be just as efficient to have one big library with cmr10,
cmr10.5, cmr11, etc. It would seem as though I could use less disk
space with that scheme, since I wouldn't need cmr10 magstep 1 and
cmr12.

4. Is the issue that I would need more .tfm files, and the device
drivers have an easier time taking the cmr10.tfm file and increasing
its dimensions for magstep .5 as opposed to looking for a cmr11.tfm.

There seems little question that there must be some advantage to the
magstep scheme (I don't really believe I am smarter than Donald
Knuth). So could some kind guru please clear up my confusion.

Thanks in advance

Neal

  
-- 
Neal Beck 
Dept of Politcal Science, UCSD
beck@ucsd.edu
Dislaimer: The Regents pay me (a bit!) to distribute my opinions.

jg@prg.ox.ac.uk (Jeremy Gibbons) (05/10/91)

nbeck@weber.ucsd.edu (Nathaniel Beck) asks

> 1. Why should I want to keep cmr10 magstep 1 if I have cmr12 magstep
> 0?
> 
> 2. Don Hosek has provided one answer to that question - he prefers
> the scaled font for some purposes. 

> There seems little question that there must be some advantage to the
> magstep scheme

One advantage of the magstep scheme is that you can print out a copy of
your document magnified by some power of 1.2, and (there's a good chance
that) you'll already have the fonts around. (One reason you might want to
do this is for making camera ready copy for reproduction; you get an
effective 360dpi resolution from your 300dpi laser printer, after scaling
it back down photographically.)

(A more useful scheme for those of us in the old world would be to use the
fourth root of 2, which is slightly less than 1.2. This is because A4 paper
dimensions are in the proportion 1:root2; I would then have a good chance
of being able to print 2 A4 pages at .707 magnification on a single A4
sheet, without doing much Metafonting.)

Jeremy

*-----------------------------------------------------------------------*
|  Jeremy Gibbons (jg@prg.oxford.ac.uk)   Funky Monkey Multimedia Corp  |
*-----------------------------------------------------------------------*

dahosek@biivax.dp.beckman.com (05/11/91)

In article <nbeck.673824646@weber>, nbeck@weber.ucsd.edu (Nathaniel Beck) writes:
> I recently posted a question about cmr11 which Don Hosek was kind
> enough to answer. But on thinking about fonts and scaling I became
> even more confused and am hoping that some guru out there can make
> matters clear again.
 
> Emtex (and other pc tex's and latex's) comes with a series of fonts
> or font libraries, with a series of standard (say 300dpi laserjet
> fonts) at magstep 0, .5, 1, etc. Now my naive calculation tells me
> that cmr10 magstep 1 should look pretty much like cmr12 magstep 0
> and my lfonts.tex chooses cmr12 magstep 0 for its basic 12 point
> type. 

> Here are a series of questions.
 
> 1. Why should I want to keep cmr10 magstep 1 if I have cmr12 magstep
> 0?

plain TeX users do \magnification\magstep1 to get an overall
magnification of the document (actually a half-truth, since in
plain TeX, 6.5truein will not be affected by the \magnification
and that's the specified default \hsize. In LaTeX, \magnification
is disabled). Furthermore, many DVI drivers allow additional
magnification at the print stage which allows one to increase the
affective resolution of the document. 
 
> 2. Don Hosek has provided one answer to that question - he prefers
> the scaled font for some purposes. But at that point it would make
> more sense to think of cmr10 magstep 1 as cmrnew12 magstep 0 so it
> could easily be used in the application preferred by Don Hosek

See above.
 
> 3. Since I am taking up several megs with magstep .5, etc fontlibs,
> wouldn;t it be just as efficient to have one big library with cmr10,
> cmr10.5, cmr11, etc. It would seem as though I could use less disk
> space with that scheme, since I wouldn't need cmr10 magstep 1 and
> cmr12.

Well, one way of doing this, especially if you're exclusively a
LaTeX user, is to go through lfonts.tex and record all the fonts
that are used and only keep those PKs on the disk. For a TeX
distribution I put together at Beckman, I installed all fonts
refered to by preload.ori (I use lfonts.new -- see
[anonymous.tex.inputs.latex-mainz] on ymir.claremont.edu) plus
all plain TeX fonts preloaded plus all plain TeX fonts
immediately available at magsteps 0, .5 1-5 and all fonts at
magstep 0. The resulting font set, LHARC'd  fit on 6 360K
floppies (although the actual distribution channel was 1 1/2
1.44M floppies).
 
> 4. Is the issue that I would need more .tfm files, and the device
> drivers have an easier time taking the cmr10.tfm file and increasing
> its dimensions for magstep .5 as opposed to looking for a cmr11.tfm.

There's a negligible performance difference. Most drivers would
open cmr10 and read it in its entirety anyway.

-dh

-- 
Don Hosek // Quixote Digital Typography   714-625-0147
     dhosek@ymir.claremont.edu
On contract to Beckman Instruments        714-961-4562
     dahosek@beckman.com