Their example says that sfcode for a is 1000, thus spacefactor becomes 1000. Suppose the next character was a period, whose sfcode is 3000, thus making f=3000. Now the subsequent space will be increased. This is good since, say, the period was the end of the sentence. Now the following example has the letter A, whose sfcode is 999, thus f=999. This character is followed by the period, whose sfcode was 3000, and now he says f=1000. What?? I don't get it. How did it become 1000?
Perhaps you can also comment on my method to find the correct spacing after a period for txfonts at 12pt:
From the first reference, at the bottom of pg75 after the period he has spacing should be 4.44444 plus 4.99997 minus 0.37036. For cmr 10,
Code: Select all
fontdimen2=3.33333
fontdimen3=1.66666
fontdimen4=1.11111
fontdimen7=1.11111
if (f <= 1000)
then (Interword space = fontdimen2+fontdimen3-fontdimen4)
if (f >= 2000)
then (Interword space = (fontdimen2+fontdimen7) + (fontdimen3*f/1000) + (fontdimen4*1000/f))
Since it ended with a lowercase character then period, I assume f=3000. Using the 2nd equation, I get (3.33333+1.11111)+(1.66666*3000/1000)+(1.11111*1000/3000), which is exactly what he got.
Applying this to txfonts at 12pt (whose values I determined using that bit of code you showed me previously):
Code: Select all
fontdimen2=3.00000
fontdimen3=1.79999
fontdimen4=0.72000
fontdimen7=0.72000
(3.00000+0.72000)+(1.79999*3000/1000)+(0.72000*1000/3000)=9.490336.
So I conlude that I must set xspaceskip=9.949034pt to get the correct spacing after the period. This is very close to the 10pt I tried yesterday by eyeballing it!