[fontforge-users] FontForge will not open the Terminus raster font

[Anton Shepelev]

Hello, all

FontForge will not open the wonderful Terminus font:

http://terminus-font.sourceforge.net/

saying that

terminus.fon is not in a known format (or uses
features of that format fontforge does not sup-
port, or is so badly corrupted as to be unread-
able)

After I found out that Visual Studio no longer sup-
ports raster fonts, my only hope for comfortable
coding has been with FontForge:

http://www.electronicdissonance.com/2010/01/raster-fonts-in-visual-studio-2010.html

Can you help?

[Js jS]

you have not stated what you would like to do and what it is that you need
help with?

Post by Anton Shepelev
Hello, all
http://terminus-font.sourceforge.net/
saying that
terminus.fon is not in a known format (or uses
features of that format fontforge does not sup-
port, or is so badly corrupted as to be unread-
able)
After I found out that Visual Studio no longer sup-
ports raster fonts, my only hope for comfortable
http://www.electronicdissonance.com/2010/01/raster-fonts-in-visual-
studio-2010.html
Can you help?
--
Please, do not send replies to the list to my e-mail.
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Js jS]

ok, i had a closer look

you can "import" the "bdf" file from terminus, but you cannot generate an
opentype or truetype font from it, only raster font

try using "courier new" font with visual studio, i have always used it with
notepad++ for programming and it is pretty legible (for me anyway)

Post by Js jS
you have not stated what you would like to do and what it is that you need
help with?

Post by Anton Shepelev
Hello, all
http://terminus-font.sourceforge.net/
saying that
terminus.fon is not in a known format (or uses
features of that format fontforge does not sup-
port, or is so badly corrupted as to be unread-
able)
After I found out that Visual Studio no longer sup-
ports raster fonts, my only hope for comfortable
http://www.electronicdissonance.com/2010/01/raster-fonts-
in-visual-studio-2010.html
Can you help?
--
Please, do not send replies to the list to my e-mail.
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Anton Shepelev]

Post by Js jS

Post by Anton Shepelev
http://terminus-font.sourceforge.net/
saying that
terminus.fon is not in a known format (or uses
features of that format fontforge does not sup port,
or is so badly corrupted as to be unreadable)

ok, i had a closer look

Thanks!

Post by Js jS
you can "import" the "bdf" file from terminus, but you
cannot generate an opentype or truetype font from it, only
raster font.

Why not? A vector font with each pixel drawn as vectorized
square is not impossible, although I should have to create a
separate version for each font size, see for example, this
wonderful collection of classic raster fonts in TTF:

http://int10h.org/oldschool-pc-fonts

Post by Js jS
try using "courier new" font with visual studio, i have
always used it with notepad++ for programming and it is
pretty legible (for me anyway)

I am very picky about fonts, that why I am here, after all.
Courier is legible, but I dislike the look of it. I use PT
Mono currently.

[Dave Crossland]

Hi

I guess you may need to find on github etc a script to draw the bitmap data
in OpenType vectors, or write your own.

Post by Anton Shepelev

Post by Js jS

Post by Anton Shepelev
http://terminus-font.sourceforge.net/
saying that
terminus.fon is not in a known format (or uses
features of that format fontforge does not sup port,
or is so badly corrupted as to be unreadable)

ok, i had a closer look

Thanks!

Post by Js jS
you can "import" the "bdf" file from terminus, but you
cannot generate an opentype or truetype font from it, only
raster font.

Why not? A vector font with each pixel drawn as vectorized
square is not impossible, although I should have to create a
separate version for each font size, see for example, this
http://int10h.org/oldschool-pc-fonts

Post by Js jS
try using "courier new" font with visual studio, i have
always used it with notepad++ for programming and it is
pretty legible (for me anyway)

I am very picky about fonts, that why I am here, after all.
Courier is legible, but I dislike the look of it. I use PT
Mono currently.
--
Please, do not forward replies to my e-mail.
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Anton Sherwood]

Post by Dave Crossland
I guess you may need to find on github etc a script to draw the bitmap
data in OpenType vectors, or write your own.

Meanwhile I've been working sporadically on an algo to fit "most
graceful" curves to certain old favorite bitmap fonts. When it works,
maybe someone with more skill would be interested enough to make it
user-friendly.

[Jose Da Silva]

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit "most
graceful" curves to certain old favorite bitmap fonts. When it works,
maybe someone with more skill would be interested enough to make it
user-friendly.

I'll be interested in looking at it (when time permits).

[Dave Crossland]

Graceful like Spiro?

Post by Jose Da Silva

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit "most
graceful" curves to certain old favorite bitmap fonts. When it works,
maybe someone with more skill would be interested enough to make it
user-friendly.

I'll be interested in looking at it (when time permits).
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Anton Sherwood]

Post by Dave Crossland

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit
"most graceful" curves to certain old favorite bitmap fonts.

Graceful like Spiro?

I am using similar curves, but the interesting part of my approach is in
deciding by how much to miss the input points.

The goal is an outline font which, at appropriate low resolution,
matches the original bitmap dot-for-dot. To do this, the midline of the
stroke need not touch any grid point.

My current thinking is to fit a piecewise Euler spiral to the input grid
points exactly and then relax it: for each input point, try a path that
meets all except that one, and shift the missing point to a compromise
between the grid point and the point where the test arc comes nearest --
with the constraint that the path must not miss any input point by half
a unit, nor come within half a unit of any "white" grid point.

[Jose Da Silva]

Post by Anton Sherwood

Post by Dave Crossland

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit
"most graceful" curves to certain old favorite bitmap fonts.

Graceful like Spiro?

I am using similar curves, but the interesting part of my approach is in
deciding by how much to miss the input points.

I think if you also have some user configurable decisions such as corners or
rounding, this might give your routines some flexibility from a font
designer's point of view. Some text looks better if you round the corners,
others look better if you have solid angles. Maybe a configurable radius
(tiny, medium, large, or a value like N.NN).

Post by Anton Sherwood
The goal is an outline font which, at appropriate low resolution,
matches the original bitmap dot-for-dot. To do this, the midline of the
stroke need not touch any grid point.

Low resolution fonts will be a challenge.
The URL mentioned earlier may be a worthy test-ground for testing.
http://int10h.org/oldschool-pc-fonts

Post by Anton Sherwood
My current thinking is to fit a piecewise Euler spiral to the input grid
points exactly and then relax it: for each input point, try a path that
meets all except that one, and shift the missing point to a compromise
between the grid point and the point where the test arc comes nearest --
with the constraint that the path must not miss any input point by half
a unit, nor come within half a unit of any "white" grid point.

See if LibSpiro can help you, this way you can avoid re-inventing the same
wheel. My recommendation is work with the latest libspiro tag point as the
code is fairly robust and forgiving, or for additional flexibility, you can
try libspiro HEAD (which also includes 'a' and 'h').
FYI - At the moment, I'm not satisfied with libspiro in terms of the
handling the endpoints (which is why I've turned-off tests 10 and 11 in the
latest commits, but from your point of view, i don't think it should affect
you much for what you plan on doing. To better explain this, I was
attempting to turn one of the tests 90 degrees, and realized that libspiro
was not giving me the expected results when I tried starting with '[', and
therefore also 'a'. You should be okay if your start-point is a corner or a
curve, like in the example 'a' shown in the introduction.
I also suggest sticking with G2 curves 'c' to keep the results simple and
probably leave you with more control of expected results. G4 curves are a
bit more complicated (you can view it as 2 G2 curves between points).
Treat '['...']' as straight lines.
The existing libspiro code should be capable of quadratic output if you're
looking for that. It just needs a switch added, but we have been pretty
busy putting-out other more imediate fires in FontForge.

Do you have a planned license? GPL3+? Other?

[Anton Sherwood]

Post by Jose Da Silva

Post by Anton Sherwood

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit
"most graceful" curves to certain old favorite bitmap fonts.

I think if you also have some user configurable decisions such as corners
or rounding, this might give your routines some flexibility from a font
designer's point of view. Some text looks better if you round the corners,
others look better if you have solid angles. Maybe a configurable radius
(tiny, medium, large, or a value like N.NN).

Heh, well, I had in mind to fit each stroke separately,
then bring them into FF and deal with overlaps by hand.

Post by Jose Da Silva

Post by Anton Sherwood
My current thinking is to fit a piecewise Euler spiral to the input grid
points exactly and then relax it: for each input point, try a path that
meets all except that one, and shift the missing point to a compromise
between the grid point and the point where the test arc comes nearest --
with the constraint that the path must not miss any input point by half
a unit, nor come within half a unit of any "white" grid point.

See if LibSpiro can help you, this way you can avoid re-inventing the same
wheel. [...]

Is there a Python wrapper? My C is very rusty.

Can LibSpiro tell me, for a given spiral and a point not on that spiral,
where it comes nearest?

Post by Jose Da Silva
I also suggest sticking with G2 curves 'c' to keep the results simple [...]

Yes; though I would like to see what happens if the program runs until
it converges with G2, and then runs to improve that solution with G3 and
so on.

Post by Jose Da Silva
Do you have a planned license? GPL3+? Other?

yah, GPL, I guess.

...
Possibly amusing:
https://bendwavy.org/wp/?p=1986
https://bendwavy.org/wp/?p=3529
https://bendwavy.org/wp/?p=3554

[Jose Da Silva]

Post by Anton Sherwood
Heh, well, I had in mind to fit each stroke separately,
then bring them into FF and deal with overlaps by hand.

That works too. Fontforge already has several bells and whistles ready, so
don't need to reinvent that wheel either. Good idea.

Post by Anton Sherwood

Post by Jose Da Silva
See if LibSpiro can help you, this way you can avoid re-inventing the
same wheel. [...]

Is there a Python wrapper? My C is very rusty.

I have not had time to look through fontforge/python.c, so not sure what it
has already. I know I would have to add 'a' and 'h' handling in any case.
Checking google, seems someone already created a python wrapper.

I'd be interested to hear feedback if you do want to try it out.
https://github.com/perey/PySpiro

Post by Anton Sherwood
Can LibSpiro tell me, for a given spiral and a point not on that spiral,
where it comes nearest?

The way you use spiro, is by creating 'knot' points where you want curves
to cross through, for example, if you want spiro to cross through:
{x0=0.0,y0=0.0}, {x1=2.0,y1=3.0}, {x2=3.0,y2=7.0}
spiro will attempt to find the best curve fits to cross through those points.
Sometimes, some curves are just not possible, and libspiro will let you
know it tried, but was not able to find a solution to fit your choice of
points.

This example doesn't have straight lines, but it's a short example.
Using a test input pattern (tests/call-test1):
80, 738, '{'
749, 540, 'o'
671, 309, 'o'
521, 396, 'o'
377, 333, 'o'
467, 231, '}'
spiro makes six attempts and comes out with a solution like:

moveto(80,738)_1
mark_knot()_0
curveto(172.314,664.426, 287.04,622.654, 403.914,606.06)
curveto(462.352,597.763, 521.42,594.908, 580.064,588.226)
curveto(638.708,581.543, 698.554,570.643, 749,540)
mark_knot()_1
curveto(776.321,523.404, 800.431,500.639, 815.305,472.344)
curveto(830.178,444.049, 835.075,410.104, 826.082,379.43)
curveto(817.089,348.755, 793.992,322.521, 764.639,309.862)
curveto(735.286,297.204, 701.137,298.343, 671,309)
mark_knot()_2
curveto(643.661,318.667, 619.475,335.4, 595.839,352.199)
curveto(572.203,368.998, 548.225,386.017, 521,396)
mark_knot()_3
curveto(493.059,406.245, 461.39,408.353, 433.721,397.395)
curveto(419.886,391.915, 407.262,383.305, 397.36,372.199)
curveto(387.457,361.092, 380.322,347.504, 377,333)
mark_knot()_4
curveto(371.073,307.121, 378.034,278.522, 395.545,258.566)
curveto(413.056,238.61, 440.626,227.952, 467,231)
end

If you do:
./configure --enable-verbose_lib
make check
then run:
tests/call-test1
you'll see the info shown above.

Post by Anton Sherwood

Post by Jose Da Silva
I also suggest sticking with G2 curves 'c' to keep the results simple [...]

Yes; though I would like to see what happens if the program runs until
it converges with G2, and then runs to improve that solution with G3 and
so on.

Libspiro starts out with an assumed expansion, but then expands it further
for best fit. For example, it will start out by giving space for 2 more
points to make a G2 curve, and will start-out with an array expansion of an
extra 4 points to make a G4 curve, but as it tries to fit a solution, the
array can/will expand in some places

Post by Anton Sherwood

Post by Jose Da Silva
Do you have a planned license? GPL3+? Other?

yah, GPL, I guess.
...
https://bendwavy.org/wp/?p=1986
https://bendwavy.org/wp/?p=3529
https://bendwavy.org/wp/?p=3554

Interesting.

[Nathan Willis]

Post by Anton Sherwood

Post by Dave Crossland

Post by Anton Sherwood
Meanwhile I've been working sporadically on an algo to fit
"most graceful" curves to certain old favorite bitmap fonts.

Graceful like Spiro?

I am using similar curves, but the interesting part of my approach is in
deciding by how much to miss the input points.
The goal is an outline font which, at appropriate low resolution,
matches the original bitmap dot-for-dot. To do this, the midline of the
stroke need not touch any grid point.

Anton,

Have you read what Bigelow & Holmes wrote about their experience converting
Apple's system bitmap fonts to TrueType in the 90s, hopefully?

The tools available have evolved considerably since then, of course, but
I'm sure their insights to the issues would be worth seeing. There are a
lot of design decisions that crop up when you're out in the weeds, for
instance, like the drastic differences you get in proportions between
different bitmap-point-sizes. Today we could account for that by having
optical-size variants, which would be fascinating to explore as a project.

Nate

Post by Anton Sherwood
My current thinking is to fit a piecewise Euler spiral to the input grid
points exactly and then relax it: for each input point, try a path that
meets all except that one, and shift the missing point to a compromise
between the grid point and the point where the test arc comes nearest --
with the constraint that the path must not miss any input point by half
a unit, nor come within half a unit of any "white" grid point.
--
*\\* Anton Sherwood *\\* www.bendwavy.org
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Anton Sherwood]

Post by Nathan Willis
Have you read what Bigelow & Holmes wrote about their experience
converting Apple's system bitmap fonts to TrueType in the 90s, hopefully?

I've read something about it, but not recently!

Post by Nathan Willis
The tools available have evolved considerably since then, of course, but
I'm sure their insights to the issues would be worth seeing. There are
a lot of design decisions that crop up when you're out in the weeds, for
instance, like the drastic differences you get in proportions between
different bitmap-point-sizes. Today we could account for that by having
optical-size variants, which would be fascinating to explore as a project.

Fortunately ;) the font at the top of my agenda, that of the PLATO IV
terminal, has only one size. But I have pondered what's the Right Way
to approach Apple Geneva, for example.

[Js jS]

here is a rough python script that may get you started

instructions are in the file at top

you will need a bdf file

i used one from
https://sourceforge.net/projects/terminus-font/files/terminus-font-4.40/terminus-font-4.40.tar.gz

copy one of the bdf files from the downloaded archive into the same folder
as your font file

the script is really rough ... it throws an error on empty glyphs ... you
have to create all glyphs before you run it ... just edit one glyph (any
glyph) ... put in a single point ... and then copy the glyph to all the
other glyphs in the font file (select ... copy ... select all ... paste)

set em = 4096

then run this script by copying into the script window and clicking OK

Post by Anton Sherwood

Post by Nathan Willis
Have you read what Bigelow & Holmes wrote about their experience
converting Apple's system bitmap fonts to TrueType in the 90s, hopefully?

I've read something about it, but not recently!

Post by Nathan Willis
The tools available have evolved considerably since then, of course, but
I'm sure their insights to the issues would be worth seeing. There are
a lot of design decisions that crop up when you're out in the weeds, for
instance, like the drastic differences you get in proportions between
different bitmap-point-sizes. Today we could account for that by having
optical-size variants, which would be fascinating to explore as a

project.
Fortunately ;) the font at the top of my agenda, that of the PLATO IV
terminal, has only one size. But I have pondered what's the Right Way
to approach Apple Geneva, for example.
--
*\\* Anton Sherwood *\\* www.bendwavy.org
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Js jS]

i figured how to create characters before drawing into them, so now you
just get the bdf file, start new font and run the script

Post by Js jS
here is a rough python script that may get you started
instructions are in the file at top
you will need a bdf file
i used one from https://sourceforge.net/projects/terminus-font/files/
terminus-font-4.40/terminus-font-4.40.tar.gz
copy one of the bdf files from the downloaded archive into the same folder
as your font file
the script is really rough ... it throws an error on empty glyphs ... you
have to create all glyphs before you run it ... just edit one glyph (any
glyph) ... put in a single point ... and then copy the glyph to all the
other glyphs in the font file (select ... copy ... select all ... paste)
set em = 4096
then run this script by copying into the script window and clicking OK

Post by Nathan Willis

Post by Nathan Willis
Have you read what Bigelow & Holmes wrote about their experience
converting Apple's system bitmap fonts to TrueType in the 90s,

hopefully?
I've read something about it, but not recently!

Post by Nathan Willis
The tools available have evolved considerably since then, of course, but
I'm sure their insights to the issues would be worth seeing. There are
a lot of design decisions that crop up when you're out in the weeds, for
instance, like the drastic differences you get in proportions between
different bitmap-point-sizes. Today we could account for that by having
optical-size variants, which would be fascinating to explore as a

project.
Fortunately ;) the font at the top of my agenda, that of the PLATO IV
terminal, has only one size. But I have pondered what's the Right Way
to approach Apple Geneva, for example.
--
*\\* Anton Sherwood *\\* www.bendwavy.org
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Jose Da Silva]

Post by Js jS
i figured how to create characters before drawing into them, so now you
just get the bdf file, start new font and run the script

Interesting method to vectorize the font.
Somewhat similar to building a brick wall.
I could see that cleanup is going to be a fair bit of deleting unneeded
lines, but this is a good starting point, and you do indicate it's very
preliminary.
I would recommend building-up on your idea, and having an intermiediate
stage.
As this is getting technical, I would recommend moving the discussion to
the developer list, and leave the user list to concentrate mainly on fonts.
...however, in terms the middle stage, this is the suggestion:

suppose the font is an 8x8 pixel square.
create an array (8x2+1)x(8x2+1).
then for your main draw routine, replace it with something to fill in the
new array.

for looping through x and y do....
# print "ox,oy",ox, oy
a(x*2,y*2)=a(x*2,y*2)+1 #top edge
a(x*2,y*2+1)=a(x*2,y*2+1)+1 #right edge
a(x*2+1,y*2+1)=a(x*2+1,y*2+1)+1 #bottom edge
a(x*2+1,y*2)=a(x*2+1,y*2)+1 #left edge

after you fill in the new array, you can now cleanup and remove overlaps
where you see a(nx,ny)>=2, as these are overlapping lines, while 1 is most
likely an edge and therefore what you want to keep.
after cleanup, this next step is harder, but anything with 1, you can test
straight rows and figure-out a way to make your short segments into one long
line.

[Abraham Lee]

Hi, All!

I realize this is in response to a slightly older thread, but I thought I'd
share my results anyway.

Post by Jose Da Silva

Post by Js jS
i figured how to create characters before drawing into them, so now you
just get the bdf file, start new font and run the script

Interesting method to vectorize the font.
Somewhat similar to building a brick wall.
I could see that cleanup is going to be a fair bit of deleting unneeded
lines, but this is a good starting point, and you do indicate it's very
preliminary.
I would recommend building-up on your idea, and having an intermiediate
stage.
As this is getting technical, I would recommend moving the discussion to
the developer list, and leave the user list to concentrate mainly on fonts.
suppose the font is an 8x8 pixel square.
create an array (8x2+1)x(8x2+1).
then for your main draw routine, replace it with something to fill in the
new array.
for looping through x and y do....
# print "ox,oy",ox, oy
a(x*2,y*2)=a(x*2,y*2)+1 #top edge
a(x*2,y*2+1)=a(x*2,y*2+1)+1 #right edge
a(x*2+1,y*2+1)=a(x*2+1,y*2+1)+1 #bottom edge
a(x*2+1,y*2)=a(x*2+1,y*2)+1 #left edge
after you fill in the new array, you can now cleanup and remove overlaps
where you see a(nx,ny)>=2, as these are overlapping lines, while 1 is most
likely an edge and therefore what you want to keep.
after cleanup, this next step is harder, but anything with 1, you can test
straight rows and figure-out a way to make your short segments into one long
line.

I wrote a fairly extensive python script that follows a procedure similar
to the one suggested (even though I came up with it on my own) to convert
the Terminus fonts into vectorized pixel representations. Each pixel was
drawn as a square and in the end I was able to clean up the glyphs by
removing all the interior points/contours and removing extra
points/contours along the edges (e.g., change 3 colinear segments into a
single segment). I call the font family "Vecterminus" to comply with SIL
OFL. You can find the OTF files on my github project page:
https://github.com/tisimst/vecterminus (couldn't think of a better place to
put them).

There are "medium" and "bold" files for each pixel font (12px, 14px, 16px,
18px, 20px, 24px, 28px, and 32px), but I haven't attempted to make italic
and bold-italic variants yet. The files are true to the originals in every
way, except they are vectorized.

I've done some quick testing to see how well them are rendered at various
pt sizes. Doing so, led me to find some optimal display sizes, as shown in
the file
https://github.com/tisimst/vecterminus/blob/master/match-font-to-size-chart.txt.
There's a chart for Notepad++ users (which I use the most) and Word users
(which, I don't think I would ever use for coding, but who knows, maybe
someone does). Not sure why there's a discrepancy between the two charts,
but that's what I found from my own testing. The matched pt sizes assume a
screen resolution of 96dpi and look really nice at their intended sizes,
but not so great at other sizes.

I initially left them scaled to their low res sizes (e.g., em = 12 for 12px
font), but Windows struggled rendering them, so they are all scaled to
em=1000, regardless of original px size.

Anyway, have fun using them!

Best,
Abraham

[Abraham Lee]

A little addendum...

Post by Abraham Lee
I've done some quick testing to see how well them are rendered at various
pt sizes. Doing so, led me to find some optimal display sizes, as shown in
the file https://github.com/tisimst/vecterminus/blob/
master/match-font-to-size-chart.txt. There's a chart for Notepad++ users
(which I use the most) and Word users (which, I don't think I would ever
use for coding, but who knows, maybe someone does). Not sure why there's a
discrepancy between the two charts, but that's what I found from my own
testing.

This seems to only be the case on my laptop (sigh). Other monitors show the
"Word" matched sizes correctly for 96dpi screens:

12px - 9pt
14px - 10.5pt
16px - 12pt
18px - 13.5pt
20px - 15pt
24px - 18pt
28px - 21pt
32px - 24pt

In case you don't have 96dpi screens, but something else, here's the math
to know the right display sizes:

dpi = 96 <---pixels per inch
pt = 72 <--- pts per inch

ratio = pt/dpi = 72/96 = 3/4 = .75

Thus, 12px*ratio = 9pt

So, if your screen resolution is different, just replace 96 with whatever
you monitor resolution is to determine the appropriate sizes for each font.

Best,
Abraham

[Nathan Willis]

Post by Anton Sherwood

Post by Nathan Willis
Have you read what Bigelow & Holmes wrote about their experience
converting Apple's system bitmap fonts to TrueType in the 90s, hopefully?

I've read something about it, but not recently!

This is pretty good, and is online:
http://cajun.cs.nott.ac.uk/compsci/epo/papers/volume4/issue3/ep050cb.pdf

Post by Anton Sherwood

Post by Nathan Willis
The tools available have evolved considerably since then, of course, but
I'm sure their insights to the issues would be worth seeing. There are
a lot of design decisions that crop up when you're out in the weeds, for
instance, like the drastic differences you get in proportions between
different bitmap-point-sizes. Today we could account for that by having
optical-size variants, which would be fascinating to explore as a

project.
Fortunately ;) the font at the top of my agenda, that of the PLATO IV
terminal, has only one size. But I have pondered what's the Right Way
to approach Apple Geneva, for example.
--
*\\* Anton Sherwood *\\* www.bendwavy.org
------------------------------------------------------------
------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
fontforge-users mailing list
https://lists.sourceforge.net/lists/listinfo/fontforge-users
http://fontforge.10959.n7.nabble.com/User-f8781.html

[Anton Shepelev]

Post by Dave Crossland
I guess you may need to find on github etc a script
to draw the bitmap data in OpenType vectors, or
write your own.

Indeed. The author of the Ultimate old-school font
pack said he used therefor the Bits'N'Pical convert-
er: https://github.com/kreativekorp/bitsnpicas .

[Anton Shepelev]

Post by Js jS
you have not stated what you would like to do and what it
is that you need help with?

My mistake, but you have gussed it aright.

[Anton Shepelev]

FontForge will not open the wonderful Terminus
http://terminus-font.sourceforge.net/
saying that
terminus.fon is not in a known format (or us-
es features of that format fontforge does not
support, or is so badly corrupted as to be un-
readable)

If that is because Terminus is a multi-size and mul-
ti-style font, Fony should help, but its homepage is
no longer alive:

http://hukka.furtopia.org/projects/fony/

It seems, however, available elsewhence, e.g.:

http://www.softpedia.com/get/Multimedia/Graphic/Graphic-Editors/Fony.shtml

[Anton Shepelev]

If that is because Terminus is a multi-size and
multi-style font, Fony should help, but its home-
http://hukka.furtopia.org/projects/fony
http://www.softpedia.com/get/Multimedia/Graphic/Graphic-Editors/Fony.shtml

Fony is alive and thriving. Its new home is here:

http://hukka.ncn.fi/?fony