let me tell you how it is, in detail, this glyph database of mine.
used a program to extract data from the 8x16 font files (*.F16) that I
1) The first 8x16 font file (cp437.f16), after being analyzed by the
program I have (which was developed by Mateusz Viste), generated a text
file with all 256 glyphs on a 8x16 matrix each.
2) The following step was to edit this resulting text file (with nothing
else than the regular DOS EDIT) to insert the Unicode codepoint for each
glyph. This first text file was renamed to "UNICODE.TXT".
3) The following step was to extract cp850.f16. Another text file with
256 glyphs. Then, I had to:
A) Remove all the redundant glyphs. For instance, all characters in
the lower half of the table because they're the same on every codepage
(i.e. pure ASCII) and they had already been extracted from cp437.f16,
along with those linedraw, shade and block characters which had also
already been extracted from cp437.f16.
B) Insert the Unicode codepoints for the new glyphs.
C) Merge this text file into UNICODE.TXT.
...And so on, repeating step #3 on all the other 8x16 font files. I did
that all in 2006; I don't remember how many codepages I had developed by
then. I can tell you that, by now, they are 192 font files (cp437 + 191
files). Yes, the lower half of the other 191 files is dismissed, so
we're talking about 128 glyphs per each new font file. Then, there are
about 20 to 30 redundancies per font file which are also dismissed, like
linedraw, shade, block characters and a few others. Let's say, then, to
make it easier, that I have around 100 new glyphs per font file. That
gives me 19,100 glyphs. I know that I have 1,745 glyphs in the database.
Some of them were drawn directly into the database, not being extracted
from any *.F16 file. I still have much work to do to catch up with 5
years of new codepages to be extracted.
I do admire Czyborra for the work he has accomplished but his database
sticks purely to Unicode only. There are a few codepages on FreeDOS
containing glyphs which aren't available on Unicode yet (and I'm already
considering its latest version, 6.0). Two of them are lithuanian (cp777
and cp778), containing precomposed latin letters such as R-tilde and
A-ogonek-acute. They are available on my database in the range of
address called "PUA" (Private Use Area: E000h-F8FFh). Eventually, when
such characters are included into Unicode, all I will have to do is to
change its codepoint number in the glyph database. Furthermore, there
are many other precomposed latin letters on codepages that I have
created for FreeDOS in order to assist american indigenous, african and
oceanian languages. I've been lately working on precomposed acute- and
grave-accented cyrillic vowels and accented georgian vowels which will
also find their way into the database on the PUA.
Furthermore, there's another point - this one being visual: the shape of
the 8x16 glyphs on our databases are very different. It would result on
an ugly printed text if a single database contained glyphs with
different shapes (if I was to consider merging both databases). It would
be like reading a text file with, let's say, Courier New and Times New
Roman letters in a same word. Anyway, I'll talk to him to see if he
allows me to merge all glyphs belonging to scripts different from the
ones I have addressed so far, in order to avoid the "Courier New/Times
New Roman" visual effect. That would spare me from "reinventing the
wheel" and start working on scripts for which the glyphs he has already
made while I would focus on letters for scripts with which I have worked
Em 7/5/2011 01:24, Marco Achury escreveu:
> I found very interesting that all the required sizes are multiple of six,
> so is wise that glyph designer to use a matrix with 6 pixels...
> Only 9 pins high-res is exact multiple of eigth
> A simple way is to reduce the character with is to look for 2 consecutive
> columns with the same bits, and delete one (as contain redundant
> You need to deleted 2 columns on each glyph to get a 6 columns matrix.
> But to get beauty characters with good simetry is better to do such
> job by hand.
> How many characters has you database?
> A good text editor with column mode (as crimson editor) will help
> El 07/05/2011 12:35 a.m., Henrique Peron escribió:
>> Alain, the answer is to interpolate the glyphs but I don't know how to
>> do it - by the way, at least on what concerns Epson printers, I think
>> that there's a native way to interpolate low-resolution glyphs. Anyone
>> wishing to develop the printer driver would also have to check Panasonic
>> Em 6/5/2011 23:01, Alain Mouette escreveu:
>>> In the printes, fonts should have
>>> Low Res 9 pin: 72/6 = 12 pixels
>>> Low Res 24 pin: 180/6= 30 pixels
>>> Hi Res 9 pin: 144/6= 24 pixels
>>> Hi Res 24 pin: 360/6= 60 pixels
>>> This for the whole line (glyph + spacing). Can tou imagine how to
>>> convert your database to work with these resolutions?
> WhatsUp Gold - Download Free Network Management Software
> The most intuitive, comprehensive, and cost-effective network
> management toolset available today. Delivers lowest initial
> acquisition cost and overall TCO of any competing solution.
> Freedos-user mailing list