I thought, once, that it would be a good idea to document the progress of my typeface development. It has since evolved into a traveller's log, I gather. Are we headed the same way, stranger …?
November 5, 2003
I added a section about specific glyph designs to the resources page after reading Type, handwriting, and lettering webpage after Adam reminded me that it exists. At first I thought I'd read that webpage later, 'cause to be honest I've read enough to be able to create something usable—I just need to create it. But then I thought Aw fuck it, I'll read it and get it out of my mind, so I did. Now I know how to make perfect Islandic glyphs, among other things, but really, I have more important stuff to do. (Well, I made three of the those four glyphs so I can now forget about them.)
There is a section on handwriting which I found quite amusing for some reason. I've never been interested in calligraphy and hopefully never will, but … for a former computer geek I still have a pretty nice, legible handwriting style, but after looking at the sample by Alfred Fairbank I thought it would be fun experimenting with tidying it up a bit. I went ahead and ordered both A Handwriting Manual and A Book of Scripts from the library. It could be a fun project to document my progress of this tidying up by posting samples in my diary or something once in a while … if I'll get time for that, that is.
October 30, 2003
I'm half through the book; reached chapter 5. I'll write my first opinion. Why? I don't know, I just feel like telling the world my thoughts.
The first thing I noticed when I started reading the book is the nasty typeface it was typeset with. I had these weird, skewed—what do you call it—bowls? in cbdpq glyphs with some very thin transitions. After discovering a hyphen that looked like a diagonal I thought What is this shit? Now I'll quote a paragraph from the preface:
Printing can be beautiful as well as effective. My own interest in typography stems in part from a love of books, which is nourished by their feel and look as well as their content. George Bernard Shaw said, “Well-printed books are just as scarce as well-written ones; and every author should remember that the most costly books derive their value from the craft of the printer and not from the author's genius.” When we design and build systems for printing, we can have a profound influence on what written documents look like. Will they be easy to read, or tiring? Will they communicate efficiently, or slow the reader and provoke errors? Will they give pleasure to their owners?
Basing myself on this paragraph I say that this book derives it's value from the craft af the printer and communicates bullshit, slows me, the reader, and provokes not errors but sadness. Is the author trying to prove the point that well-printed books are scarce by providing the world another bad example? Hmmm …
I don't agree with Shaw's second part of the quote about the values. That typography-centric attitudes is pathetic, but … it's not clear what he means by value, so, fuck it. The only excuse Rubinstein has is that he let someone design his book. But why would he do that being able to write a book on type and text composition? I don't know.
So, I started reading the book in an irritated state. The ornaments on every page next to the page numbers continue to piss me off. Instead of falling into depression I thought I will experiment on myself and see if the quote Eric Gill on legibility that I have in resources section holds true. Reaching chapter 5 I can say that I have accustomed myself to the afwul typeface, and I don't feel that it slows me down or whatever. It has become transparent in my mind. I confirmed the axiom (heh, for myself, that is).
The first third of the book is boring and it wasn't very useful for me. There is much background information/theory that I either know already or don't care about for my current purposes. From one third to one half of the book is one chapter which was more interesting to me, though not much. I might use the idea of creating tables for weight and contrast to adjust the—fuck, I don't know what to call it—the standard stems tables to have a better idea of where to make pixel jumps.
At the end of chapter 4, Rubinstein quotes Frederic Goudy with “A design devoid of emotion, rtythm, and expression, yet technically excellent, merely betrays the fact that it has been produced by one who has nothing of value to express.” Compare this with a quote from chapter 1 by Beatrice Warde: “Printing should be invisible.” How can something be invisible if it's expressive and isn't trying to express invisibility? My Goudy buddy seems to live in a typography-centric world and wants to express something (what?) with his typeface design. Well, I don't—I prefer to use words in most cases, and I want a typeface to help me express myself using those words. The absurdity in Goudy's statement can be seen in that if I think the word in itself doesn't express whatever it does, I need a typeface to help me do it—then how about we make books with every occurrence of word “big” in a bigger size, “broader” with broader spacing, “grayscale” in gray, “hard” illegible, “uppercase” in uppercase, “emotion” in some special emotional type (whatever that means …), and so on. Legibility suffers; Goudy can lie down and die, if he hasn't already.
I like the quote by Fernand Baudin which begins chapter 5: “Readability is not wholly built in the letterforms as such. One half of it is in the spacing between the words, the lines, the columns; in the geometry of the text and the margins i.e. in the visual editing.” I have discovered this on my own a long time ago—design is about spacing. And I don't mean just the text, but anything visual. Most web- and GUI-designers have no fucking clue about spacing, and by looking at how they handle the space it's easy to see through their visual skills (apart from the technical and practical). Now that we are at it, not talking specifically about typefaces, I'll mention another important design concept—
Hmmm, why am I doing it my the progress log for my typeface? I don't know … I want to write down my thoughts and I don't know where else to write them. I'm not going to write a book on this topic, so I don't feel like it deserves it's own section somewhere on my webpage.
of invisible sections. I don't know how to explain it wery well, but … I have this theory that more sections bring more complexity and make the brain work needlessly. It's about simplicity, as with everything. If you draw a line on a sheet, you get 2 such sections; if you put a paragraph of text in the middle of a sheet, you get 4 sections (out of margins). If you put another paragraph under the first, you'll add 1 more, or, if the paragraphs aren't aligned vertically, up to 3 more. So, the simplicity is achieved by minimizing the number of sections and by maximizing their similarity. A webpage like CNN.com has an awful lot of invisible (and visible) sections—it sucks. Most of the web sucks. You suck. Now go play in the rain, pinch your partner's ass, burn your favourite book, throw the pizza out of the window, drown your pet in the bathroom, punch the person you always wanted to punch, kill the innocent insect you never wanted to touch, brush your boots, kiss a stranger, make thirty pushups on your knuckles (forty if you are female), look at the sun through a telescope, paint the sky blue (after getting blinded) …
21.10.03
I ordered Digital Typography by Rubinstein a few weeks ago from my local library—today I got it home. It was recommended to me by William Adams back in August, who also pointed out how gross my “s” and “S” were. I knew about their ugliness, of course, but it took me a while to pull myself together to design a decent “s” glyph. I haven't touched “S” yet … I'm a bit afraid, I guess—it's not an easy task! Actually, I'm a bit scared of all the difficult glyphs that I have yet to finish, but I know I can do it, I just … know.
Anyhow, I think I won't work on the typeface for a little while because I want to read the book and because I want to write all the people who I feel I need to keep in touch with. So, I just need some time to do what I have to do. I'll return with new enthusiasm and let the shit hit the fan!
20.10.03
The sad part is that it's more interesting to talk to Adam and girls than making my typeface. The fun part is calling them all arrogant assholes.
14.10.03
Um, I wrote the other day that I had a lot of thoughts. However, I later noticed that I didn't write a lot of thoughts down, so a reader might be wondering if I just pretend to have thought something. Unfortunately, I forget a lot of my (good) thoughts before I manage to write them down. Sometimes they emerge later on only to disappear again before I can capture them. This doesn't really bother me, though—I'm not in a hurry.
Adam Twardoch began a conversation with me last week. He looks like a cat from a long forgotten Russian cartoon from my childhood. Ah, yes, not so forgotten after all: «Ребята, давайте жить дружно …» was what кот Леопольд used to say. He wants to help me.
It's nice when people wish to help you when you don't ask for it. It's nice when people can spare a bit of themselves instead of being controlled by their stupid fears. It gives me a little hope in humanity. I'm not the most helpful person in the world myself (ask my parents if you for some reason don't believe me), and I don't think every person has something to spare in some particular period of their lives, but when everyone in this thread on Typophile agrees that it's too much to spend five minutes to help a stranger, I want them all to get shot at point blank range, in the back. They shit in their pants because they think it has become too easy to do what the man in need politely asks for. Pathetic. Truly pathetic. Idiots. To all of them I have the following to say:
Fuck you! Not as typeface developers, but as fucking human beings. You are welcome. The world would be a better place without you. I hope you won't get an opportunity to raise any children.
This is just my opinion and everyone is welcome to ignore it. I don't want to defend myself here or try to argue about who is right. I don't care who is right—the point is that I want to voice my disagreement because no one else does. And if I misunderstood the whole situation I'll even apologize. (I have yet to read all those replies …)
Back to Adam. He might become a source of some of my motivation to continue the work on my typeface by simply giving me his attention. I don't know how long this is going to last, though. It's somewhat unfortunate that I have a lot of other things in my life to worry about at the moment to be able to get the most out of it.
A thing I forgot to mention last time about my font is that I tried to make all digits of same width. My long-term plan is two divide my typeface into two: one for text and one for GUIs. The GUI version will have digits of same width, I'll make the overloaded apostrophe symmetrical, and there might be other similar things that I will adjust in the future. Even though not all of my digits are finished, I experienced a problem with the ones that were (mostly) done: The width of “1” didn't want to match others with my tried hinting methods. I thought about it for a bit, but since I still didn't care about digits that much and it was an unexpected problem, I chose not to pursue this further and just delta-hinted the two sizes that I mentioned the other day. I'll return to this later.
I finally redesigned “V” and “W” glyphs. But, but, but. It's no good. Through my trials they both aqcuired a style which is mix of Trebuchet MS and Lucida Sans. Before I looked closer, I thought Trebuchet MS had a symmetrical “V”, but the stems actually differ slightly in width. I didn't (and still don't) really see the point in this, as a normal eye will never see the difference, so I thought I'll make the difference somewhat bigger (more Lucidish) and went with that. However, when I looked at the final result, from afar, this difference became like a needle in my eye. I made the difference big enough to see it, but small enough to make it look like an unintentional error by the designer. So! The next thing I'll do is make two versions of both “V” and “W” (and perhaps “M” too, when time comes to that)—symmetrical and a more asymmetrical than the current one—and after thinking for a few years …
I also took a shot at ©, but I don't think it turned out that well. I'm not sure where to put it on baseline … The hinting is not very good at some sizes, and this glyph finally convinced me that I'll have to create (a lot of) standard stems for special cases. (This sounds like an oxymoron, but what can I do?)
If anyone is using my current font and wonders why “g” remains unhinted, it's because every time I see it (about once in every sentence), it motivates me to redesign it. If I hint it right now, it will be passable in my book, and I won't have the motivation and patience to bring it to another level. On the other hand, if I slowly let it get on my nerves, at some point I'll explode and make it perfect. It's just a trick I use on myself. And even if I don't make it perfect after one go, I'll know how to further improve it just like it happend with “V” and “W” glyphs.
On a completely different note: I recently went back to using Acrobat Reader 4 because later versions were being dog slow, and when I opened the FontLab manual today, I noticed that the text isn't rendered very good compared to how I recall it from the latest version (both without the ClearType thingy). I'm not sure, though, but I can't be bothered to reinstall another version. In any case, I think it's unfortunate.
Yes, I have gotten over being disappointed with myself; now I just have to learn how to get over being disappointed with other people, especially people older than me. Nothing unusual, the kid is growing up. Please move on.
09.10.03 – A lot of thoughts
I have accumulated a lot of thoughts on the subject of font development in the last month. There are some things that bother me that I think I'd like to address. However, don't expect some formal, professional opinion—I will state my thoughts from a very personal point of view and it does not mean that my perception of reality (whatever that is …) is going to be similar to that of others (other typeface developers).
The (digital) typeface/font community is fucked-up. I just don't like it as a whole. I don't know … maybe it's because it's so small that it feels like an elite clique. Of course, it could just be me, as I seem to have hard time connecting with people and eventually always find myself in solitude, feeling like an outsider, but this community is just … too fucking baroque. Sorry, but I can't find a better word. I mean, there are nice people out there who do whatever it is they do with typefaces, but as a whole it doesn't work. For me.
Maybe because it's a mix of three very different ones: graphic, scientific, and corporate.
But enough about that. I not really in the mood to pointlessly bitch about something pointless.
As I wrote earlier, there is one FontLab tutorial on the web. Since FontLab has become the best tool to do the job in recent years, I think it's a hint that the community isn't what it's out to be, especially when that one tutorial sucks. It feels as if it was written by a dying dinosaur. Maybe I should write a tutorial myself …
Speaking of dinosaurs, I think there are too many of those and too few of young type designers. I mean, young and serious type designers like me. I actually don't want to be considered as such, but my point is that I don't see young people creating something other than mumbo-jumbo bullshit that no one in their mind should care about, or use.
Also, I'm tired of the dinosaurs who tirelessly talk about font piracy. If I ever meet one, I going to hit him/her in the stomach and politely tell them to shut the fuck up.
Now, about my typeface. It is obvious that my enthusiasm to work on my typeface has dwindled. This is not really surprising, at least not to me, because it has become quite usable to such a degree that the motivation to improve things has mostly disappeared. I think this is a very natural reaction when you are doing something as a hobby without getting paid. Once you are somewhat satisfied with what you have made so far, ironing out the details becomes somewhat boring. Unless you are a perfectionist, which I certainly am not.
I haven't really kept track of what I've done since my last update, so I'll just mention a few things from the top of my head.
As planned, I've been redesigning glyphs. This is hopefully going to be my final redesign. I am quite satisfied. I really like glyph “a” now. The Latin letters are almost complete and so are the Russian.
There a few designs that I have decided to change. I don't like the usual “g” design in sans-serif typefaces like Tahoma, Helvetica, and Lucida, so I thought making one like in Trebuchet MS and most other serif typefaces. In Russian, I've made an alternative design of “д” and “л”. I don't recall where I got the idea of the smooth corner (perhaps from some old bitmap fonts), but I really like it. I also did some research on the Euro sign and I like the end result.
While redesigning glyphs, I've been delta-hinting 12 PPM and 15 PPM sizes. With my Win2000 setup it's sizes 7 and 9 whatever (I don't know which measurement Windows uses). It looks fine.
I tried out MS Font Validator and did find some errors like points sitting one on top of another, but it reports error for every glyph that has overlapping outlines which are acceptable (and rightly so) according to the specs. The program also reported some minor errors that I don't remember right now and which I haven't bothered to iron out.
I've also tried my font on GNU/Debian. I guess it looks okay, but there is a general problem with Freetype rasterizer, even with patented algorithms turned on. Compared to what I see in Win2000, glyph diagonals are rendered worse. I think they have too much contrast …
There are two problems I have with hinting: 1) The diagonals get fucked, so I have to delta-hint many sizes or redesign the glyph (something I don't want), 2) Diacritics don't want to be aligned in composed glyphs.
11.09.03
I've read on Apple's AAT Font Quality Specification that delta-hinting doesn't work under transformations and rotations, so I'm having second thoughts about using deltas for contrast adjustment. I think I'll go through with delta-hinting of size 14 PPM to see just how legible I can make it, and then I'll try to rehint the font with normal instructions. I don't really care about transformations, but I'm not that fund of using deltas either.
Reading another page by Apple reminded me of using anchors for accented glyphs, so from now on I'll create anchors along with redesign of the glyphs. Untill now I've been using offset positioning of diacritics, but that page states that this method isn't very good at gridfitting, which my experience confirms. I'm going to try and see if I can get better results with anchors.
I've added a few links to resources section.
It seems Microsoft's Visual TrueType program is more powerful at hinting than FontLab. Maybe I should acquire it? They write that you can get it free of charge if you sign the license agreement and fax it to them. Well, maybe I'll do that at some point just to check it out.
Stubled upon Microsoft Font Validator and joined the required MSN community to get it. Waiting for my membership. I wonder if it can be of any use …
Found (incidently) an article on Polish diacritics. This will be quite useful when time comes to make those.
10.09.03
Continuing redesigning Latin glyphs with some optical illusions in mind that I've read about.
Switched design tactics. I'm getting tired of hinting and rehinting glyphs, so I decided to try focusing on designing nice shapes and not worry about hinting.
I've been often thinking about anti-aliasing at normal sizes. I mean, the whole issue is somewhat controvesial because some people find it to be distracting. I myself am still ambivalent about it, because I can't decide what I think is more legible for me. Clearly, the font can be made very legible without shades of grey, but that requires etheir drawing of bitmaps or very heavy hinting.
In everyday reading on the Web with my font, I've noticed that anti-aliased text is harder to read when the contrast between text and background is low. It seems that with my current hinting diagonals and round parts of the glyphs have too low contrast. This kind of anti-aliasing is obviously bad for legibility, but I think that text with subtle anti-aliasing is more legible than without anti-aliasing. Even though I've been trying to surpass legibility of Tahoma, I personally consider my font being worse because it's too fuzzy at the current stage. The only way I know to fix this is delta-hinting. I think delta-hinting is usually used for fixing rare anomalities that pop-up in some cases with normal hinting, or for aliased glyphs as it's the case with Microsoft's fonts, but I'll use it primary as a way of increasing contrast. I could also increase it by having contrast-friendly glyph designs, but I don't want to compromise on that because I want my font to be multi-purpose, unlike Tahoma/Verdana.
So, I gave delta-hinting a shot on size 12 PPM and I like what I see. This was only a test because I first need to finish designing the glyphs.
06.09.03
I've started redesigning Latin glyphs to be more round. It was planned from the very beginning to be done at some point, and right now I felt like doing it. To make round corners it's not possible to achieve good results using one off-curve point as I've been doing it until now, but two is enough. It's not very easy to find the right roundness moving two points, but I'm getting hang of it. The difference in glyph shapes is quite noticable at large sizes.
I also felt like having better contrast. I really like Trebuchet MS, so I kind of ripped of the way stems differ. I ended up reducing vertical stems of small letters, increasing horizontal stems of capital letters, increasing curved horizonal stems of both small capital letters, increasing bearings of capital letters. So far I think it turns out very nice.
I mark redesigned glyphs with magenta colour in FontLab. Blue colour is used for composite glyphs.
Other than with general roundness, I'm dissatisfied with many glyph shapes: JKMNRSVWfgjkrstvwy (if only poking in the Latin range). I've experimented a lot with “M” and “W” and came to the conclusion that these letters suck major ass when hinting. As for the others, they are not hard to improve, but I haven't had time to give them more thought.
01.09.03
I haven't done much progress other than a small amout of fine-tuning. I'm kind of getting bored and kind of have less free time.
I have tried to find good Cyrillic typefaces, both on the Adobe website, AGFA Monotype website and just on the web in gerneral. I didn't have much luck. A lot of artistic ones, but nothing legible.
I found a few articles in Russian on typeface creation, optical illusions, russification and such: 1, 2, 3. Some of these kind of articles radiate this “Oh, typography is art, you are supposed to appricate it …” attitude. I can't stand this bullshit.
25.08.03 - Languages
Ukrainian should now be complete. And also a few other languages I'll never understand. Added macron vowels for Japanese Hepburn romanization.
I've delta-hinted the smallest size (PPM 8) so that it doesn't look spaced out. I never thought that a font could remain so legible being so small. I personally don't need to have the typeface legible at this size, but I'm sure it may be useful in some circumstanses for some people (like PDA devices or whatever).
25.08.03 - …
Well, I've done a few things. I've increased bearing around glyphs, which is fine at small sizes, but the font gets too spread out at large sizes and becomes quite awful. I'll have to experiment with that more. The issue here is the hinting of bearing, because it kind of gets in the way, but without it the spacing between the glyphs gets screwed up.
I've also redesigned Latin glyphs with diagonal stems to look more like those of Lucida. They do look nice, but I'm somewhat hesitant about deciding on their legibility. Maybe I'll revert back to old style.
The typeface should now cover English, Danish, German, Russian, and French. I still haven't got around designing a few Russian letters because I haven't bothered to do reseach on them, and I also haven't finished making numbers. The lack of numbers doesn't actually bother me that much yet, because I don't look at numbers all day to get sick of the drafts. Of course I'll do them, but I'm not exactly eager to do the job (it's not easy making them “just right”).
Somewhat unrelated to the topic of typography, I've found a free keyboard layout creator program from Microsoft that does what it says. It's useful for two main purposes: 1) you want to have a Danish Dvorak layout (and a phonetically corresponding Russian Dvorak layout), 2) you want to be able to easily access some Unicode glyphs like em-dash (—), bullet (•), elipsis (…), correct quotes (“bla”), or whatever your bloody heart desires. The program itself is not as easy to use as Keyboard Layout Manager (which is not free and has some limitations), but it does the job. It's bloaty as hell, but you shouldn't really expect less from Microsoft.
What else? I've been playing a bit with 国字. I haven't done much actual designing, but I looked at the possibilities of creating something that resembles brush strokes in FontLab. It seems that the best solution is to draw a line outline with Pen tool and then use Expand Strokes function. However once you expand the lines, there is no longer an easy way to correct the overall shape of the glyph. A solution is to have the original lines as separate glyphs, but I don't like it that much. I haven't yet given any thought to the metrics of Japanese glyphs and how they should relate to metrics of European glyphs. Currently I'm trying to find out what some special Japanese symbols mean (from CJK Symbols and Punctuation Unicode range), because I for now stick to the goal of designing only the glyphs that I need, personally.
15.08.03 - …
Otlegacy tool would have been useful if it didn't fuck up my font. Will have to talk to the developer …
The most bizzare problem with hinting of my font: wide, symetric letters like “m”, “ф”, and “ш” get cut off at some sized. I don't want to delta-hint them but currently I don't know any other way to fix this. Something tells me that fooling around in TrueType Stem Control window will help.
Okay, the FontLab manual explains how to autohint TrueType fonts. It's hidden away in Transform Range function.
Now I do understand what device metrics (hdmx and vdmx) do.
14.08.03 - “Argh!”
If you press Ctrl+R in FontLab, … Don't. And don't save your file after that.
Design status: Since last report I've tuned the font metrics. The x-height is now larger and should make the font more legible. Fixed the ugly “s”. I've been playing with a combination of hinting and glyph design of some glyphs. There are some ploblems with hinting space around glyphs and mirroring, but it's not very important right now. I need to finish Russian alphabet, numbers and a few important diacritics. I think I almost settled on the design of Latin glyphs.
I had to disable generation of [vdmx] table because it resulted in too wide vertical spacing after I added a lot of high and low diacritics. TrueType-specific metrics had also to be set manually for best results.
I mentioned autohinting feature of FontLab a few days ago, but it turns out it only works for Type1 curves. You can probably dance around this limitation by converting curves back and forth, but I don't want to do that.
My list of bookmarks on the subject of fonts has grown alot. Will sort them out in resources section sometime later.
13.08.03 - “Aha!”
FontLab keeps crashing without CacheTT. I think I'll report the bug.
I've been wondering for some time why some of my glyphs that have the same height, e.g. E and H, when hinted, display with one pixel difference at some PPMs. It turns out that because I changed global caps height and changed the vertical positions of the top control points (together with alignment zones), the hinting align points were linked to an old alignment zone (or the wrong one, because I placed another alignment zone in it's place). So, if you are changing alignment zones of an already hinted font, delete the old align points and place new ones. Boring.
I've submitted my font to freshmeat.net today. This will surely get the word out in the Free Software community. Personally, I was quite disappointed with Vera typeface family from Bitstream, mainly because they didn't have Russian glyphs, but also because the design itself was to Verdanish for my taste. I hate Verdana. I tiny little bit. Let's leave it at that.
But it's the reason why I chose not to extend Vera. Vera; Vera, Verdana, hmmm … the standard convention of typeface design rip-off. I'm not impling anything, but this state of affairs amuses me.
Anyway, it's don't give rats about the community, but I'm going to observe the reaction with interest. Last time I spoke to RMS, he said … argh, I can't remember. He agreed that there is a shortage of Free quality font and welcomed my grand ambition to utterly destroy the hopeless situation of looking at on-screen text. Don't ask him whether he actually said that—he will deny it up and down; it's of no use to try. The sucker won't budge. Trust me.
Narrowed down the crashing of FontLab down to exporting VOLT data option. Disabled it, as I've no use for that right now. I do plan to play around with Volt later, if I have time and motivation. The problem with my motivation is that I don't need all those fancy OpenType features to enjoy reading the languages that I'm reading. Those features are mostly useful for typographic design or DTP, and those subjects doesn't interest me. I just want to read, and I want to do it effortlessly—without pain. I want my typeface to be transparent.
I want to say something else about Vera, and I'm going to be very honest here: Someone got paid for creating an ugly typeface. (Ugly by my own standard, okay? Now, piss off.) Someone else hyped it up. Many geeks rejoiced when it was released. The development seems to have stopped, the mailing list is dead, the long-promised design guidelines haven't seen the light. There seems to be a consencus that Vera is nice which means the majority of users have low expectations. All this doesn't amuse me. It makes me want to puke.
Bla bla bla …
12.08.03 - General punctuation.
I was doing general punctuation glyphs today. It turns out Trebuchet MS has an error: instead of “ it has ‟.
I've been doing more hinting today. I have discovered that glyphs made by mirrored components also mirror the hinting, which is expected, but because hinting isn't symetrical in most cases, it gets fucked up. This means that if you want properly hinted glyphs, you should decompose them and hint manually. Meep.
FontLab keeps crashing when I generate my font after working with it. I think this has something to do with CacheTT. This isn't fatal, but irritating.
There are problems with hinting composite accent glyphs because the components are hinted independantly and the summetry becomes difficult, if not impossible, to achieve.
11.08.03 - Design refinement.
Verdana and Tahoma have the same stem widths, but Verdana has wider glyphs. I like Tahoma better; Verdana looks a bit out of proportions.
I figured out (by a mere coincidence) how font substitution works in Windows 2000. I installed Summersby as my system font and made MS PGothic work as a substitute for all other glyphs that I haven't created yet. Both of these feats require a small portion of registry exploring and hacking. This substitution seems to be system-wide, as it works even in EmEditor.
Hey, this substitution is pretty cool. It should actually have been working with Tahoma, and it did, until I outsmarted myself by extracting MS PGothic out of msgothic.ttc and than deleting msgothic.tcc (I did it because I didn't like having MS Gothic, MS Gothic UI, MS PGothic in the font list when I only ever used one of them) at which point the substitution stopped working.
Found more forums: Font Tools Community and Microsoft WEFT users community.
After hinting glyphs, the flawed parts of the glyph design become more evident. So I've been looking through the designs of other fonts, mainly Tahoma, Trebuchet MS, and Lucida Sans. Legendum and Vera Sans don't have good designs so they aren't very useful in this case.
I don't understand what device metrics (hdmx and vdmx) do.
10.08.03 - FontLab forum.
Some interesting threads on the FontLab forum: OTF Russian and Greek and cedilla vs. comma accent and Edge sensitive kerning of OT Cyrillic fonts and substitue two specific characters by two other.
I did some further hinting and I also fixed some glyph designs. The typeface is actually beginning to look nice … so I've decided to put up some screenshots.
I'm pissed off at the lack of visual hints in hinted TrueType fonts. I don't know whether it's a FontLab limitation, but I want to see how the best fonts are hinted. By the way, now that I've got hang of hinting, I think I should try playing with Autohinting function.
Current issues with my typeface design: the stems are a bit too think, the spacing is a bit too wide, hinting isn't all that good at small sizes (in black & white mode it's downright ugly), x-height is a bit too small.
Very soon I'll start using it as my main system font to motivate myself making it better.
10.08.03 - Just for the record.
I hate most typefaces with passion, together with their creators. The world has enough useless, broken, ugly fonts already.
I found Ray Larabie's FontLab tutorial. I think it's the only FontLab tutorial on the Web.
09.08.03 - Playing with shapes, hinting.
Okay, I played with normal glyph shapes trying to get hang of using TrueType curves and control points. So, I recreated some glyphs with a simple design, using as few control points as possible. They turned out okayish, but far from perfect.
After a while I recreated the basic Latin alphabet, caps and small letters. I started out with caps, so they ended up being uglier and less consistent. Both vertical and horisontal stems need ajusting. The thickness is the design is a bit much for a normal font.
I had the most difficulty designing letter “S”. While designing the glyphs, I used Tahoma, Threbuchet MS, and Lucida Sans as references. I was looking both at the shapes and at how the TrueType curves were used.
Then I played a bit with spacing. No real problems here, and not really important at this stage.
I added some diacritical marks that I need for reading the stuff I'm reading, i.e. I made ring which I need in Danish (å) & breve which I need in Russian (й). The problem is now how to automatically generate composite glyphs. I still havn't figured out how to make Generate Glyphs function assign Unicode code points to the composites. Sure, I can do it manually, but doing it for every font is masohism. There are OpenType features that can do combining, but I haven't looked into that yet.
Having made some basic shapes that didn't look like mousewriting, I thought it's time to learn hinting. So I read the FontLab manual and hinted some glyphs. It's easier to do than I thought at first, but some interesting issues with “i” and “j” did crop up. The dot in those letters doesn't alway stay right above, so there are still many techniques to be mastered … interpolation of key points seems to be the solution. Anyway, I haven't figured how to do global hinting, for the font as a whole. I'm not sure it's possible, but the FontLab manual mentions it together with Alignment Zones. So after my first encounter with hinting, I get the impression that after some getting used to, it's easy but time consuming and somewhat boring.
I've also read the papers by the developer of Gentium typeface. See resources page with links. Interesting stuff, will come in handy at some point. Not now though, as I'm still making rought strokes and am not focused on the details.
I figured FontLab must have a good forum somewhere, so I found it. The bastards are on MSN. I don't want to register with Microsoft to participate. So far potentially interesting or useful threads: Generate Glyphs tricks and Working with different types of kerning and Italic tricks and Extended characters in kerning, preview and hinting. Now I'm reading the relatively high traffic general section of the forum …
21.07.03 - Resources.
Well, in the last two days I've done more browsing around the Net. I've found a forum for typeface developers. To be frank, they do sound like a snobby bunch most of the time, but I'm not really complaining as I haven't seen other such forums. Anyway, it might come in handy.
Someone might get the impression that I'm even snobbier than them—that's true. I'm the epitome of snobbiness … but! I'm not a bunch; that's always a plus.
It was indeed handy. Someone gave a link to an interesting article on designing кириллицу in three parts 1, 2, and 3 (in Russian). Something similar for 仮名 and 漢字 could be extremly useful.
19.07.03 - Кириллица and beyond.
I wanted to sketch some Russian glyphs, so I thought I'd assign some code points to Latin glyphs like A, O, P, etc. This was quite easy to accomplish.
Then I thought it would be a good idea to see if it is possible to automatically generate composite glyphs like á, ó, é, ú, etc.—
Currently I'm using Tahoma as a general reference to see how the font is contructed. In Tahoma those Russian glyphs are made as composites instead of just being referenced by the code points. Making them composites gives more possibilities, but if the glyphs are exactly the same, I think it's better to just add multiple code points.
The possibilities of composition are mainly moving, scaling, and mirroring (besides the obvious composition of several glyphs into one). I created Russian И and Я by flipping the composites of N and R. If memory serves me right, typographycally this is probably not such a good idea, but because I'm just sketching things and exploring the possibilities, this will pass.
Doing all these rough editing could be nicely automated for the other fonts. Right now it's not very important, but later, this progris riport can come in handy for exactly that purpose.
Getting back to automatic composition of áóéú. I think it's best to define so-called anchor points to tell the program where exactly to put composite glyphs. Then you can generate new glyphs, but they don't automatically acquire Unicode code points. Will have to look into this later.
Drawing the rest of Russian alphabet, I found out that you cannot make composites out of other composites (I was trying to compose Й).
FontLab has crashed once yesterday and once today. Not good.
I'v been browsing GNOME font mailing list—it has some information about Bitstream Vera typeface. The list seems to be mostly dead. I sent an email with some questions.
I discovered that some glyphs (from Latin-1 Supplement: Ã, Ô, etc.) in FreeSans font were not composites when they should have been.
I rediscovered the page with some useful OpenType tools. The guy does some nice things with the OpenType features in his fonts and has a tutorial that supplements the FontLab manual.
18.07.03 - The sketch.
I have uploaded it to my typeface section, so go have a look at the monstosity.
I wanted to experiment with assigning multiple Unicode code points to the same glyph, so I downloaded some charts. There is some fascinating stuff in them.
Also found a few interesting FAQs there, starting with Ligatures, Digraphs and Presentation Forms. So I made fi ligature and experimented with capabilities of OpenType ligature substitution. I made it so that it's only substituted automatically in the beginning of the word, because the in the middle of the word it can be a no-no if the word is a composite where f and i are not supposed to be tied (I don't remember where I read that). I tested this feature in EmEditor (the Unicode editor I use to edit these pages and everything else), but the substitution didn't happen, however it did work in Photoshop 7. The only other font I saw do this kind of substitution is Sylfaen, but it substitutes all instances which can be unfortunate (also only seen happen in Photoshop). The last paragraph in that FAQ should be noticed.
Oh well, stumbled upon Proposal to encode Klingon while browsing unicode.org. It got rejected—no shit!
17.07.03 - First steps.
Like I said in my blog, I've started reading FontLab manual. Yes, I have decided to make the typeface in FontLab. It's actually not a big deal, as I have read that both PfaEdit and FontLab don't modify/delete hints of untoched glyphs. The whole hinting process seems rather complicated after reading the manual, but I'm definately going to hint the fonts.
I am quite uncomfortable about drawing with splines (or whatever you call it) and control points, and while the manual is very good, it lacks some step-by-step examples. So it will take some time for me to get used to it.
I know there exist some free fonts like FreeSans and Bitstream Vera and Dustimo that I could use as a starting point, but I have decided to create my typeface from scratch and only use the other fonts as references. Since FreeSans also strives to be a Unicode typeface, I might copy-paste a few things in the future.
I only managed to make one glyph—“A”. I tried to make “O”, but it was harder than I thought, so we can say that I didn't make an O but a ring. It's 駄目.
Facing such simple difficulties, an idea of making a rough sketch of all the glyphs that I need popped up in my head. I still don't have a particular design in my mind, other than sans-serif, so I think it's the best I can do right now.
The license is most likely going to be GPL.
Another problem I'm facing is naming the typeface. I haven't fount any good names yet. For now I'm calling it Summersby.
