FontForge as a Typeface Design Environment--Good Idea, Bad Idea

Steve Tiano's picture

I've been having a running conversation on the Design forum about my burgeoning interest in typeface design--I am a book designer and page layout artist. Initially, of course, FontLab Studio comes up as the software package of choice for professional typeface design. Some have said, Type Tool is also something to start with and then, as necessary, make the switch later to FontLab.

Some mention was made of the open source package FontForge. I'm on a dual processor G5 PowerMac running OS 10.4.10. After understanding what was necessary to get the X11 server package installed, it was an eay matter of installing and running FontForge, even though it is UNIX-based and I am not a techie.

But I was also told "given that there is a learning curve for any of the programs," I "risk wasting time by going with [FontForge]. ... some software people use it, but not the design community."

How do folks here feel about that? Is it accurate? Should I just stay away from FontForge, start with TypeTool if the investment is presently too much to get FontLab (but assume that professional typeface design will eventually mean moving to FontLab Studio)? Does anyone do viable, professional creative work in FontForge?

Thank you.

dan_reynolds's picture

I vote for FontLab. I know that it is expensive, but most professionals use it, and its development is directly influenced by feedback from said professional designers. I've been using it for about three years, and I love it.

Gary Long's picture

I bought Typetool 3 to discover if I had the stamina to design and produce a complete typeface, and found it surprisingly good considering the modest price---easy to learn, decent tools. I will move up to Fontlab in due course, but if you don't need Opentype features, it might be all you need, especially if you're making type for your own use.

twardoch's picture

Gary,

also, if you use Windows, you may install the free Microsoft VOLT or the free Adobe FDK for OpenType (AFDKO) to add your OpenType features. A combination of TypeTool 3 + VOLT or TypeTool 3 + AFDKO will actually let you build professional-quality OpenType fonts. Since TypeTool 3 is based on the FontLab Studio 5 core, the technical quality of the fonts made in TypeTool 3 will not be any less from those made in FontLab Studio 5 (esp. if one only uses autohinting in FLS5 rather than manual hinting, which most people do these days -- FLS5 lets you do both auto and manual, and TT3 only does auto, but uses the same algorithms as FLS5).

You can find the links to the software I mentioned at
http://www.font.org/software/

Regards,
Adam Twardoch
Fontlab Ltd.

blank's picture

I tried using Typetool, and while it is a great program, I was really just using it to compile fonts that I was building in Illustrator due to some of the Typetool drawing limitations. But I find that drawing letters in Fontlab is often easier, and keeps me from running into the performance and file management issues that come along with keeping the letters in Illustrator. This week I decided to just move on and upgraded to Fontlab, which wasn’t to hard on my wallet now that Fontlab offers academic discounts.

Something to keep in mind about Fontforge is that, being an esoteric open source program, the amount of quality help and support that you’re likely to get for it is something like nil. If you have trouble with Fontlab, you can just post here and eventually one of the Fontlab programmers will probably answer your question.

cuttlefish's picture

Have you who have posted so far really given FontForge a fair shake? I'm transitioning from my decade old copy of Fontographer to FontForge, and while it hasn't exactly been smooth, the tools in FontForge offer a lot of significant improvements. As far as technical support, if you sign up for both the users and developers (for bug reporting) mailing lists, you have direct access to the program's author who takes very quick action to address any problems, and if it takes altering the program to effect the fix, he'll do so without hesitation (first with a patch, and at least once a month with a full build).

anagnost's picture

Steve,

First of all, I don't think learning FontForge can be considered a waste of time even if you decide not to use it for your actual work. In general, I am sure that having some experience with "alternate" applications is always extremely useful, because it allows you to understand the difference between the peculiarities of your application of choice (e. g. Word or FontLab) and your general task (e. g. word processing or font design) and thus to use your favorite application more effectively.

Regarding FontForge itself: I used it for several years, and now I am also participating in its development by submitting patches. I have also some experience with FontLab, and I am sure that FontForge is better at many aspects. In particular, I heard many font designers (especially those who had previously got familiar with Fontographer) to complain that FontLab is very inconvenient as a spline editor. But drawing glyphs with FontForge is a real joy, so that no external tools are required. Just one example: FontForge has an option to mark extrema points and points of inflection on the spline. Thus if I see that handles of some of my points expected to be positioned at extrema are not exactly horizontal/vertical I can easily correct this, while in FontLab I would have to run FontAudit to notice this problem. BTW, I often see slightly distorted extrema even in costy commercial fonts, and I am sure this would be not possible if their designers had used FontForge.

Furthermore, FontForge is much more powerful than FL in designing OpenType layouts. In particular it allows to define anchor points for combining marks ('mark' and 'mkmk' features) which neither FontLab nor AFDKO support. Well, most designers currently do that with VOLT, but VOLT is only for TrueType fonts, while FontForge supports the OpenType-CFF format as well. FontForge also has a nice intuitive GUI for adding OT features, which is much clearer for me than feature files with their obscure syntax.

I should mention also the main (to my mind) FontForge's disadvantage, which still makes using it for production quality TrueType fonts a bit difficult. FontForge's TTF autoinstructing code is really bad, and the developer was not interested in fixing this problem for years. However, things are changing now, as there are volunteers working to improve this functionality. On the other hand, FontForge already may be very useful even for creating/modifying TTF instructions: first, it allows to modify them directly (unlike FontLab), and second, it has an excellent TTF debugger which is extremely useful for anyone who would like to understand how exactly his/her instructions work.

William Berkson's picture

>in FontLab I would have to run FontAudit to notice this problem.

You don't have to run FontAudit; you can just leave it on, and it will mark extrema, non-vertical handles, etc.

Steve Tiano's picture

Thanks for that response, anagnost. That's kind of what I've been hoping to hear.

As I'm really the most elementary kind of a novice at this, I can't really draw a lick, and I'm seeing this as a long-term project, I'm actually starting by reading some history of type stuff. I know the typefaces I like the looks of; now I want to see how they fit in to various ways of categorizing them. And then maybe the next step will be to start sketching by hand some characters. After that, I imagine, it will be time to bring FontForge into the mix.

I'm sure I'll be keeping this forum informed.

blank's picture

Aganost, based on your post, I tried installing Fontforge and firing it up, something I haven’t done in a few months. Being on a new laptop, I had to install the program fresh. It does not, however, start, as one of the libraries is missing. So now I am faced with two options: just stick with Fontlab, or uninstall Fontforge and start installing older Fontforge packages until I get one that works.

And now I remember why I don’t bother with free software. It’s only free if my time is worth nothing.

anagnost's picture

> So now I am faced with two options

For me, the most obvious option would be just to install the missing library...

> It’s only free if my time is worth nothing.

You should understand your problem has nothing to do with free software. Many open source applications have been successfully ported to Windows (or MacOS) and can be installed and used just like any other native programs. But FontForge is a Unix application, which means you need some Unix knowledge even if you are running it in your Windows environment under Cygwin (I suppose the same would be true for most X11 applications under MacOS). Nobody never promised that installing a Cygwin application would be as easy as a native Windows application, but having a Cygwin port is just better than nothing.

So if you are not prepared to some small difficulties, which, unfortunately, are unavoidable when you are running an application written for another platform, then, indeed, it is better for you to stick with your favorite software. Otherwise I am sure your problem can easily be resolved. But please don't blame something you haven't even tried.

aluminum's picture

I have used neither. I've only use Fontographer, and ANYTHING has to be better than Fontographer.

I'm about to start getting a bit more serious about trying to create some typefaces and I'm likely going to go with Fontforge. Not because it's better but because it's open source.

I've been trying to move that way for a while. Use the Gimp as much as I can (though still love Fireworks). Learning Inkscape as I go (but holding on tight to Freehand). Etc.

So, that's just me. I like using Open Source when I can.

That said, if I were doing this professionally and time was money, then I'd probably break down and use both apps as needed. ;o)

I will say the 'learning fontforge is a waste of time' argument is silly. Using software isn't rocket science. What's important it to learn what the software DOES, at which point the particular flavor of software you are using is rather moot.

aluminum's picture

anagnost...thanks for working on FontForge! As a lowly user of open source software, I appreciate your efforts!

"And now I remember why I don’t bother with free software. It’s only free if my time is worth nothing."

I think that's MS's memo on the subject, isn't it?

As a comparison, we just purchased a quarter-million-dollar license for MS Sharepoint here at work.

Documentation!? Ahahahahahaha!

Support? Sure! Just pay us 70k a year! Ahahahahaha!

Point is that the quality of support for software really has no direct correlation to the price (or lack thereof) of the actual software product.

A lot of open source software is a pain to get running. So is a lot of commercial software. ;o)

blank's picture

ut FontForge is a Unix application, which means you need some Unix knowledge even if you are running it in your Windows environment under Cygwin (I suppose the same would be true for most X11 applications under MacOS).

I’ve been running on some variant of UNIX or Linux, from Solaris to OS X, and putting up with free software, for most of the last decade. And I don’t deign to work on Windows, much less Cygwin. And yeah, I could just try to track down the missing library and try to get stuff working, but this sort of stuff is why I just stick with commercial software most of the time. I’m sure that Fontforge can be a great tool, but I can’t begin to describe how many hundreds of hours I have wasted dealing with some little thing here or there when it comes to free software, which is why I jumped ship from Linux to OS X and eventually stopped using most free software entirely.

William Berkson's picture

>I will say the ’learning fontforge is a waste of time’ argument is silly. Using software isn’t rocket science. What’s important it to learn what the software DOES, at which point the particular flavor of software you are using is rather moot.

As I made the argument that learning Font Forge is a *risk* of wasting time, of course I don't think it is a silly argument. If your main interest is in making fonts, not learning software, then you are going to want to minimize your time away from design time. Designing fonts is extremely time consuming, and for me at least learning FontLab and getting to use it with some fluency has also taken time, and I'm still learning.

FontLab is not the most complex application--from what I hear Pro Tools for sound engineering is way more complex--but it does take some serious time to get on top of. Given the complexity of font technology now, I can't imagine that other options are a lot simpler, when you include interpolation tools, including multiple masters, as well as spacing, kerning, hinting, and open type programming.

Also, the idea that what software application you use is just a 'flavor' and 'moot' as a general proposition does not seem to me defensible. Publisher and InDesign are both page layout programs, but I don't think even MicroSoft would say that Publisher is just as good for full-on professional work.

I haven't learned other font design applications, so I can't compare their merits. But I can report that at TypeCon earlier this month there were lots of workshops on FontLab, some using the revived Fontographer, and none on Font Forge. I never even heard Font Forge mentioned. I wouldn't be surprised if there are zero professional type designers who use it, judging by conferences and Typophile discussions. In any case there are certainly very few among those doing latin scripts.

If Font Forge is actually distinctly better, and you can do the same stuff with it, or with it plus other usable programs, then it would absolutely be worth learning and using. Otherwise, I don't see how it's a good investment of time.

As font design is very time consuming, if Type Tool and the full Font Lab are actually better than Font Forge, then worrying about the $99 seems to me like really false economy. I realize that in poor countries this might be a lot of money, and the money might be a significant factor. But in rich countries it's not.

To me the main question is really how good Font Forge is, not the cost of Type Tool.

aluminum's picture

"To me the main question is really how good Font Forge is, not the cost of Type Tool."

Yep, and that's a valid and pragmatic approach.

I'm not saying FontForge is better or worse...just that it's not necessarily a waste of time. Just learning Fontographer taught me specifics about font creation that I can apply to any other font creation software.

Personally, I'm just trying to wean myself from total dependence on Adobe. That's just a personal thing, more than anything. ;o)

William Berkson's picture

>wean myself from total dependence on Adobe

Just to avoid confusion: none of the available font drawing applications is owned by Adobe. Adobe's FDKO set of utilities for font development is free.

aluminum's picture

"Just to avoid confusion: none of the available font drawing applications is owned by Adobe."

Understood...just saying why I like to try opensource when I can as 90% of a visual designers toolbox these days is owned by Adobe.

anagnost's picture

> If Font Forge is actually distinctly better, and you can do the same
> stuff with it, or with it plus other usable programs, then it would
> absolutely be worth learning and using.

William,

I appreciate this approach very much. What I really don't like is arguments like "nobody uses it" or "it's an esoteric open source software". You should try it before criticizing it. Or just browse the online documentation at http://fontforge.sourceforge.net/overview.html and see how powerful FontForge is.

And yes, I can easily believe no professional designers are using FontForge, but this proves nothing for me. There are several known reasons (I would not like to discuss them here) for which most folks concerned with graphics and design are not familiar with Unix and Unix software. I also think the problem with TTF instructions I have mentioned previously could affect FontForge's popularity too: of course it would be not a real problem for those who have access to VTT, but... there are so many "professional" designers who prefer just to export an autohinted font and then claim their production is perfectly hinted. Unfortunately, this is impossible with FontForge (but probably will be possible soon).

And, finally, there are very few real professionals in font design anyway (I believe most of them can be found at this forum). I know enough font foundries which promote themselves as "professional" but actually sell so poorly designed amateurish fonts, that the fact they use FontLab really doesn't matter. On the other hand, I can list several open source font projects implemented at a high professional level (my favorite is the Junicode font) which use FontForge. I think their designers can be considered real professionals, even if font design is not their main business.

You can have a look also at my own Old Standard font. If it has any flaws (and of course it does), FontForge is surely not responsible for them.

Steve Tiano's picture

Well, I am by no means a techie. And I found that once someone told me in two or three straight declarative sentences how to get X11 off the systems disk, I had no problem installing it on my G5 PowerMac. And FontForge downloaded and installed easily.

I can see where this discussion's gone already, so I hesitate because I don't want to start a whole platforms war. I just want to say that, in my experience--as a non-techie--that's why Macintosh rocks and has been my choice 17 years: it's easy to get going and just start learning how to do the work you want to do. Rather than learning how to use the computer, and install and configure jazz. I know the Windows machines at my day job are constantly down or slowing down--even just for minutes can be a drastic inconvenience.

That said, since this is simply a long-term project and book design/layout is really my work, I gues FontForge is definitely a reasonable choice for me at this stage.

dan_reynolds's picture

William points out above that FontLab is NOT owned or distributed by Adobe. I use Adobe products every day, but part of me can understand why some designers might want to "diversify" their software portfolio, so to speak, and use products from other companies. No problems there.

What makes me wary about a lot of OpenSource-talk is the "values" that are often implicit, and which are so often brought up. Values? I guess that I just have different values.

FontLab is a small company. I don't know how many employees it currently has, but I'm thinking 10–15, tops. And that is spread across multiple countries and continents. I don't think more than a few ever work in the same room.

FontLab, and other companies like it (inclucing many in the type design industry), IS the little guy. They are the people on the forefront bringing technology forward, and trying to make a living from it, too. I think that that is worth supporting. Here is why my values come into play ;-)

Personally, I think that it is more worthwhile to support these 10–15 guys spending their working careers trying to make a great product than to support some vague, hobby-level "opens ource" movement. Especially if I'm making a living using these products.

aluminum's picture

Yea, I understand FontLab isn't adobe and have nothing against FontLab. They seem like a great company and would say that supporting them is likely a good thing.

Just saying that I like to try to use Open Source when I can...even when it has to be alongside its commercial brethren for the time being.

"Values?"

I think the main value I subscribe to is just diversity. It's good to keep alternatives alive out there.

"Personally, I think that it is more worthwhile to support these 10–15 guys spending their working careers trying to make a great product than to support some vague, hobby-level “opens ource” movement."

Personally, I think it's important to support both.

blokland's picture

'[…]But I can report that at TypeCon earlier this month there were lots of workshops on FontLab, some using the revived Fontographer, and none on Font Forge […]'

The fact that an application is the choice of the majority does not per definition proof that it is the best tool for every purpose. For instance Windows is also the choice of the majority, but the operating system does not seem to be very popular in general on this forum (which as such is interesting, if only because of the fact, that the Windows platform represents the largest market for font producers and besides that, Microsoft has done a lot for both font technology development and the font community in general [like sponsoring events]).

I don't know if DTL FontMaster was mentioned at TypeCon, but presumably not. That does not mean that it is not a good, or perhaps for certain tasks even better and more powerful font production tool than any other available on the market. FM was simply not developed for the 'mass market' (nor is our organization very much suited for extensive [low level] customer support).
I don't expect our customers to attend workshops either; the most work on a large scale and are very capable of solving or circumventing inevitable font production related problems themselves. But sometimes even FM users come to us with unexpected problems, like for instance recently with a request for support for glyphs with higher Unicodes, i.e. not in the BMP, which was added subsequently to the program then.

If one wants to judge a font production tool, in my opinion at least two parts of the production have to be separated: first, the conversion from design into outlines and second, the conversion from the outlines into font formats (for a more detailed separation see: http://www.fontmaster.nl/pdf/FM_docs/Automating%20using%20FM.pdf
Concerning the first part generally a couple of Bezier editors are comparable when it comes to functionality (although not in detail, of course), but when it comes for instance to manually digitizing this functionality is only supported by DTL IkarusMaster.
The generated font formats must be reliable, according to the specs and the results reproducible. Concerning the latter, because OpenType layout features files can be used in batch in FM, all GSUB features supported by the AFDKO will be added automatically to every font that has applicable characters. This way it is not only unnecessary to do the same work for all the 'common' GSUB stuff, but it also assures consistency throughout a library (which is important for the success of the OpenType format itself). It actually does not make much sense to me to do this per font.

By the way, I don't think that the platform on which an application is running, should be an issue; if only for the customer support for the generated fonts one must have knowledge of and control over at least the mainstream operating systems and applications anyway.

The Ikarus based file structure of FM was developed for batching (I will show this in detail at the ATypI TypeTech Forum at Brighton). For the most needed functionality, like merging glyph data, generating composites, checking glyph data, generating fonts, et cetera, extensive batch functionality is built-in and mostly controlled by simple command files.
If one wants to make full use of this batch power, for instance for generating all current font formats for all typefaces in a library in all applicable code pages at once with the call of just a single command file, some reading and good organizing is necessary. But not everybody wants to draw letters on paper and subsequently manually digitize these, or needs the batch power, or want to spent more money on a font production tool, or likes a modular structure, or is willing to use a tool that is not the choice of many colleagues (which means less direct assistance from others in the field, or perhaps even no workshops at for instance TypeCon [although this can be organized on request, I reckon]).
Others only want to work with apps for Mac OS X, or want to add more or less comparable batch functionality themselves through Python scripting, because it fits their purposes better or perhaps because they just simply like to do it. Maybe some will wait for certain already available technology to appear in some way in their favourite application(s). And there will be probably plenty of other rational and irrational reasons and explanations for the use of certain tools.

In short, it is a relatively small but complex, diverse and interesting market for font production tools. Objective judgement of the offered technology looks even more complex.

Dan Gayle's picture

One thing that constantly nags me is the design environment itself in Typetool/Fontlab. Some people complain about the use of the tools themselves if they come from an Illustrator background, but I'd honestly rather have the look be more similar.

The handles and all the little controls and everything just seem too small. I can't seem to see exactly what is happening all the time because the controls are too small. I honestly like the "feel" of FontForge for drawing glyphs if only because it's easier to see what I'm doing.

And I'm not blind, by the way, I just want to see what I'm doing better.

William Berkson's picture

>I just want to see what I’m doing better.

In FontLab you can make the size of nodes big or small, colored or black & white. Go to Tools/Options/Glyph Window/Appearance and check or uncheck the appropriate boxes.

With all due respect, when you tell me you prefer FontForge, but don't know the FontLab interface, I am not moved to try FontForge. When I hear of experienced type designers singing the praises of Font Forge, I will start to take notice. Personally I am skeptical that FontForge can be that good without a base of professional type designers interacting with it, and the programmers to respond. But I'd be glad to hear more.

Mr. Blokland:
I didn't mention FontMaster as this thread was started by a question whether to use to use a free program, rather than TypeTool, which only costs $99 US. FontMaster Utilities are considerably more expensive than the full FontLab. And from the description on your site--and here--they have seemed to be targeted more to foundries than individual type designers.

I have heard about the FontMaster utilities, but not a lot. It was mentioned here on Typophile that Miguel Sousa at Adobe was impressed by the Kern Master. And the very high quality of DTL fonts of course speaks well for the tools. I don't think many people have seen your tools demonstrated, so I would think for most they are a question mark.

raph's picture

Steve Tiano: in answer to your original question, yes, FontForge is a perfectly viable tool. I use it along with my own Spiro tools for the fonts I create. But I'm very strange: to me, reading the code and understanding the way the geometry happens under the hood is very important. I'm also quite capable of fixing things myself when they're broken, and indeed have sent patches back to George.

It's also more than a bit frustrating and maddening. The UI is by no means up to the standards of modern apps, even other free software ones. Do not generalize from FontForge to the rest of the free software world. For some things, such as production operating system kernels and compilers, the open source products leave their proprietary competitors in the dust. In other categories (databases, for example), open source has staked out the low end (MySQL, Postgres), while leaving only the very high-dollar end (Oracle) to the proprietary players. The font software business is totally different because the market is so small. There just isn't a large population of people who both care about the quality of the tool and are able to make big improvements to it. Unlike, as I've mentioned, the situation with compilers.

But I find myself agreeing strongly with the main point I see articulated on the thread. What's really valuable to the type community is diversity. A few years ago, it wasn't clear there would be any viable font editing programs. Now even the venerable Fontographer is back, and people who do prefer free software for whatever reason also have a usable tool. People on the high end can use the IkarusMaster and family. This is all good. If I were setting out to make a living being a professional font designer and nothing else, I would no doubt just go with FontLab and be happy with it.

It will be interesting to see which, if any, of the major font platforms have the flexibility to adopt the Spiro technology, which is another good test of responsiveness. I have no time right now due to personal stuff (I got to spend three hours today with a lawyer, yay), but when I do get time again, I'm going to try to get it out there as a Web app, which may turn out to be an important ecological niche. (also, whether a new program succeeds in that niche probably depends more on how important web collaboration turns out to be, and whether the program is well done, much more so than whether it's based on fresh or stale technology, but in any case it ought to be fun to do.)

k.l.'s picture

Hello William -- FontMaster Utilities are considerably more expensive than the full FontLab. And from the description on your site—and here—they have seemed to be targeted more to foundries than individual type designers.
I don’t think many people have seen your tools demonstrated, so I would think for most they are a question mark.

Since the original question was about an inexpensive tool for the first steps into type design, the answer may be TypeTool or FontForge. (I never used FontForge because it requires X11.) For the beginning, I think, the simpler the font editor is, the better. This allows to focus on the essence of what, referring to the technical side, a font is: (a) a collection of glyphs (b) that have space to the left and right side and (c) possibly may be kerned.

As regards FontMaster: Today one doesn't need the entire FM suite to generate font but can do this e.g. with the BezierMaster module, which is about the same price as, or less than, FLS.
Yet the interesting thing about FM of course is batch font generation and the way it deals with OT features. This may not be required for single style fonts, not even for four-style families, but as soon as just a few more weights are involved, it can save time and trouble since one doesn't have to care for every technical setting in every single font. (See the other thread about the horror story cause of a missing hyphen in one style's font name.) I touch a few similarities/differences between FLS and FM here.

As regards other aspects touched in this thread:

As said earlier, FLS's popularity has the effect that people can easily exchange data or tools, and people can help each other. However: All font editors allow to import/export PST1 (or even OT) fonts. So no font editor is an island.

I see two important factors:
(1) The designer must feel comfortable with the drawing tools since she/he spends a lot of time with them.
(2) What matters in the long run are the OTF/TTF files which are finally produced.

When drawing, use whatever tool you like, can afford, etc.
Production is another issue. You can use any of the mentioned font editors to produce fonts. Some examples: FLS only. FLS + AFDKO2, in case you plan to do things that the older AFDKO-version does not support. TypeTool/FontForge + AFDKO, as the least expensive combination. And why not draw in FLS, TypeTool, FontForge and produce with FM, or draw in FM and produce with FLS? Whatever gives the desired results.
When producing fonts, you'll sooner or later change the point of view anyway. You will not ask any more whether it's better to check this or that box in the FLS or FM interface but make up you mind about how the font file and its tables shall look like and wonder, how can I achieve this with the tools at hand? And then it ceases to be an either-or and will turn into a both-and as regards choice of tools. For example, currently my fonts consist of some tables generated by FLS, others generated by AFDKO2, plus additional table tweaks done with FontTools/TTX.

As an aside, even unconventional combinations are possible. InDesign is a good kerning interface, and FontLab Studio and KernMaster work together nicely.

Karsten

William Berkson's picture

Thanks a lot, Raph and Karsten, that sheds a lot of light on the subject.

>It’s also more than a bit frustrating and maddening. The UI is by no means up to the standards of modern apps, even other free software ones.

Raph's view here I think confirms that for a beginner the $99 for TypeTool is a better option than Font Forge, unless you are a programmer or the price is prohibitive.

On the Pro end, my impression is that many, like Karsten, use multiple tools to produce their fonts.

anagnost's picture

William,

what exactly in Raph's comment convinces you to think so? The only negative moment he seems to stress is the fact that FontForge's UI doesn't correspond to the standards of modern apps. Well, I would call this rather an advantage. I never could understand why MS and other large software developers (even open source developers) change the UI of their applications in each new release, although an ideal of good appearance was already achieved in early nineties. For me old style interfaces (like in MS Office 95 or in some old Motif-based applications for Unix) look much more solid and elegant. And indeed I like traditional FontForge's UI very much.

typovar's picture

My experience with FontForge is positive. I suggested a simple improvement in the UI and within a few days it was solved. There is a lot of tech-talk in the mailinglists, but if you want some special feature added to the programm, it is possible.
I'm no programmer, but being able to look underneath gives me a feeling of “keeping in controll”.

Steve Tiano's picture

Well, I again need to thank yo folks for some enlightening discussion. Believe it or not--partly due to being quite busy; unfortunately, not all of it busy-ness connected to making books--although FontForge is now safely installed and opens just fine, I haven't even had the time to do a tutorial with it. I am beginning to get some sense of the sketches I'd like to start making of letterforms.

Would it be advisable to begin sketching on paper or just start up FontForge for a tutorial and see where it takes me? Or should I do some sketching in, say, Adobe Illustrator?

cuttlefish's picture

Definitely sketch it out on paper.
Many designers do start drawing directly in the font design programs, but sometimes those results can wind up looking overly mechanical unless substantial care is taken (see some of my own early works).
Really, though do whatever you're comfortable with.

Gunnlaugur SE Briem has some handy notes on type design, including workflow and tracing sketches, none of which are tied to any particular platform (except maybe the bits about autotracing)..

typovar's picture

What would you like to do first?
Learn how to use FontForge or create a typeface?

Maybe start up FontForge and try it, is a reasonable thing to do if you want to learn how it works. Once you get an idea of it, you could start creating something new. Then starting with sketching on paper is the best to start with.

It all depends on 'what you want'.

Daniel Denk's picture

The sole problem with Open Source, has always been that the only software that is truly embraced -- is the software that has the most thriving and active community.

In regards to commercial font creation tools -- there are really only two company options available that are worth considering.:

- FontLab
- Font Editor

FontLab has a more robust mindset in the professional paradigm, and taking over FOG was a good move. BUT - being a Windows user, I am still a little miffed that there hasn't been a priority placed on getting the memory architecture upgraded and patched for Windows.

Font Editor reminds me more of Macromedia when they started-out years ago. Macromedia were always the kid on the block that understood community support and the value in providing useful and educative resources to their users. Unfortunately, when Adobe took them over, some of that approach was washed in with the mix and degraded a little bit.

Look. The main issue here regarding any platform are two things, really.

1) Cost
2) Trainability

If you can't afford it -- chances are, you'll look at something else. If you can't train on it, or train someone within your organization -- chances are, you'll look at something else.

Those are the only two entities that I see having both those factors in mind -- though, FontLab is the higher priced option.

Personally, I think if Open Type were finally treated as the default standard for once, instead of tiering the approach and restricting between apps (TypeTool and ScanFont vs. FontLab, etc) -- then FontLab would have no problem creating more buyers. But, that's just my opinion.

PS:
I'm still not willing to get out of my beloved FOG just yet. I would just like to see them update it to standards, and include Open Type support.

dberlow's picture

a font is: (a) a collection of glyphs (b) that have space to the left and right side and (c) possibly may be kerned.
I wish it was this easy:) But for the beginner, I'm compelled to 1/2 agree. Point out the "Y" issues for his beginning, and I'll fully agree.

Cheers!

twardoch's picture

Daniel,

what is "Font Editor"?

A.

typovar's picture

Font Editor:
Font Editor Pure Java font rendering
We are building a pure Java™ font-rendering technology - targetted at J2ME, Personal Java and Java 1.1 environments.

This supports anti-aliasing, hinting, kerning, background blending, variable-strength emboldening, italicising and condensing - and will allow access to a range of TrueType™ fonts.

There's also High Logic's FontCreator. It's also affordable but there's no information on OpenType features.

blank's picture

Font Creator would be a lot more useful if it ran on Macs. Windows-only software isn’t especially useful in the design world.

Dan Gayle's picture

Now, correct me if I'm wrong, but wouldn't Python, rather than Java, be a better platform independent language to build from due to the already available scripting code such as FOG? I mean, if you're going to start from scratch with the code, you might as well use what little is available already that is a)open and b)professional, am I correct?

I'm not a programmer at all, but that's what my impression is. I was told if I want to learn a programming language from scratch, I should start with Python because is it relatively easier to develop, but also scales well.

Thomas Phinney's picture

Disclaimer: I'm not a programmer. But I am product manager for a bunch of stuff that includes a number of applications, tools and libraries of code.

That being said, I gather there are many things to love about Python, but at least some kinds of operations are realtively slow in Python, in ways that become apparent when one attempts very complex operations.

I'm watching a large Python library get rewritten in C++ because of the ~100x performance improvement in certain key areas involving I/O. So, "scales well" is not something that I'd immediately ascribe to Python. I strongly suspect there are significant performance-related reasons why an application like FontLab is unlikely to be done in Python.

Regards,

T

canderson's picture

what is “Font Editor”?

Yeah, what is this thing? On my Mac, it simply doesn't work. If the UI functioned as advertised (it does not) it would still be of little value. In my experience a lot of Java(tm) applications are not platform independent, if only because they are not tested on all platforms. It's hard to tell if it's a fake/stub app. or if it actually will function.

I'm sorry if this is off-topic, maybe this deserves it's own thread.

raph's picture

Thomas: I am a programmer, and if I were doing a font editor as a traditional desktop app, Python is very likely the language I would use.

Yes, it's slow for some things, but for this application there are a number of factors to mitigate that. First, font editing doesn't require a tremendous amount of actual computation, nor the movement of huge amounts of data. Heck, if Fontographer could work on the 8 MHz Macintosh Plus of 1986 (the year of its first release), even a 100x slowdown would still be reasonable on modern hardware.

Fortunately, the situation is better than that. If you can get the heavy lifting to happen in libraries, it's full speed. The antialiased vector graphic rendering should be done in a cross-platform library such as Cairo, or in the native graphics engine of the OS, such as Quartz. (One frustration for me is that the comparable bindings to the native graphics library aren't as mature on the Windows side as on Linux and Mac, but with luck that situation will improve). Some numeric computation, especially when dealing with vectors and matrices, can also be shuffled into NumPy or SciPy, which are every bit the match for handrolled C++ code.

It's even possible, though still somewhat cutting edge, to run Python code in the browser. Within the Microsoft world, there's IronPython, which should run inside Silverlight. On the free software side, it's possible that integrating Mono with a browser should provide similar functionality. But none of that stuff is quite ready for prime time yet, so I'm doing my Spiro stuff in straight JavaScript (which actually isn't all that different from Python, as a language).

I think it's fair to say that Python scales well compared with other scripting languages, but if you're writing hardcore systems-level apps, none of the scripting langauges scale anywhere nearly as well as C or C++. Java is somewhere in the middle of that, and has carved out a good niche for server-side stuff, partly because it supports threading natively. But none of its strengths apply very well to the specific task of font editing. In fact, I don't see anybody in their right mind doing real client side or browser-centric development in Java these days.

yuri's picture

I'd like to say few words about some other font tools that weren't mentoined in this discussion.

- RoboFab. It is a set of Python libraries and tools and you have to know how to use it, but it can actually create a very powerful and flexible font production environment (used with FLS or without it).
- Superpolator. A tool that nobody should miss if ever working on the font family.
- MetricsMachine. Simply the best tool to edit font metrics.
- ScanFont. Not just a "tracing" tool but a very good in converting fonts made as EPS files into real stuff. So draw anywhere, get your font build as a collection of outlines (in a single picture) of any size and convert it into font in few clicks.

Not all of these tools are already released (for both platforms) but it could be a good idea to find a way to try them.

Steve Tiano's picture

Does it reach a point where the wheel is being reinvented with all these other pieces of software entering the fray? I mean, even I have to admit, if I were to get heavily involved in what is now just an "exploration," I think I'd have to move over to FontLab Studio, ultimately, because Font Forge does not give you Open Type. And, surely, that's where a viable commercial font needs to be from the start nowadays. Or am I incorrect on this?

aluminum's picture

I think FontForge does opentype. But, then again, if you're at the point where you think your face is ready to be moved into a commercial product, just grab FontLab at that point, import your work and finish the OpenType version there.

Yuri...thanks for the info on all those tools!

William Berkson's picture

Yuri, thanks for coming on to comment.

I am interested in Superpolator, which I have also heard highly praised elsewhere. How do its capacities compare to the Multiple Masters functionality? Does each do something the other doesn't, or does Superpolator do everything that Multiple Masters does and more?

yuri's picture

William, have you seen this: http://superpolator.com/ ?

William Berkson's picture

Yes, but I didn't see a direct comparison with Multiple Masters.

I haven't yet worked with either, but I soon will need to. I am looking for simple answers to the questions: what are examples of what superpolater can do that Multimaster can't? And visa versa? How easy is it to use one verses the other?

anagnost's picture

As I have already stated, FontForge actually does much more of OpenType than FontLab or AFDKO. It may be comparable only with VOLT, but, unlike VOLT, supports both TrueType and OpenType-CFF fonts.

vanblokland's picture

William:

Superpolator vs Multiple Master - in short, Superpolator can take any number of masters on an axis, have any number of axes and place masters at off-axis places. While it won't magically reduce the number of drawings, it allows you to construct different "design spaces". For instance 6 masters total, Light, Medium, Black, and then condensed versions each. Any construction -- 9 masters on 3 axes, 4 masters on 1 axis, whatever your design needs. It can recreate any MM design space and then some. Glyph specific masters - you can insert a master for a single glyph if you have to. Superpolator uses the same math as GX variations. Previews, animations, export to pdf.

Multiple master is fixed on 2 master/1 axis, 4 master/2 axes, 8 master/3 axes and I believe even a 16 master/4 axes setup. If your design happens to match that, great. Usually things aren't so neat and need intermediates, fixes for specific glyphs, lots of previewing. In Multiple Master, to get the "middle" right, the weights that stand to be used most often (most often anyway) you have to work through the extremes. You can't edit the "regular", you have to tweak the light and the black to change the middle - this forces the extremes to be a lot less extreme. With Superpolator you could set up light/regular/black on a weight axis. While this means one more master to draw, it's right down to business. The extremes can be more extreme.

Superpolator runs on OSX, is written in Python, and is plenty fast ;) But a lot of people do great work with MM as well. Whatever makes your designs tick.

If you need more info, I have some keynote slides I can send.

William Berkson's picture

Thanks for coming on the thread to explain your product.

The fact that Superpolator gives more flexibility than Multiple Masters in doing and modifying the middle and more extreme weights has sold me on it.

I've already done the middles on three optical sizes, as these are the most important, and some heavier and lighter weights, with more to come. It sounds like it will be be extremely helpful as I go forward on the project.

In fact, it even sounds a little fun :)

Syndicate content Syndicate content