The RoboFont Editor thread

eliason's picture

It's the talk of ATypi 11 and the Twitterverse, so let's talk about it here.
RoboFont was released last week. It is a new font editor that runs on a Mac. It works with the UFO format. I'd love to hear testimonials about what it does and doesn't do, what it does better than preexisting tools, and so on.

John Hudson's picture

I've been using Frederick's UFOstretch tool, and have been very impressed with it, so am looking forward to trying RoboFont. My primary workflow, though, relies heavily on Windows-only software (VOLT and VTT), so RoboFont will need to be very significantly better than Windows FontLab before I'd consider making it my primary design tool (and then I'd have to invest in a lot of new hardware, because I couldn't face full time type design on a 13" MacBook screen). Where I see RoboFont being most useful for me in the immediate future is for making edits to sources that are already in UFO format for use with UFOstretch, MetricsMachine, Prepolator and Superpolator.

eliason's picture

What I wonder most about is what the built-in spacing and kerning features are like.
I suspect that most beta-testers and early adopters use Metrics Machine.

blank's picture

I want to know where the demo version is. I have never used Robofont and have not seen it in use in over a year. I have used UFOStretch and that gives me confidence that Fred can make a good font editor, but that confidence does not compel me to spend €450 on a program I might not like. I will be choosing between Glyphs and Robofont in the coming weeks, and if Fred has a higher price and no demo then Georg wins.

John Hudson's picture

At present, if you have Metrics Machine installed then RoboFont will automatically send your UFO to Metrics Machine for kerning, rather than using the built-in spacing tools. Someone has already suggested in the forums that this should be an option or preference, rather than default behaviour, which makes sense I think: whenever there are two ways to do things, the user should be able to choose which to use, even if one is presumed to be more powerful than the other.

The built-in kerning tool looks fairly basic, but with good group support. It doesn't look like it can be set up to work the way I kern in FontLab, with lists of context words for each pair. I find trying to kern looking only at pairs or at key letter contexts rather than real words unsatisfactory, as real words reveal patterns better and make it easy to spot systemic problems in spacing.

[Like other tools, RoboFont makes a distinction between spacing and kerning as activities. This reflects how data is stored in the font -- as sidebearings and advanced widths, and as adjustments --, but I remain intrigued by the possibility of a holistic approach to spacing that would automatically calculate optimal sidebearings and kerning values based on how a designer visually spaced strings.]

eliason's picture

compel me to spend €450 on a program I might not like

For the record it's €400, but your excellent point stands.

Cristobal Henestrosa's picture

[Oh, nevermind. What Eliason said.]

dezcom's picture

As Arty Johnson used to say on Laugh in, "very interestingk"

butterick's picture

A few days ago, I paid my €400 for Robofont. I don't know much about it. I haven't used it yet. Maybe I won't for a few months. Maybe I won't use it on release-quality work for a year.

The type-design world is a lot smaller than the graphic-design world. The economics of software development are commensurately more difficult. For me, that €400 is not a payment for software that's necessarily useful in its current form, but rather a down payment on what Robofont can become. If nobody invests in Frederik's work, Frederik will find something else to do with his time.

Also, that €400 is indirectly a way of supporting Erik, Just, and Tal's work on Robofab and UFO. I get a ton of value from Robofab. I paid nothing for it. The fact that Robofont relies heavily on Robofab & UFO gives me confidence that it will go somewhere. Maybe I'll turn out to be wrong. But I'd hate to see a worthy project die of starvation in its infancy.

Glyphs may be great. I haven't tried it. Maybe I will. But Georg Seifert is a consultant for Fontlab, so he has a sponsor for his work. I'll continue to be a paying Fontlab customer, so I'm indirectly supporting his work too. It seems inevitable that ideas from Glyphs will find their way into FLS.

In general, I think that fontmaking tools are underpriced, relative to the value any professional type designer gets out of them. Frankly, if Fontlab had asked me for that €400 a year ago to fund the work on FLS 6, I would've gladly forked it over.

Before you say "that's crazy," nobody in the design world seems to have a problem shelling out to Apple for new laptops and Mac Pros every few years. But mostly, your hardware sits around waiting for you to do something. Better productivity is determined largely by software quality, not hardware quality. And if no one wants to pay for better software, it will never appear.

blank's picture

Matthew, you are right about the need to support new software development. That’s why I have licenses for Superpolator, Prepolator, Metrics Machine, Font Remix Tools, and UFOStretch. But I cannot afford to buy everything. And there is also a very large market of font software consumers who are not type designers and do not see a reason to support new software, and somebody has to reach out to them. That is how I started out, and Fontlab had demo versions of everything ready for me to start working with.

twardoch's picture

I've had the privilege of talking to Frederik Barlaen for a few hours after his excellent presentation in Iceland. RoboFont is a solid piece of work.

My short description of it would be: it focuses on drawing glyphs, and acts as a switchboard for UFO-based projects. It also is extensible through Python, so plugins can be created.

RoboFont largely seems to follow a traditional approach for designing type: you have a font window, a glyph window, a metrics and kerning window. So any user familiar with Fontographer or FontLab Studio will find many similar concepts. It does introduce a few novel ideas as well.

It currently does no hinting and generally is free of any "auto" functionality which is design-related. So it does have auto save but does not have auto kerning, automatic generation of Unicode codepoints from glyph names, automatic creation of composite glyphs or automatic building of groups/classes. This seems to be Frederik's deliberate choice, as he listed "no knowledge" as one of his principles. In his view, as it seems, design choices should be left entirely to the designer. (Of course if the designer is a coder, he can put his knowledge into Python code as a plugin, but Frederik seems not to intend to offer such functionality "built in".)

RoboFont can generate OpenType PS .otf and works natively with the XML-based UFO format. As far as I understood Frederik, it is a specialized tool which is designed to work best within a suite of UFO-based tools such as MetricsMachine, Prepolator and Superpolator. Once FontLab Studio adds native support for UFO, RoboFont will be able to work hand-in-hand with FLS as well.

It is a very welcome addition to the set of tools available to type designers. If there's even just one thing that it does better than existing solutions, then I believe it's worth supporting.

I'm thrilled that there are people who decide to develop type design applications (this was the reason why I joined Fontlab Ltd.: I was amazed that a single developer, Yuri Yarmola, spent years to create an application in which I could do all my font-related work, an application that I still use every day, sometimes loving it and sometimes hating it).

As it was said here, the market is quite limited, especialky if we count full-time type designers. I think that while products such as FontLab Studio are an effort to create a comprehensive solution for type design and font mastering, there is also a need for more specialized products that address certain aspects. If products from various vendors can interoperate seamlessly, then even better.

UFO is a great example of a format which can be used for interoperability. Fontlab Ltd. supported the RoboFab developers financially in the past, and we will be improving our support for RoboFab and UFO in near future. Immediately after FLS 5.1 is released, we'll continue our work on FontLab Studio 6, which will support UFO version 2 directly.

I'd like to thank Frederik and Tal for the fascinating conversations that I had with them in Iceland. It's pure pleasure to spend time with smart people who are passionate about the software technology of making type. Respect!


Mark Simonson's picture

I've been using Robofont off and on since late in its beta stage and a bit more since it was officially released. Altogether not a lot of time, but enough to get an impression of it.

It's difficult not to compare it to Glyphs, which it feels similar to in a lot of ways. In comparison, its UI not as polished as Glyphs, but it feels more like the core of a powerful, extendible system of tools. Knowing Python (or access to someone who does) seems essential to getting the most out of Robofont. I've got some experience with Python, but it's a bit daunting to me at this point.

I've been able to feel my way around Glyphs, Robofont seems less intuitive. Trying things sometimes leads to unexpected results, and the documentation is a bit sketchy. It does little more than explain how to use each feature and tool and could use some editing.

As I said, I haven't used it a lot yet, but I have generated fonts with it and am happy with the results. It feels like the start of a big thing for font makers and I'm committed to learning to use it more fully.

Typography.Guru's picture

I would choose RoboFont if I would use or aim to a UFO-based workflow. Glyphs does a lot more to help regular users with the design process and the nasty technical details. But that does not mean that Glyphs is just a beginner’s tool that would not be suitable for professionals.
Both offer the great OS X working experience, which FontLab’s products don't/can't have at this point.
But don't forget that these are 1.x releases. I expect some healthy competition between the two and great improvements in the near future. Which is something I don't really expect from Fontlab anymore. I like the FontLab guys and I know they are on a rather small budget, but it took SIX years from FontLab Studio 5 to FontLab Studio 5.1, which brings nothing more than Lion compatibility and some bug fixes. I know there are great plans for FontLab Studio 6, but when i do the math, when will this come out? 2017?
I hope Georg and Frederik make a lot of money with their tools so they can hire other developers to extend their software and people doing good documentations how to use them.

Aaron Moodie's picture

Frederik tweeted last night that demo version would be available soon.
Looking forward to it.

dberlow's picture

Back to your oars then.

blank's picture

Back to your oars then.

I can paddle for another week or two.

Cristobal Henestrosa's picture

Demo available now.

dezcom's picture


blank's picture

So I like RoboFont, but the feature set is very limited. Of course, that will not be the case once a lot of extensions have been developed by users. I think that most individual users will never have the time to replicate a lot of the basic tools found in FontLab Studio, Glyphs, and Illustrator. This brings me to my question for RoboFont users:

TL;DR: Do you plan to share your RoboFont scripts/extensions? Or will you be holding them back as a proprietary advantage over other type designers?

Mark Simonson's picture

I've been wondering the same thing.

Goran Soderstrom's picture

I'm also curious, because in this thread David Berlow writes:
I have moved to Robofont full time for the integration it brings to my work as a designer, producer and hinter of full functionality fonts for print and web, if that's another clue.

Which raises this question: How do you truetype hint with RoboFont? Am I missing something very crucial here?

John Hudson's picture

Goran, I also wondered what David's statement meant, which I guess turns on what he means by integration. I can see RoboFont sitting at a nexus of other tools, so perhaps that is what he meant.

benkiel's picture

Butternick and others interested in directly supporting Robofab: the paypal donation button on the homepage of goes to fund the work on the development of robofab and the UFO spec.

dezcom's picture

Do you use glyphs now regularly? How does it fit in with your workflow? I am looking at both Glyphs and Robofont (which is more expensive).

Any thoughts?



pvanderlaan's picture

RoboFont may lack some of the functionality that other font editors out there offer, but for a version 1 release it is remarkably feature rich I think. Does anyone remember InDesign 1.0?

dberlow's picture

TL;DR:?: Do you plan to share your RoboFont scripts/extensions? Or will you be holding them back as a proprietary advantage over other type designers?

I’m not sure which end is in sight for whom, but I’m hoping that there are going to be opportunities for type designers, architects, programmers, engineers and web developers in this area.

…as long as Microsoft, Apple and Adobe adopt positions as they have and continue to evolve their user agent parenting skills — at lease some hints and some deep and complex font stacks are going to be needed;)

GS: "How do you truetype hint with RoboFont? Am I missing something very crucial here?"

We were missing both that, a web font format, and a letter drawing module for the exiting python-/xml-based environment which already contains great spacing tools and tools for developing and managing stylistic and optical design spaces.

We still are missing a suit of tools which address each of the existing, and a couple of new, Truetype tables.

JH: I can see RoboFont sitting at a nexus of other tools, so perhaps that is what he meant.

The more sitting at the oars the straighter the course?

petrvanblokland's picture

While I am fully aware that from the typedesigner-user's point of view it is legitimate to compare FontLab, Glyphs and RoboFont on existing functionality and $€. But so far in this thread there are some basic issues missing in the discussion.
The comparison is valid if the applications are defined as complete out-of-the-box entities. A typedesigner (who does not want to do any scripting and has no means of letting someone else do that work), has to deal with the available user interface, tool set, feature set and the list of pending bugs. This makes the designer victim of the development of applications that are often merely production tools and not intended for design iterations, as is the case width equivalents for publishing, as Illustrator and InDesign. Being not (partly) designer of your own tools puts a designer in a vulnerable position.

RoboFont is not an out-of-the-box application. The fact that Frederik and I and others have been advertized as the “missing link” means that RF is not an application-island, even while its basic functions are great by themselves. The application can better be described as another item in the growing set of applications that allow to edit/manipulate/produce UFO fonts. The existence of and full commitment to the UFO standard guarantees that all UFO applications perform specific functions on the same font files, and it also means that they can do that simultaneously. This binds them together as one mega-UFO-application with already many great functions inside and counting.
The mere fact that Frederik added a basic spacing-kerning editor is therefor a service to novice designers who want to start with just one lego-box instead of buying the entire profi-set all at once. If you need better then add another UFO app to the family.

The advantages of building a modular mega-UFO-app are numerous and obvious:
- Spreading functions over different modules reduces the need to buy them all at the same time, or you can skip the functions you don't need all together. Or can do better yourself. Functions can be added over time as they become available.
- Decentralized development also means that a group of people (Frederik, Tal, Erik, me, …) can work on the functions that they really need and are extremely capable of making and maintaining.
- Constructing an open structure like this allows any development, competition and new technologies to be added to the overall functionality, without the need for the main developer(s) to put it on top of their list. This is very different from having one application that can only do most of the functions on a average level. And upgrading one function implies upgrading the entire application, which often is not attractive.
- The fact that RoboFont is written 100% in Python, goes way beyond any speed discussion (besides, btw, that the speed is great as it is now).

To grasp the advantage of full Python apps needs some insight in how object programming is organized. As Frederik did this in the proper way for the entire application it allows any(!) user to dive into the deep functionality of RoboFont and overwrite/redefine/add functions, even to the core of the program, by means of inheriting the existing classes. This way there is a total melt between the scripting and the coding of the main application. What richer environment for developing your tools do you want? And even if that technology is beyond the reach of the average typedesigner: the reason that is can be done, it will be done. Think of a near future with an rfStore, that has lots of free nice-to-have scripts and documentation and how-to's, besides commercial available scripts and tools as profi-versions of the free ones.
In the current RF version the 4 tabs in the main editor window are actually mini-applications, melted into the editor window, each with specific tasks. User scripts can hook into the application on the same level, adding more tabs that have full (!) access to the existing classes, functions, user-events and dataset of RoboFont. Over time more and more documentation will become available how to do this, along with a number of example scripts (which by themselves are useful mini-apps)
Where RoboFont is only released a week ago, already some of the existing UFO-app developers think of making their existing applications available as RF-plugin.

A bit about (any) pricing in general and about the € 400 in specific. It is common knowledge that nowadays the production cost is by no means related to the market price of fonts. Exactly the same is true for applications. The developer needs to sell a huge amount of licenses in order to come anywhere near the investment. And the only way this can be done it to cut on (professional) functionality and ignore pending bug lists. So, when an application like RoboFont is created, which allows other professionals to plugin their knowledge without any boundaries, any price for such a tool should be fine (if €400 – the equivalent of saving a modest 2 days of work in the lifespan of the application – is a problem, then it's the relevance and quality of the design work that is to be questioned, not the price of the tool).
Matthew did put it exactly right: it is a down payment, taking a share in a future that is likely to come (and in which all users take part), supporting the development of many great additions. It is not the payment for an application license.

dezcom's picture

"Does anyone remember InDesign 1.0?"

I remember PageMaker 1.0. Illustrator 1.0, and Quark XPress 1.o in their time. It was a different world then and this is a much more vertical market product.

Mark Simonson's picture

Chris: I wrote a bit about my experience with Glyphs in the "FontLab vs. Glyphs" thread: It seems more like FontLab or Fontographer in its intent compared to RoboFont (i.e., stand-alone solution), but it is easy to use it as a component of a UFO-based approach to font development. For that matter, FontLab can be, too, although you need helper scripts to get things in and out of the UFO. Not as seamless right now, but coming in a future version from what I have seen.

dezcom's picture

Thanks, Mark!

nina's picture

Petr, thanks for the insightful post. It's all becoming a little more tangible (and enticing).

pvanderlaan's picture

I remember PageMaker 1.0. Illustrator 1.0, and Quark XPress 1.o in their time. It was a different world then and this is a much more vertical market product.
That’s not what I meant. InDesign 1.0 was almost impossible to use as a professional publishing tool since it lacked many important features (such as colour-seperated output). At that time InDesign was highly criticised but very soon it caught up.

Angus R Shamal's picture

Err, a note to Frederik, Tal, Erik and Peter and whoever is involved in RoboFont:
You know, you could also just learn to answer emails from customers who have very simple questions, or better yet provide a descent downloadable documentation/manual, instead of a lengthy post on Typophile that still doesn't explain how to use the program within the so called 'workflow'.

And I'm not talking about illustrating typefaces.

This notion of expecting type designers to know what UFO is, to have experience with coding/scripting and visiting a conference or whatever, prior to paying a premium price for a program that doesn't come with the essential know-how-to documentation, seems a bit arrogant to me, wouldn't you say? It's this tendency to expect and assume that others speak the same "language" and have the same resources, that bugs me. This attitude is very evident in the initial user experience with RoboFont.

And stating that designers that don't know or wish to know scripting or coding are in a vulnerable position, is very true, and more so in the case of RoboFont. But why should it be? You essentially calling such designers "incomplete" or "lacking". While I could say the same things about your program package being incomplete and lacking.

I'm one of those "incomplete" type designers who for whatever reason IS interested in knowing more and getting into the celestial prophecy that is UFO. To see what news it has to offer to my prehistoric design and production workflow, and maybe even make me a better person.

But first I would like to know how to use the damn thing! Not over time, and not by visiting typophile or google.

So, get off your high horse for a minute, and learn also how to treat those customers whom you haven't met in a conference, and maybe just maybe write/offer a manual while you're at it.

Angus Shamal

vanblokland's picture

Err, a note to Frederik, Tal, Erik and Peter and whoever is involved in RoboFont:
You know, you could also just learn to answer emails from customers who have very simple questions, or better yet provide a descent downloadable documentation/manual, instead of a lengthy post on Typophile that still doesn't explain how to use the program within the so called 'workflow'.

Hey, how about writing that friendly email before blowing up on Typophile? I don't think we've ever refused a call for help. Dude.

Mark Simonson's picture

Angus: I agree with some of what you say about Robofont, but you don't need to know anything about scripting or coding to take advantage of UFO. It's an open font format, so you're not stuck with using only one program to work on your fonts (if that's a problem, and it definitely can be).

dezcom's picture

Manuals are not a trivial pursuit. They are time consuming and expensive to produce. I am glad they got their new products out to the type design community for comments before they went to the expense and delay a manual would cost. I see it as type designers sharing their tools with the community rather than the money grab you seem to see.

petrvanblokland's picture

"But first I would like to know how to use the damn thing! "?
It's damn simple: how to learn to play violin? Just pick up the damn thing and start playing. A lot! And more!
Ever heard about someone who learned to play music from a manual?
But no worries, I think it was my first and likely my last contribution to this thread.

Angus R Shamal's picture

Hey, how about writing that friendly email before blowing up on Typophile? I don't think we've ever refused a call for help. Dude.

Eric, check your records — and by "your" I mean the RoboFont website records. More specifically emails sent on September 19. If you need the order number, mail me.

And I'm not your Dude.

Chris > as far I know this wasn't a beta-testing version I paid for, but an official release. It's basic customer service I'm talking about, and I don't think a PDF manual is expecting too much.
And please don't put words in my mouth, no one accused anyone or anything being "money-grab". I paid willingly to a promising program/workflow I'm interested to pursue.

Peter > Again, I'm not talking about drawing type or designing type. Those tools seem fairly straight forward as most of the other tools. And I'm not a beginner, but just new to RoboFont or UFO etc..

But what about OpenType features? Each program has it's own slightly different way of scripting them — which is the way of RoboFont?
How to work with smart sets? where to import standard ones from, etc.
How to implement work with classes?
How to integrate the UFO workflow? maybe with some examples.
..and a bunch of other stuff that seem to differ from other programs, like FontLab for example.

I see your Documentation section on the website isn't that far off from a PDF version.. if it also addressed more points and included more examples, etc.

William Berkson's picture

Hey, Petr, I hope you will continue to post.

About the violin analogy. There may be a few natural geniuses who learn to play the violin by just picking it up. But almost everybody does have a teacher, and instructional books. So the fact is that most have do have manuals, and personal instruction, as well as trial and error and long practice. It's not "either-or," but "both-and." And every single master had a teacher, even if they learned it first by themselves. Louis Armstrong learned to play trumpet by ear, but Joe "King" Oliver was his mentor, mentored him. And Mozart learned from his father, who was a music teacher.

So IMHO, helping people to use Robofont, both with a manual and interacting here with users and modifying it in response will make it far more usable and used. The basic fact is that if you get stuck using software, and you don't do code yourself, you're screwed unless you can ask somebody how to get unstuck.

I'm not saying that you have to be the one answering questions, on line, in public, but if somebody does, it will make a better product and certainly a more widely used one. And whoever does the answering will have to deal with with uncalled for and annoying rudeness. That's just the way the internet is. Rude, but amazingly useful.

petrvanblokland's picture

Suggested template for a workflow:
“Dear Frederik,
i have sent you a mail with some questions about the use of RoboFont after my immediate purchase.
I guess that you must have drowned completely in mails after the successful introduction in Iceland. So I can imagine that an answer takes a bit longer that you expected. And I understand that you are not a large company with call centers and distribution departments.
However, I am very anxious to use this great tool, so maybe you know someone else that can help with some initial directions.

I appreciate very much that you have dedicated 2 years of your life building this application, which otherwise never would have existed. I am very much aware that my contribution to this is only the cost of 2 days of work. Wow. I guess this new platform will allow a lot of people to contribute tons of great stuff to the type community.

Looking forward to your response,
Yours Etc.”

kaibernau's picture

Also, you may go to

See where it says "RoboFab"? Click on the text there. Tells you all the basics about RoboFab scripting.

See where it says "UFO"? Click on the text there. Tells you all you'd ever want to know about the UFO format. (Sidenote: Who here is yelling at FontLab for not providing a detailed public spec of their format?)

(Edit) Also when you direct your browser to it will tell you where to find the necessary information about OpenType feature scripting. Straight from the horse's mouth, as they say.


dezcom's picture

> as far I know this wasn't a beta-testing version I paid for, but an official release. "

These guys are not Microsoft or Adobe. They are your neighbors working from home without a big bank account and guaranteed market share--cut them a little slack and give them a chance to grow.

Höfe's picture

[comment by moderator]

Account blocked for gross misconduct, unwarranted all caps shouting, and too damn many typos fer chrissakes! Friggin' lunatic…





pvanderlaan's picture

A very concise introduction to RoboFab (the Python scripting API which works in numerous applications including FontLab, Glyphs and RoboFont): (It is even printable too!)

And on a slightly related note: the next edition of the Robothon conference, scheduled for 8–9 March 2012 in Den Haag, will be largely dedicated to RoboFont, UFO workflows and the UFO 3.0 spec. See you there, Angus?

dezcom's picture

Where will Robothon be held, Paul?

Jackson's picture

Where will Robothon be held, Paul?

scheduled for 8–9 March 2012 in Den Haag

eliason's picture

(That's The Hague to us Amerikanen.)

Angus R Shamal's picture

I was pointed back to this thread.. I can see why.

Kai and Paul > thanks for the reference links and info! I'll certainly make time for March for the next Robothon. [edit] Hopefully I'll be on the same page by then.

It's probably too late and the damage is done and all that, but I realize I stepped on some toes and rubbed on some people the wrong way. For what it's worth, it's not the program itself I was aiming at, which on itself reminded of Glyphs' UI which I'm very fond of. That's what attracted me to purchase RF in the first place, to finally be able to get into the UFO workflow in perhaps a more complete manner.

It's the fact that I felt a bit lost when first opening and trying out the program, with no immediate support document to turn to or never getting answer to my email etc, that left me disappointed. Maybe that email got lost in traffic of responses, i don't know. Maybe I'm the only one who felt lost opening the program at first. Or maybe I should have been better informed before even purchasing it in the first place. Maybe maybe maybe.
In any case, my point was that there might be different customers out there, some of which might need a better introduction and maybe a guide into a new type of program and workflow.

dezcom's picture

That is big airfare, too ;-)

Syndicate content Syndicate content