Many problems surrounding free fonts of which more people should know they are solved, on to interesting type-centric discuss’n?

Schrijver's picture

Or am I wrong?

With regards to discussing free fonts on the crit board:

Could the different concept of authorship exemplified in FOSS add an interesting dimension here? As in, it makes perfect sense to bring in a typeface you did not design yourself yet still try to fix it?

There are many existing FOSS fonts out there, of which a large part could presumably use tender loving typocare. Wouldn’t it be a very efficient way to increase the quality of FOSS type when, for example, your spacing in OpenOffice looks off, you just bring up the font here and let the community help you in fixing it?

Point 2

I didn’t see a who’s going to Robothon thread yet, but I am, more of you who are interested in the foss type world?

Nick Shinn's picture

Perhaps if I could be convinced that a multi-authored typeface would have some new emergent quality.

But I haven't heard that argument. In fact, the assumption seems to be that quality can be "increased".

Si_Daniels's picture

2 issues with the above and an observation.

"Experts" - the experts inability to maintain the quality of extensions to Vera for example, seem to suggest that their expertise or their influence is limited.

"initiated by the typographic community" - seems to suggest more that more than a handful of well intentioned part-timers and amateurs initiated any given open source project.

I feel any significant growth in open source fonts will be driven by the type of people (students, beginners, savvy marketeers, special interest folks) who have traditionally given fonts away for free. Moving to OS licenses is a good move for them and the community. Rather than just post fonts sans-EULA or with a EULA full of holes that they made up or copied from someone else, OSS licenses are a step up in that regard.

James Arboghast's picture

Hi Eric. I'm the director of Alt Indie Typo, the privately run group working on the "free, open source font superfamily, starting with a linear sans serif" you mention. I recently retired from Typophile but my involvement in the FOSS movement necessitates a posting on this thread.

There will soon be a spiffy looking distribution channel for FOSS fonts, maintained by people who as far as I can tell are serious about designer’s IP.

I have very little involvement with the Open Font Library, but enuff to know it is run by good-intentioned peeple who lack experience making fonts. Mainly they have experience making open source software. You only have to examine some of the 160-odd free fonts in the openfontlibrary to see the general lack of quality and almost complete lack of usable, useful fonts in the collection. So far there isn't much at all in the collection many web designers would even consider using.

There is an effort initiated by the typographic community ‘to make a free, open source font superfamily, starting with a linear sans serif’. It’s proceedings are not public, which may not accord to the working ethic of the FOSS community, but would that matter if they manage to produce great fonts that are licensed under a free license?

Alt Indie Typo operates as a closed group in order to get the exacting task of designing our text type series done and get all the variants and weights of font made. We leave discussion to venues like Typophile and openfontlibrary.org

I drafted a management framework for the purpose of obtaining a specification for what web designers want in the way of open source web fonts, and the response has been disappointing in the extreme.

Web designers and the Openfontlibrary.org are very keen on having a high quality open source text font superfamily yet they have so far totally failed to generate a spec for the work. This situation is unworkable and impractical so Alt Indie Typo has gone ahead and drafted its own specification for the text font superfamily we will build. I just completed an A - Z, a - z draft of the serif variant this morning to get an idea of how the final design might look.

Could the different concept of authorship exemplified in FOSS add an interesting dimension here? As in, it makes perfect sense to bring in a typeface you did not design yourself yet still try to fix it?

Nope. That approach smacks of naivity and lack of experience at developing fonts.

There are many existing FOSS fonts out there, of which a large part could presumably use tender loving typocare. Wouldn’t it be a very efficient way to increase the quality of FOSS type when, for example, your spacing in OpenOffice looks off, you just bring up the font here and let the community help you in fixing it?

"tender loving typo care" ? I'm compelled to ask you a question: have you designed any typefaces yourself? Your example involving off spacing in OpenOffice seems naive and attuned to the open source software community, and your solution even more naive. If the spacing of Alt Indie Typo's text font superfamily looks off in OpenOffice it will be because OpenOffice isn't handling the fonts properly or because some users can't handle the fonts properly. Proposing to "fix" a well-developed font by bringing it here to typophile's crit board is an ad hoc reaction that could be avoided altogether with a few judicious tracking adjustments in OpenOffice.

I didn’t see a who’s going to Robothon thread yet, but I am, more of you who are interested in the foss type world?

I dwell in Melbourne, Australia and have no interest in Robothon. My interest in the FOSS type word is limited to the bare minimum required to communicate with a group of peeple who, in my experience, have bitten off more than they can chew.

For example, there is a basic and serious problem with open source collaborative font building projects. Version control software only works on plain text files. That's fine for text projects like Wikipedia but hopeless for vector font project file formats like Fontlab's .VFB files. So technically no way exists for open source collaboration to take place with a version control system to allow a typeface design or font build file to be wound back to a desired point in its development history with a revert command the way Wikipedians do every day at Wikipedia.

There is a great free font editor, a great free font scripting library, and a very handy free-as-in-beer SDK (though I am just echoing the consensus here, can not claim personal authority on these matters :-o

Compared to Fontlab, FontForge is not great, and since you're not an authority on these matters I don't understand why you say FontForge is great.

Take it all in good faith because it's meant that way.

j a m e s

Schrijver's picture

I agree with the prevalent opinion that a font which would be stuck together in a Wikipedia style direct democracy scenario would be awful. But a font is not comparable to a Wikipedia page—rather, it could be comparable to Wikipedia itself: a product driven by the vision of an individual, but coming to substance through the contributions of the community.

This is my own theory which I am propagating to both sides of the fence:

Multi-authorship is overrated, and is the least interesting part of foss.

Multi-authorship is silly

The driving force for multi-author projects such as Wikipedia is our incredible skill level in conforming to the others around us.
This is a powerful mechanism: because we have such a stubborn drive towards consensus, that even a completely random 'crowd-sourced' population can create a coherent cultural artefact which, even more strikingly, comes across as if it could have been written by a single author. (dixit a FoF, will look up)

Yet here is the important part: very often the wrong conclusion is drawn from this observation.

  • There is a diverse amount of people with a diversity in tastes, opinions and longings
  • These people are able to collaboratively create a coherent artefact

Leads people to conclude

  • The resultant artefact necessarily reflects the diversity in tastes, opinions and longings of these people

This is simply not true. Rather, you can say:

  • The resultant artefact necessarily reflects the common ground this group of people is able to find

Which might make it a handy artefact, and very interesting from a sociological perspective, but not really interesting in other ways.

distributed authorship is great

But multi-authorship is not really a prevalent model in open source development.
And IMHO neither should it be in foss type design.
In my view it is not about loads of people working together on one design, but many people working on their own versions and sharing the more general smart ideas people come up with.

In open source, just like in the arts the need individuals feel to differentiate from others plays a very large role.
Inspiring to me was the surprisingly analytical description of this model in Eric Stephen Raymonds 'Homesteading the Noosphere', and more exactly the chapter Open Source and Ownership.

This is much more interesting than democratic collaboration IMHO, as it is more close to how science or art works: you produce something immaterial that you freely spread (‘knowledge’, ‘artistic ideas’) and find alternative economic strategies to enable you to keep doing so.

This approach has arguably brought us some nice things.

The big added quality of this approach is, that you can build on the work of others. Sort of like type revivals, but more directly, because you can quote your contemporaries.

You could see it that way: an individual or a small team is still responsible for a design, but they are feel to pick and choose the bits they like and want to adapt from other designs. This might seem a strange idea for a type designer, because it seems you could just go and knock about a new typefaces borrowing bits and pieces from others; but the reason this works is social pressure. Because you are in a small transparent community with a steep barrier of entry, you will only want to adapt somebody others work when it is obvious to your pears that you are using it in an apt way.

The reasoning is that you can create with higher sophistication and adapt to smaller niches if you don’t have to reinvent the wheel all the time. Like when you use a JavaScript framework like jQuery: you have a tried and tested piece of code that abstracts away from the browsers implementations so you don’t have to deal with all the mundane details in your code.

The type equivalent could be making an open source ‘framework family’, an unassuming but robust set of glyphs upon which others can base their work, yet are subsequently able to push back specific refinements that also apply to the master.

That actually sounds like a good idea IMHO. Will look into that—the first step could be just stripping Gentium from all it’s baroqueness :-)

Basically it is all about the power of forking—the social code hosting site GitHub is all about that. In I think the OFL license really anticipates these mechanisms very cleverly: by making it very clear that if you start redistributing a modified version of a font, you should give it a new name and make clear that this is a new font.

This should prevent what of course is the worrying scenario for type designers: a multitude of ill conceived versions of their cherished designs floating around.

The GitHub example is inspiring to me: &with the UFO format as an xml based representation of a font, you have the perfect basis for this kind of granular design, because it lends itself extremely well to versioning systems such as SVN and Git. And you could even create a graphical interface to the public API of GitHub.

Schrijver's picture

Hello James

Thanks for your long reply.
We cross posted, and I think my comment addresses some of your concerns.
I think it is totally up to you which way you choose to make this font—if you heard a critique of your closed model, it wasn’t meant that way.

No I don’t design type.
And I don’t use OpenOffice.

Why I feel the need to weigh in because I am a user of both digital type and open source software,
and for me both are not living up to their potential.

I can imagine your assesment of me being naive, but I don’t think that’s the case—I think that by being neither structurally involved in Foss or in type design, I will I will indeed overlook all kinds of small problems that occur in both fields—but IMHO also some of the dogma.

I have a practice as a conceptual artist for which I make use of both old and new media.
And I feel the digital world does not yet offer me the possibilities to express myself the way I would want to.

I use both closed and open software under OS X.
And for me this combination is extremely productive.

Because for day to day stuff I use proprietary, and ‘everything just works’. At the same time, when I have a very specific need that’s not met by the market, I can use an enormous wealth of open software through MacPorts.

Which for Unix standards is still pretty much ‘everything just works’.
No this would be nice and jolly, but the tech geek world
lacks any fundamental notion of either the function or the value of visual communication
So why would I care?
Because tech geekery is shaping technology, which is shaping the media I can express myself in.
So I decided any mutual understanding between the two worlds would be a good thing.

__
art directing the internet in 2009

Schrijver's picture

@MOD, I seem to be unable to edit the above post, is that because I started this thread? If so, could you maybe remove the hard returns inside the paragraphs? They really break the flow… Thanks!
__
art directing the internet in 2009

Nick Shinn's picture

Eric, can you give a practical example of how distributed authorship, involving free copying of glyph outlines, might work in type design, and what use that might have?

Schrijver's picture

Well my own main preocupation with type is as a tool to connect to the visual part of our cultural heritage.

Since these typefaces should preferably be free, and seeing how arriving at only a few decent free typefaces already proves to be such an ordeal, I’d say the hopes for a decent amount of historic free revival are slim.

But with this sort of distributed authorship I could imagine, for example, the following scenario:

A library has a specific and rich collection of late 19th century books from one specific press, that cut its own type.

They decide to digitize the collection. And they commision a font designer to design digital type for it, because they want to digitise it in a way that is at the same time accesible and flexible, and fitting the visual presentation of the originals.

The type director, or lead designer, starts studying the portfolio of typefaces. He takes all the time, effort and dedication as he would in a normal assignment. But he is not studying one typeface: he is studying a collection of type, looking for its essential characteristics, finding out which individual differences he wants to emphasize and which ones he will attach less value too.

He proposes a master plan [sic], and if the client agrees, the distributed part starts. At first he will just have a few assistants helping him set up the main skeletons for the prototypical faces of the collection. But after that point he will start to look for specific designers that can start working on specific problems that they specialise in.

This is where the forking starts, when different designers start working on different parts of different designs. The lead is the one who is minding the grand plan, but at any given point is able to use his very specific judgment on details if necessary. He is also the one who can pull in changes from one font and push them back to others.

You will end up with a collection of typefaces, the richness of which the lead designer could not have created on his own. Not only for lack of time, but also because the unique skills the contributors have contributed, skills not found combined in one person. Yet on the other hand, the type collection has all the qualities you would attribute to a traditional type family as designed by an individual. The different fonts are not described as individual entities, but in their relationship to each other, starting from the prototypical design bases created by the lead designer, who is also the one who has been making sure the collection works as a whole.

__
art directing the internet in 2009

Schrijver's picture

@sii
With the experts I meant typophile, and the initiative was referring to James’ project.
Surely that is more than a handful of well intentioned part-timers and amateurs?
I was just trying to recapulate what I have been reading in these threads: the willingness of the typophile community to offer their critique and insight regarding foss typefaces.

As it stands I don't think there are many type experts inside the world of open source, so I guess we assess the situation in a similar way?

I feel any significant growth in open source fonts will be driven by the type of people (students, beginners, savvy marketeers, special interest folks) who have traditionally given fonts away for free. Moving to OS licenses is a good move for them and the community. Rather than just post fonts sans-EULA or with a EULA full of holes that they made up or copied from someone else, OSS licenses are a step up in that regard.

I fully agree. And I also see how this development could be good for the type community by getting rid of shoddy licenses. At the same time, since you say it’s more or less the same phenomenon as with previous free fonts, the impact can’t be that huge on the professional community? Because this whole free-as-in-beer free-as-in-freedom distinction has nowhere near the relevance it has for the Foss community.

But the benefits to Open Source would be huge IMHO.
Because a talented type student might not be able to make a typeface like a veteran designer, he will make a typeface which is 100x better than that of someone who just feels like drawing some letters on a rainy sunday afternoon.

So if all these eager to prove themselves cultural professionals start using OSS licences as opposed to giving away proprietary freebies in exchange for e-mail addresses or whatever scheme they come up with, wouldn’t that constitute a great starting point for a Foss world that is more aware of the value of visual communication?
__
art directing the internet in 2009

Dan Gayle's picture

@james
Version control software only works on plain text files.
Rubbish. I use version control every day for binary files like .jpegs, .psds, and .ai.

@Schrijver
One of the major flaws of FOSS is that there isn't a single authority to accept blame for something that is utter crap. What is also true, and not a flaw, is that in the FOSS or Open Source community ALL individuals are supposed to accept responsibility for something that is utter utter crap and then FIX IT. That works great for code.

But it doesn't with a font.

A font is so delicate, so demanding of consistency, that having any more than one or two different designers working on it breaks the flow. A font is not a grab bag of glyphs, it is a collection of highly synchronized and cohesive designs that are meant to be a solitary whole.

That, by its very concept, precludes the development of fonts by a group.

Ray Larabie's picture

Consider this wacky model for a collaborative font project.

One font designer when actually builds and tunes "master glyphs". For example: an A is a master, an Aring is not. The font would be designed using a minimal, modular system. No glyphs with accents. No unecessary flips (brackets, less than, backslash etc.) Any characters that can be constructed automatically or by following simple instructions would be left out. A spec doc would specify all of this. Them more glyphs than can be made by following simple instructions, the less burden there is on the font designer.

A team of reliable type testers who send detailed reports back to the type designer and the rest of the team. Moving bezier handles isn't time consuming, knowing where to move them is. You need good testers who will actually print stuff on actual paper and look at it with actual eyes. If your testers just look at stuff in Fontlab and say "looks nice." then it can cause a lot of work down the line. Especially with interpolation. You want to make sure the source fonts are perfect or you can end up with he same mistake repeated 64 times, adding to the font designer's workload. So, you need real, professional typographers that happen to enjoy working for free.

A kerner. If only one person kerns, no version control is required. They just need to upload the latest kern file. With class based kerning it's not so tough, even for a large project. It's work but one hotshot kerner with skills can do it if they know how to deal (and cheat) with classes.

OpenType coding. This can be done with version control software as it can be done in a separate text file. Could be 1000 coders.

Autogenerate scripting. People who write code for accent generation, ordinals, fractions and other "mechanical" aspects of font design which can be automated. This person would be involved with writing the spec as well. They would have to confer with the font designer whether on not a glyph can be autogenerated, or if autogeneration shouldn't be used.

Font monkeys. They build glyphs based on the font designer's instructions. For example:
afii10048 - generate I_E_O decompose E and extract the midline. place the O two periods distance away from the I. Adjust length of horizontal line and remove overlaps. Copy right sidebearing from the O. That sort of thing. Font monkey may not be the best name for this position.

Interpolators. Blending regular and bold to make semibold aint no thang. The font designer probably wouldn't even need to be involved in minor interpolations. Just the testers. It's often more of a cleanup job than a design job. The font designer would have the final tweak on each interpolation but the initial cleanup should be done by someone else.

It's something I've thought about for years. Just as a "what if" scenario. Multiple font designers aimlessly poking around at a font sounds like a nightmare. It might work in an office environment but over the net it's a wheelbarrow of pain.

Nick Shinn's picture

That monkey-to-maestro ratio could be an issue.
The deal is usually, you do pro bono, you get some creative jollies.

Ray Larabie's picture

@Nick That's one of the many reasons I don't want to get involved in a free collaborative project. I know very well who's going to get stuck welding a hundred friggin' ogoneks all weekend. Perhaps it would work if the laziness level of everyone involved was about even. Then it'll self regulate.

Dan Gayle's picture

@typodermic
You want to make sure the source fonts are perfect or you can end up with he same mistake repeated 64 times, adding to the font designer’s workload.
This hasn't worked well for Utopia, Charter, or whatever Vera was before it was Vera, and no one is going to claim that they weren't as perfected as a typeface gets.

It's proven, give out a goldmine, and the "font monkeys" will start digging out coal.

Quincunx's picture

> "That’s one of the many reasons I don’t want to get involved in a free collaborative project. I know very well who’s going to get stuck welding a hundred friggin’ ogoneks all weekend."

Yes, to me it sounded a bit like what happened when mass production was introduced in factories. Alot of different people doing one little part of the whole. You know, the conveyor-belts and whatnot.

But when designing a typeface a coherent end product is the goal, plus that it is oft times very personal. I don't really see how some of the proposed workflows would work in that sense.

Scott Thatcher's picture

Oo, Oo! I want to be a font monkey!

It’s proven, give out a goldmine, and the “font monkeys” will start digging out coal.

But seriously, it seems that one problem with the additions to Utopia and the others has been a lack of continued involvement by a type designer who is responsible for giving continued instructions and for ensuring overall quality. At least that's what I hope has been true when I look at some of the modifications I've seen--others haven't been so bad. "Let the monkeys loose" is not a good strategy in most every circumstance. Does anyone know of an instance where typodermic's scheme has been more fully implemented with an open font?

Concerning the problem of creative unity and producing a cohearant whole, I'd be willing to believe that such a workflow wouldn't be optimal, and it wouldn't replace the work of talented type designers, but could it have a place in helping to produce workhorse fonts that aren't terrible? For example, I wish I knew more details of the TeX Gyre fonts. It seems that they're paying to support someone who's working to make reasonable, semi-automated extensions of the Postscript fonts, with some testing and oversight for the final product. Is this an example of the suggested model?

st

Schrijver's picture

But seriously, it seems that one problem with the additions to Utopia and the others has been a lack of continued involvement by a type designer who is responsible for giving continued instructions and for ensuring overall quality

I think that’s exactly what should happen were such a collaborative effort to work?

IMHO the moment that can happen is when there is an economic potential for this workflow—I don’t see where this niche might be just yet.

A lot of influential open source projects are very much associated with their lead developer. jQuery is John Resig, Ruby on Rails is David Hanemeier Hanson. But Resig works for Mozilla and Hanson for a web consultancy firm (37signals). Mozilla makes browsers, so they benefit by web apps taking off, 37signals makes web apps with the Rails framework, so they benefit from it being popular.

So I guess there has to be an investing party that sees a potential for foss fonts improving the experience of their core product (like SIL with Gentium, no? isn’t that ultimately meant for improving the experience of bible reading?)—and believes that a collaborative workflow benefits the development process…

PS I love the related foss concept of the Benevolent Dictator for Live :-)
__
art directing the internet in 2009

Si_Daniels's picture

Technically I think Gentium was a student project (Victor did the MA program at Reading) not specifically tied to SIL's activities.

Ray Larabie's picture

This hasn’t worked well for Utopia, Charter, or whatever Vera was before it was Vera, and no one is going to claim that they weren’t as perfected as a typeface gets.

Were they made using the method I described? It's not just about perfect letterforms. It's about good planning, a coherent spec and an organized team or professionals.

@scott. I should have explained that everyone involved would have to be a talented font designer. Everyone involved would have to be an experienced professional for a project like that to work. Don't get me wrong about font monkeys. There's no difference between font designer and font monkey. That's just their position on the project. So someone could be a font monkey on project A and font maestro on project B. Getting a team like that to work with no money might sound implausible. That's because it is.

That's what I believe to be the ideal method of creating a large collaborative font family. If you take the same formula, replace most of the people with inexperienced amateurs and one professional, you'll likely end up with one frustrated professional. I'm not saying you can make a collaborative font using other methods but if anyone can poke holes in my plan and suggest how it could be improved, go ahead.

Make no mistake: I don't want to get involved in a collaborative project. I just had this idea in my mind about how such a thing could work so I thought I'd share it.

Thomas Phinney's picture

A kerner. If only one person kerns, no version control is required. They just need to upload the latest kern file. With class based kerning it’s not so tough, even for a large project. It’s work but one hotshot kerner with skills can do it if they know how to deal (and cheat) with classes.

Even with just one person kerning, if the glyph complement is still expanding at the same time, version control could be a big issue.

OpenType coding. This can be done with version control software as it can be done in a separate text file. Could be 1000 coders.

See my comments on glyphs above. Also, there's a lot of potential for feature interaction, which requires that feature ordering be defined up front, and... stuff.

Cheers,

T

Ray Larabie's picture

Even with just one person kerning, if the glyph complement is still expanding at the same time, version control could be a big issue.
Those glyphs would be defined in the spec. Since the font designer has the only version of the font, they just have to load the latest kerning table. Once you get into the thousands of glpyhs, there's won't be much that can't be added to an existing class.

See my comments on glyphs above. Also, there’s a lot of potential for feature interaction, which requires that feature ordering be defined up front, and... stuff.

The font designer wouldn't need to be involved in that once the project starts. Since all the proposed features would be have been defined in the spec, what would the font designer have to do other than follow the spec. In fact the testers would be responsible for making sure it all works, the font designer wouldn't even have to be involved at all.

If the project were an unplanned feature creep monstrosity, then yes, that would be a problem. But, if it were properly planned and spec'd out, it wouldn't be.

Nick Shinn's picture

...feature creep...

But surely extra features are exactly the kind of customization that users would want to add, such as "afrc" (alternative fractions) or "pcap" (petite caps), or "hist" (historic forms). This last would be implemented as a Stylistic Set now, as it is not supported by layout applications.

Ray Larabie's picture

well, add 'em to the spec then. I think once the initial project is "completed", people can chip away at it for centuries.

Thomas Phinney's picture

@typodermic/Ray:

That's a fair point: if the spec is thorough, right down to exact glyphs, then the rest of the stuff will follow pretty well. Even occasional spec changes will at least be well-known and centralized, so everybody can work from the same info.

This does of course necessitate a detailed spec. It's also possible that, say, some person coding features will be the first to realize that the spec was incomplete when she realizes that some particular feature interaction requires more glyphs. But that should be manageable.

Cheers,

T

James Arboghast's picture

When Alt Indie Typo's text font superfamily is complete a decent spec will be available, readable from the glyph designs and other components of the fonts making up the superfamily. If, as a model or spec, the superfamily of fonts we produce don't meet everybody else's expectations of what the spec should be, the parameters it should adhere to, here is why:

The FOSS peeple are way too slow and not up to the task. You're not dedicated enuff to what you want to achieve to even participate in the FOSS Spec Steering Group and the rest of the FOSS management framework I set up for you. It takes half a day to draft a management framework like that. But since you lack the knowledge and experience required to put micro management into action, and it seems you don't care about getting the spec hammered out, I have gone ahead and worked out the spec myself. The FOSS community left me no choice.

Is it any wonder I'm frustrated by that?

The eyesight impairment resulting from my stroke in January 2007 recently got significantly worse; currently working on fonts and everything else on a computer screen is an uphill grind. I'm doing the best I can with the help of collaborators.

@Dan: thanks for the correction about version control software. I was going on what has been said about it until your correction.

j a m e s

Dan Gayle's picture

What I think is interesting is that there IS progress being made, if only a preparatory work. Till now, to my knowledge, no one, correction—no type designer—has yet sat down and said, “IF an open licensed font were to exist, and IF it were to be halfway useable/acceptable to the knowledgeable eye, THEN this is how it would have to work.”

I applaud your efforts and your tenacity :)

raph's picture

In case you didn't see this thread, I was thinking about such things a couple years ago:

http://typophile.com/node/16620

I ended up not pursuing it, for similar reasons that James Arbogast suggested - it's a lot of work to set something like that up, and it's not clear the results would be very rewarding.

openfontlibrary.org has been disappointing, partly because they haven't reached anywhere near parity with commercial sites for just showing the font, and partly because there's no emphasis on promoting high-quality fonts above random dross. The new version may well turn out better in both regards.

I still think something really cool could happen, but somebody has to have the fire to push it. That won't be me, sadly.

Dan Gayle's picture

The fire, or the time and money?

Like @Schrijver said, getting a decent financial backing really doesn't hurt.

So the trick is, Who?

gaultney's picture

Hmmm. My (albeit) limited experience and exposure to free/open software projects, both big and small, is that 'success' often coincides with three things:

  • The presence of a single leader who protects the vision and quality of the software. In the case of really big projects, this may be a handful of people who work very closely together. Nothing goes in without their stamp of approval.
  • The existence of a clear, well documented means to contribute to/participate in the project, even if that is very basic.
  • A sensible licensing model, including how to handle copyright for contributions.

I think the same would be true for a quality open font project. A single designer needs to be the Tzar - even if that slows down the project. Whether the project has 1 or 100 contributors, someone needs to take design responsibility. That's true even for commercial foundries, where one person adds glyphs, one kerns, one write OT code, etc.

We shouldn't assume that all free/open projects are even interested in outside contributors. Many are, but some people choose open licenses as a way to distribute the software and allow derivatives, not to get others involved. With a good FOSS licensing model, people can make their own derivatives if the speed is too glacial or the designer wants to do it all themselves. (Obligatory apologies to everyone for the long Gentium development cycle, which does involve a number of people.)

One quick clarification on Gentium. Although it does support the goals of SIL International, it was originally my personal project, and later embraced by SIL. Those goals, BTW, go much further than bible reading - linguistics, literacy, literature production, language preservation and the like.

Dan Gayle's picture

@gaultney
If you could do it all over again, what would you change? Specifically, the things surrounding the development of Gentium or the organization of others working on it, etc.? Your unique insight on this would be greatly appreciated, since you're one of the few who has one of these projects, even if the speed is glacial.

Schrijver's picture

@James you speak about the FOSS community like an entity that you can communicate with as a party, this is something I saw more often in these threads—Why doesn’t the FOSS community pose a more clear brief for what they want?

But I think FOSS is no single entity, it is just the sum of all people who happen to be involved in FOSS projects. This might be a strength (because it allows it to fill very specific needs proprietary software will never be able to address), in these circumstances it shows as a weakness: you can’t ask FOSS anything, at least not in the same way you would ask a costumer or a company.

I am sorry to hear of the disappointing experience, though.

@Raph & @Victor I really apreciate your taking the time to comment on this thread. @Raph your site has been a great inspiration for me thinking about which role open source can have in type. @Victor I apologize for my simplistic articulation of your project—I think it is an unprecedented accomplishment!

But this brings me to a specific point, which I myself failed to address in the examples I gave…
It seems there is a consensus on the need for a Tzar.
I definitely agree. But IMHO the radical thing open source changes is not so much the possibility to have one Tzar with a lot of footmen, but the ability to have Tzar’s easily pick up where they deem other Tzar’s to have stopped.

And what to me is evident, is that this how culture works anyway. It is like making a revival. There was a designer with a vision. You build upon what he has done, and in the process make it your own. And the original still exists.

The thing open source can do is take that process, but allow it to take place much more rapidly, because you do not have to wait more than a century for something to pass into the public domain.

We shouldn’t assume that all free/open projects are even interested in outside contributors. Many are, but some people choose open licenses as a way to distribute the software and allow derivatives, not to get others involved. With a good FOSS licensing model, people can make their own derivatives if the speed is too glacial or the designer wants to do it all themselves. (Obligatory apologies to everyone for the long Gentium development cycle, which does involve a number of people.)

Exactly. We need to get over the fear of open source being about wanton collaboration blurring typographic vision. It can just as well be a catalyst in enabling the production of a multitude of diverse and highly individualist expressions.

Licensing your own work under on open source license to me says: I presume this can be valuable to other people in a way I don’t know yet. Or: I presume this can be valuable to other people in a way I can not concentrate on myself.

A great example I think of this reasoning by @James releasing his Fry glyphs.
Knowing he was not going to have the time to make a font out of it anyway, giving it away makes a lot of sense: you spend time making something nice.

And while we have been talking about these things on a hypothetical level, these ways of collaboration are actually taking place, showing that at least in part the lack of these kind of projects has had to do with a lack of suitable technology. With versioning systems going mainstream, and the UFO format as an XML based font description format, the technical groundwork is there IMHO.

concretely

@James donated the Fry glyphs late January, @Klepas started a project a few days later, the files were up on GitHub in the UFO format Feb 8th. While I was busy writing on typophile last weekend, I mentioned a student I tutor the Open Baskerville URL in an e-mail. The next day he (Rob) had forked the project, and had actually started drawing glyphs for it!

So this week he has been making his additions to his own fork
http://github.com/rbmntjs/open-baskerville/commits/master
And friday @klepas pulled Rob’s additions back in to the original repository
http://github.com/klepas/open-baskerville/commits/master

The workings of GitHub are also explained on the homepage http://klepas.org/openbaskerville/
Seeing how the UFO format is XML-based and concise, it is perfect for versioning.
But of course looking at text is still not a very pleasant way to browse a fonts assets.

But every different glyph has it’s own concise XML file, and translating that to SVG with XSLT or the like should be perfectly possible.

So even though there is no font yet, at least we have a proof of concept that this way of working is technologically possible.
It also shows the power of open licensing. @James donated the glyphs without knowing @Klepas would pick them up. @Klepas started the project without knowing Rob would start to contribute. And if @james is still coming to Robothon, we might be able to have a look at what his glyphs have been up to, a mere month after he let them loose, since Rob will also be attending.

For me this experience combined with the substantial way in which we have been able to explore these ideas on the forum this week, gives me the incentive to say, that for now I will try and light a fire under the open source font movement…

I’ll get back on this :-)

__
art directing the internet in 2009

gaultney's picture


If you could do it all over again, what would you change? Specifically, the things surrounding the development of Gentium or the organization of others working on it, etc.?

Presumably, if there was something I wished were different, I'd change it. I'm still directing the development of the project, so I have that control.

We solved the licensing problem - almost - with the SIL Open Font License, but there are still issues surrounding how we handle copyright for contributions. Do we require contributors to assign the copyright to us? Do we include their copyrights alongside ours? Do we offer both? Each of these has troublesome aspects and requires administrative effort.

The most challenging issue is managing those contributions. Gentium is a big enough project that I have a hard time letting potential contributors know what holes still remain and what types of contributions I might appreciate. Then how do you manage the actual contribution process (communication, submission, review, incorporation), handle formats and metadata, keep track of contributed elements? You also need everything submitted in all styles (regular, italic - with interpolation-friendly bold and bold italic). And when you hit technical problems (overlapping contours, duplicated knots, other bugs), who fixes them? It all becomes a massive investment in administrative time, which - frankly - I'd rather spend designing.

Having said that, I know that the software development community has made great strides in this area. Textual glyph formats such as UFO have held great promise, and I've tried to use them, but I've yet to see a system that reduces the admin load for larger projects. A system that does would not only serve the open font development community, but be useful for the commercial font foundries, both large and small.

V

Randy's picture

Re: Managing Contributions

First, hi Victor! It seems these discussions always break down into how fonts can't work in an open source software model. I get that. The recent swing in the conversation to version control and contribution management is welcome. It seems to me that this is where the open source community could really help a collaborative font effort, and type designers generally.

In recent discussions we font folk have been daring open source folk to provide a design brief. It might make more sense for us to give them a design brief for the web application required to pull it off. Surely the result would be useful to the type design community for commercial projects as well. If we each stick with what we're good at, it just might be possible for both parties to benefit.

A very crude idea about what such an application might look like:
I'm envisioning a web application that is the centralized repository for the font project. In my "limited" experience with collaborative font projects, removing the sprawl from contributor hard-drives seems vital. Having a central visual reference that shows missing puzzle pieces would be equally crucial. Permissions would define the roles suggested (Tzar, Monkey etc tbd). Contributors could check-out "under construction" glyphs for a limited time (like a library). When you check out the "germandbls" glyph you get a package containing: any existing versions of the glyph, notes from the Tzar, links to design resources for the glyph (see diacritics project.cz etc) and any components that would be helpful in designing the glyph (in this case the lc f and s). Tzars could communicate vision and revision effectively. Clearly thought needs to be put into the format of the files. Keeping everything online would encourage participation: an online glyph editor with package contents readily available for reference seems ideal, but difficult to implement. Package download, edit, upload seems easier, but more limiting ( perhaps a UFO->EPS->UFO or UFO->VBF->UFO on a glyph by glyph basis to allow vector drawing programs to get in on the mix? I don't know.

The main goals would be:
• For Tzars, maximize single-ness of vision and minimize admin headache
• For Monkeys, breaking tasks into well-defined, bite-sized chunks (one glyph, or group of glyphs)

Obviously developing this is a gargantuan task. But many hands make light work, right? Haha. Thoughts?

James Arboghast's picture

@Eric: @James you speak about the FOSS community like an entity that you can communicate with as a party, this is something I saw more often in these threads—Why doesn’t the FOSS community pose a more clear brief for what they want?

But I think FOSS is no single entity, it is just the sum of all people who happen to be involved in FOSS projects.

This is the essential problem. FOSS is not organized. FOSS is an unfulfilled pie-in-the-sky dream with no clear goal, and you're the latest dreamer to come along. Very boring waste of my time.

This might be a strength (because it allows it to fill very specific needs proprietary software will never be able to address), in these circumstances it shows as a weakness: you can’t ask FOSS anything, at least not in the same way you would ask a costumer or a company.

Then what is the point of having a non-organization like the OpenFontLibrary? No point answering that one because I'm abandoning this thread. It's going nowhere.

No cop, no stop
I don't care

Every one of you could be the same
Every one of you can play this game
Got a green light, got a green light
But you're going N O W H E R E !

You know the best things in life
Aren't for me
You know the best things in life
Aren't for free
—words by Richard Patrick

What are you yourself going to do about the lack of organization, lack of discipline, lack of clearly defined goals? Spin out more of your mind-numbing FOSS evangelism? The amount you've been posting on this topic is so lengthy and undisciplined --- professionals such as myself, Nick Shinn, Ray Larabie and Hrant Papazian *do not have the time to read that much blather*.

If you want this to happen, learn to be efficient and present well. I can't take you seriously. The way you added those bold text edits to your thread header was really irritating, galling, a major piss-off.

I am sorry to hear of the disappointing experience, though.

I am sorry I ever came into contact with you. Don't ever come near me again.

Bye.

j a m e s

Dan Gayle's picture

@Randy
You know, I bet a person could get on with a school to develop something like that. How many design schools could use something like this for projects? I bet a bunch would, IF it existed. Getting the edu system involved is how many FOSS projects take off. It's something worth thinking about, for sure.

Zara Evens's picture

The way you added those bold text edits to your thread header was really irritating, galling, a major piss-off.

I, for one, appreciate the typographic sensitivity displayed in these posts. If you don't like it James, go away. That is like someone demanding you stop using so many cite tags, and certainly stop putting spaces between the letters in your name. It is ruining my day, those spaces.

The amount you’ve been posting on this topic is so lengthy and undisciplined

How are your posts any different?

This is an important discussion to be had, and if anyone wants to spend the time writing lengthy essays on their points of view, then by all means let them. Please stop discouraging people from posting here.

Si_Daniels's picture

>and certainly stop putting spaces between the letters in your name. It is ruining my day, those spaces

Zara, that was a low blow! You know James is from down-under where lots of people, you know, like to, er "letter space lowercase." ;-)

Zara Evens's picture

Si, I know. I am pure evil :)

Nick Shinn's picture

I management is an issue, how about a spec that minimizes monkeywork:

- monoline with rounded terminals
- monowidth
- disconnected script

fredo's picture

>and certainly stop putting spaces between the letters in your name. It is ruining my day, those spaces

But you're cool with constant references to grunge lyrics‽

ƒ

Randy's picture

Nick: I agree that manageable projects are a reasonable first step. Though monoline/rounded terminals are not particularly easy to do well.

As for the collaboration application, I'd hope it would work well for managing a 4 style family with modest glyph set. Gentium would certainly push the limit of any manager or tool. But I can see SIL being excited about giving small language groups a tool with just enough oversite to organically generate their own extended language support or new creations - but realistically this would require an online glyph editor, training and extensive wizardry I've not given thought too :-) Such as the difficulty of localizing such an app for such a people group. Back to the *modest* glyph set!

dberlow's picture

Wow! Any sign of a spec in here?

I'll keep checking. Cheers!

Nick Shinn's picture

...monoline/rounded terminals are not particularly easy to do well.

I was thinking that the face could be generated metafont style, by applying a stroke width to a path.
That would create no-brainer round terminals. Alternatively, the pen "circle" tool could be used.

Schrijver's picture

A reproach can only hurt if it hits the mark. Whoever knows that he does not deserve a reproach can treat it with contempt.
__
art directing the internet in 2009

Schrijver's picture

for the web application required to pull it off.

@Randy

Thanks for chiming in!

I agree with you in doubting a ‘spec’ is all that important. Like if there would somehow appear a spec for a FOSS font family, and we would have some money, some computers, some drinks and a few weeks in a nice house near lake Como, type designing hotshots could just conjure up a nice and flexible type family, give it away, and forever end the need for free fonts and typographic collaboration in general.

I am interested in a distributed model of authorship, for which open source and collaboration happen to provide the means. I think these models provide benefits that can not be attained through the traditional ways of working. If I just wanted a typeface I would buy it (really!).

Like @V said, the tools needed for such a way of working could just as well be suited for commercial foundries.

But did any of you check out the links I posted?
Open Baskerville is working in this sort of way right now.

http://github.com/klepas/open-baskerville/commits/master

I am really curious of what you think of that. Because GitHub does ALL the things @Randy describes, save for visualising the UFO data as an EPS, SVG, PNG, whatever. Which really is the most trivial part to implement. I will try and give it a go the coming week.

Contributions

With regards to copyright of contributions, there are established practices in the open source world, I found this one to be pretty concise: http://www.movabletype.org/opensource/contributions.html

But I do see @V how that generates a large amount of amount of administrative hassle. Though I guess it is not necessarily designers that would be handling this. And like @Randy says, it makes sense to expect some collaboration on these matters from the Open Source world.
__
art directing the internet in 2009

Schrijver's picture

BTW for contempt read indifference, I did not want to use a word which was so much like resentful.
__
art directing the internet in 2009

Randy's picture

No I didn't visit the GIT links until now.
- -
I am really curious of what you think of that. Because GitHub does ALL the things @Randy describes, save for visualising the UFO data as an EPS, SVG, PNG, whatever. Which really is the most trivial part to implement. I will try and give it a go the coming week.
- -
First reaction: step one is opening the command line to launch fink to...
sorry, no thanks.

This is a collaborative coding model applied to collaborative fonting. Certainly there are some aspects of the coding model that should inform collaborative fonting. However, this tool would only be helpful to someone steeped in the coding world. Who is your audience? Who are your collaborators? For this to be a success it needs to look and work a lot more like the latest release of Wordpress. Sure you could blog from the command line, but 99.9% of the people would never do it without the slick UI. Make it work for muppets like me who cant operate a computer without a mouse. Please? Short of that, this effort will fizzle on the launchpad.

Second reaction: If I truly expect the Open Source community to do anything for me, I'd better be willing to enter their world a bit, if only to speak the same language. In this case that would mean bucking up and actually using GIT. Plus, it makes requests much more effective.

- - -
EDIT: BTW, I wasn't saying a design spec from the open source community is not important. I definitely agree that it is, if any useful type is to be developed. I was hoping to diffuse some of the US versus THEM mentality by pointing out that similarly, creating useful type in a collaborative way would be greatly aided by a spec for better tools (which the Open Source community could deliver).
- - -

Nick: I now understand what you meant.

Jens Kutilek's picture

But did any of you check out the links I posted?
Open Baskerville is working in this sort of way right now.

I did take a look at the Fry’s Baskerville project (you can download the files without installing git or anything).

I agree with you in doubting a ‘spec’ is all that important.

But without a spec, without somebody who wants – and is able – to judge the quality of contributions, this is going exactly in the direction I expected. In the worst case, the additions are sub-standard, like you can see in the image below. In the best case you end up redrawing someone else’s glyphs (Fry’s in this case). I'm sceptical about this kind of "open" collaboration, but of course I'd love to be proven wrong, by a well-executed project, but again this is not it.

The first line is original, as drawn by James after the scans. The other two lines are the contributions by your student. I'm sorry to say most of them are badly drawn (more visible in close-up) and some don't follow the original style. Which isn't a bad thing in itself – everyone was a beginner at some point – but you have to keep in mind these glyphs haven't been posted on the TP critique section, where probably some people would have been glad to help improve them, but they are presented as the progress in a project which seems to aim at no less than revolutionizing the way collaborative typedesign works. As I said, all contributions would be measured to the original Fry’s Baskerville, so the possibly best result would still be reproducing someone else’s work.

Do something original instead.

Like @V said, the tools needed for such a way of working could just as well be suited for commercial foundries.

Maybe, but I guess the established foundries already have their internal production systems, and I doubt they would be willing to publicize large parts of them in order to get contributions from FOSS people.

Are you familiar with FontQA? That is an open-source project for font quality assurance licenced under GPL. I'd have to ask, but I think there were not many contributions, if any at all, from anybody not directly involved in the project. Why is that? I think the audience is too small, font production is a too specialized area in which a few people already have their own solution, while the majority won't even see the need for such a tool.

It may be the same for the proposed "visual glyph versioning and font production system". It would have to be developed by people who really use it, and know a thing or two about professional font production. And then probably it would be next to useless to other people whose collaboration workflow differs.

Schrijver's picture

@Randy
If I truly expect the Open Source community to do anything for me, I’d better be willing to enter their world a bit, if only to speak the same language. In this case that would mean bucking up and actually using GIT. Plus, it makes requests much more effective.

Hear Hear! This is exactly the conclusion I came to myself, and the attitude I would like to be propagating!

Also, I think you are the first to on this thread to articulate the possibility of Open Source actually doing anything for you. Great! Because why would you want to have anything to do with it otherwise? Speaking of motivations, I tried to formulate mine in the form of a concise list @ my user blog

I would phrase one thing differently now though: the part that struck me about Open Source is IMHO a specific subset of open source, and I guess it is the part that has come in contact with both creativity and business. IMHO this is the world where the investment of entering a bit has a probable return of investment.

Ultimately, the utopian super-easy GUI programs might come along—but for that we need to have entered a bit more, and a bit longer: and we would have had to have tried to redecorate the house up to the point where we feel comfortable sleeping over…

As example GH again, they promote themselves as Online project hosting for Open Source and Business code. They have a very clear business model, because public code is free to host, but they reason that the user experience will make you also want to host your private code there. So basically, Open Source development is the freemium they offer: if that does not show how the tables have turned, what will?

Where creativity comes into play IMHO is that the whole affair is steeped in the individualist logic of web 2.0. You want your own version of a package, and you want to be able to show your friends what you have. That might sound superficial instead of creative, but the underlying needs driving it are comparable IMHO: the struggle between the need for differentiation and self-expression on the one hand, and on the other hand the need to feel accepted by others.

Part of this echoes what @sii said earlier, about how OSS fonts will probably come about because of those interested in economical and social opportunities joining in as well. In that sense Open Baskerville might also be a good example, because even though I might be pointing you at this open source project, I myself have no background in open source, and the three authors involved in OpenB up to this point have a background in design, not in open source development. Now they might be mistaken by the opportunities these projects offer them, but the fact that they do so shows a significant change.

Yet arguably, the point of discussion is not on that, but rather on the question: to what extent does this impact the practice of the cutural industry (and of open source) as we know it? Because I think these interdisciplinary developments will fundamentally impact all parties involved, even if specific members choose not to investigate these fringes.

&I do not agree with @Jens, but will come back on that.

Eric
__
art directing the internet in 2009

PS:

I was going to ask GitHub if they could maybe support some of my ventures, and seeing how I have been such a fanboy for them in this thread, I figured that it makes sense to state this now to avoid future confusion: it is not GitHub per se that constitutes the promise for these kind of models. They just happen to bring together two things I strongly believe in: empowering non-geeks by allowing them to use very specific technology, by focusing on what you can do it instead of what it does (fork it!). And propagating a distributed model for content, as opposed to a centralized one.

Schrijver's picture

I was hoping to diffuse some of the US versus THEM mentality by pointing out that similarly (…)

Another point to which I agree, and to which I would like to add a similar analogy:

*If we ask the Open Source development world to respect the amount of skill and dedication which goes into the creation of a typeface, IMHO we should try and adequately asses the skill and dedication they bring to their software development.*

Because with all the OFLB reserves flying around, isn’t it also obvious to everyone that their new site proves to be a major accomplishment in these areas? I find the design frankly to be the single best looking site I have seen in ages. Which is an even more able accomplishment when you realise they build it around MediaWiki—really, the blue and white one. They also managed to mash this up with CCHost, which should be a major technological pain in the -, and there is some superspiffy Javascript going on with the live glyph previews.

__
art directing the internet in 2009

Syndicate content Syndicate content