Font copyright law changes coming to U.S.

245

Comments

  • @Thomas Phinney and @Georg Seifert - Riddle me this . . .

    How difficult would it be to roll a 100% open sourced/crowd sourced text base font editor?

    All things being equal, it would need a text entry window (perhaps character, metrics, font info and features dropdown menu) and a live shape preview window in the UI and good amount of tutorials for how to create letters fully via text.

    Of course it'd need to have some manner to which it could import and export from other font file formats.

    As I watch modern developers code in text, there's a good deal of auto-complete and adding the proper headers many times are a one-click situation. There's no reason it couldn't load a font, auto propagate the necessary code that is required for each shape any differently than if there was a font designer who wanted to learn and type that code by hand with some guidance.

    At some point I could see Tal Leming sitting in a US courtroom demonstrating how he creates fonts in a 100% text based editor if push ever came to shove but I think this would be enough to satisfy the copyright office.

    Ultimately, if such a tool existed and could be pointed to as 'this is the font editing application that I wrote my font in' I think that solves our problem.

    At present, there simply isn't such an application and it'd be clear to the copyright office that all a designer did was an XML dump and as we saw in the letter from the copyright office, X/Y point data isn't an acceptable format.

    I think it's much easier to meet their criteria than it is to push a rope uphill as there isn't a single font designer who has a personal relationship with a US senator who would bring this to the floor for a vote, it's just not going to happen.
  • John HudsonJohn Hudson Posts: 1,617
    Joyce, right, the design patent law is not at threat, but on the other hand, I'm not sure whether it has ever actually been put to the test in a court case. Adobe were hoping that it would be during the SSi suit, but the case never went that far.

    Historically, the existence of design patent protection has been used by USCO as one of the grounds for denying copyright protection to typeface design.
  • Mark SimonsonMark Simonson Posts: 1,105
    edited August 2018
    It seems that the copyright office has decided that XML files don't meet their definition of a "computer program": 

    A “computer program” is a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.

    The 1998 Adobe v. SSi suit decided that digital fonts do meet this criteria. Given this, I wonder what they would accept? Or are they are overturning or disregarding the earlier decision? Can they?
  • Khaled HosnyKhaled Hosny Posts: 245
    edited August 2018
    @Thomas Phinney and @Georg Seifert - Riddle me this . . .

    How difficult would it be to roll a 100% open sourced/crowd sourced text base font editor?

    Hmm, MetaFont?

  • Kent LewKent Lew Posts: 867
    Given this, I wonder what they would accept?
    For cubic bezier glyph outlines, I suppose we could submit a set of PostScript instructions. OTF Layout feature files would also seem to meet the criteria, since they are instructions in nature. Including the {kern} feature rules. Right?

  • Hmm, MetaFont?

    That would count, but that's not going to be of much help for CFF and/or TrueType fonts, which I would assume account for some 99.99% of fonts on the market, even among Teχ users.
  • John HudsonJohn Hudson Posts: 1,617
    Mark,
    The 1998 Adobe v. SSi suit decided that digital fonts do meet this criteria. Given this, I wonder what they would accept? Or are they are overturning or disregarding the earlier decision? Can they?
    I'm not sure how this works in the US. The relevant document from the Adobe vs SSi suit is a summary judgement written by Judge White; this is a summary of pre-trial findings made by the judge to determine whether the case has merit. Judge White determined that the positioning of outline points was a creative act within the provisions of the copyright act. At that stage, SSi opted to settle with Adobe, so the case never went to trial. I don't know what the status of a summary judgement is in US case law, or how it relates to decisions of the USCO. My guess is that the summary judgement might only come into play if someone were to sue the USCO over the new decision and cite the Adobe vs SSi case.
  • John HudsonJohn Hudson Posts: 1,617
    edited August 2018
    Kent,
    For cubic bezier glyph outlines, I suppose we could submit a set of PostScript instructions. OTF Layout feature files would also seem to meet the criteria, since they are instructions in nature. Including the {kern} feature rules. Right?
    The USCO definition of the computer program is very broad: 'a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.' That 'directly or indirectly' clause covers a lot of different kinds of software, not all of which consist of manually input code. It seems to me that the insistence on manually written code signifies that the USCO is trying to treat computer programs as a species of literary work, rather than as, um, computer programs.

    At the time of the Adobe vs SSi suit, most of the people who strenuously denied that the software copyright extended to the outline definitions accepted that it applied to hinting, and I don't think there would be any debate that it applied to OpenType Layout (although something like .fea source code might be required under the new regulations). At the time, I and some other folk on the comp.fonts Usenet group made the case that if hinting constitutes a computer program under the USCO defintion, then the outline definitions must too, because hints apply to the outline points, and if you move the outline points around while not changing the hints you 'bring about a certain result' that is different from the result if the points are somewhere else. Hinting does not exist independent of points. I was gratified to see my argument repeated almost verbatim in Judge White's summary judgement after he visited Adobe and observed the process of creating digital fonts first-hand.

  • Kent LewKent Lew Posts: 867
    [minor correction, FWIW: Judge Whyte]
  • Kent LewKent Lew Posts: 867
    edited August 2018
    I don’t disagree with your position vis-a-vis computer code. My suggestion was more along the lines of: what if we were to try to give them what they want. (Given that they are rejecting an XML representation of the actual font software code.)
  • Mark SimonsonMark Simonson Posts: 1,105
    edited August 2018
    My guess is that the summary judgement might only come into play if someone were to sue the USCO over the new decision and cite the Adobe vs SSi case.

    That's what I was wondering, and I think you may be right.
  • Dave CrosslandDave Crossland Posts: 1,040
    edited August 2018
    Kent,
    For cubic bezier glyph outlines, I suppose we could submit a set of PostScript instructions. OTF Layout feature files would also seem to meet the criteria, since they are instructions in nature. Including the {kern} feature rules. Right?
    The USCO definition of the computer program is very broad: 'a set of statements or instructions to be used directly or indirectly in a computer in order to bring about a certain result.' That 'directly or indirectly' clause covers a lot of different kinds of software, not all of which consist of manually input code. It seems to me that the insistence on manually written code signifies that the USCO is trying to treat computer programs as a species of literary work, rather than as, um, computer programs.
    Computer programs are literary works. They even look like poems. 

    Indirectly, like Python code which goes through several layers of virtualization and real time compilation, to get executed; vs directly like a precompiled binary executable.


  • @Thomas Phinney and @Georg Seifert - Riddle me this . . .

    How difficult would it be to roll a 100% open sourced/crowd sourced text base font editor?
    Very far from trivial if you did it from scratch.

    If you don’t care about visualization while editing, then it already exists, arguably in multiple forms. ttx/FontTools being one of them.

    But I am thinking that you want a visual preview of the effects of your editing while you are doing it. That gets harder. But we have been working on it already.

    Making the tool open source is a separate issue, of course. We are happy to have open interchange formats, but are inclined to make features part of the FontLab app(s).
  • @Dave Crossland I think you nailed by suggesting fonts are data vs an executable application. I'm curious how then data files are protected by the copyright office - An example would be filing a copyright on a completed Excel spreadsheet for instance (regardless if it contained custom Macros or not).

    @Thomas Phinney you're definitely on the right track based on my suggestion. And initially I thought perhaps it could be a plug-in component to an existing font editor however that doesn't entirely prove the font was initiated in a text based editor but rather was an afterthought to a graphical editor and gives me pause that lots of font designers may wantonly check the 'I hand-typed this font in myself' box knowing that was not the case just to get the application accepted.

    That said, all the font copyrights I've ever submitted are generated via TTX and the copyright office is now saying, that output is no longer acceptable since it is as Dave suggested 'data' instead of being a fully executable application. Does a format like UFO solve this?

    Just for fun I dropped a random .OTF font into a text editor and it's clear that it's been compiled. I'm curious what a pre-compiled .OTF font looks like that is not generated from TTX -  THAT is the format we'd need to submit that could be approved as I understand it.

    Or is pre-compiled OTF source code specific to each unique font editor? I'd love a software engineers perspective on this since it's beyond my knowledge base.

    All of this brings me back to my initial suggestion that we'd need a bespoke text editor for writing fonts just so we could submit its full source code to the copyright office as an application.
  • Dave CrosslandDave Crossland Posts: 1,040
    edited August 2018
    Personal opinion only:
    all the font copyrights I've ever submitted are generated via TTX and the copyright office is now saying, that output is no longer acceptable since it is as Dave suggested 'data' instead of being a fully executable application. Does a format like UFO solve this?
    No, I don't think it matters what format of data you submit - a TTX XML, UFO XML/PLIST, GlyphsApp pseudo-plist, TruFont JSON, would all be the same: not a computer program for copyright purposes.

    Not program source code that can be compiled or interpreted and executed, but data that is parsed by programs and used to instantiate a typeface - and thus not acceptable within the copyright category of "computer program."

    But I think this category is a red herring. 
    Just for fun I dropped a random .OTF font into a text editor and it's clear that it's been compiled. I'm curious what a pre-compiled .OTF font looks like that is not generated from TTX -  THAT is the format we'd need to submit that could be approved as I understand it.

    Or is pre-compiled OTF source code specific to each unique font editor? I'd love a software engineers perspective on this since it's beyond my knowledge base.

    All of this brings me back to my initial suggestion that we'd need a bespoke text editor for writing fonts just so we could submit its full source code to the copyright office as an application.
    I don't think it matters at all how the data is authored - line by line in a text editor, or using a GUI. 
  • John HudsonJohn Hudson Posts: 1,617
    edited August 2018
    Dave,
    Computer programs are literary works. They even look like poems. 
    What they look like is irrelevant. They're not literary works because they don't exist to be read: they exist to be used directly or indirectly in a computer in order to bring about a certain result. So by the USCO's own criteria, computer programs are performative — they do things —, and it seems to me that treating them as literary works, as if the writing of the source code is the medium of the work rather than what the program does, is indicative of a bureaucracy that still has an essentially 19th Century paradigm for creative works.
  • John Hudson said: 
    by the USCO's own criteria, computer programs are performative — they do things —, and it seems to me that treating them as literary works, as if the writing of the source code is the medium of the work rather than what the program does, is indicative of a bureaucracy that still has an essentially 19th Century paradigm for creative works.
    But fonts do not do things in the way programs do. Yes?
  • I'll second Dave @John Hudson - The copyright office clearly defines that computer software must generate a result and that fully executed result is what is protected.

    @Dave Crossland - A thought experiment - What if somebody developed an open source application that when launched simply displays a character showing and the font data was fully embedded in the application? Does this mean the font data is now protected because its part of what is being rendered in the output?

    In part, as absolutely stupid as it sounds, it could become a new font format solely for the purposes of filing copyrights that is an executable application. In part, it also means not only would the file format need to be fully executable but it may require somebody build a host application that can use these files as 'fonts'

    My point being, I think once we can pin down the copyright office's requirements, it should be a trivial matter for us to generate 'fileable' software which satisfies their requirements while giving our works the same protection as current copyrighted fonts enjoy unless we find this alternate format wouldn't protect the font data when it's used in OTF, TTF, etc formats.
  • Dave CrosslandDave Crossland Posts: 1,040
    edited August 2018
    I'm not sure anything other than an OpenType font is worth discussing - you could fairly easily convert an OpenType font to a MetaFont program (sort of like what happened to the Zapf math symbols font, the students gave up trying to write mf code that replicated Hermanns design and just used it to plot outline contours) but since no one would use your MetaFont nor your application bundle, I don't see how doing so is useful to you. 

    I remain skeptical that copyright registration matters much in the first place, since unregistered copyrights exist, and I can't imagine that - as Joyce was saying earlier - when actually attempting a licence enforcement action, it really matters if it's registered or not.

    That being said, how does anyone register a contemporary application for copyright? Thomas? :)
  • I hope I'm not about to embarrass myself...  I'm business side and I know very little about the technology.  But I also know that stupid questions from lay people can be useful in this sort of conversation.  So, a thought experiment:

    I'm really stuck on this idea that fonts are mostly data.  I'm sure that, objectively, fonts have a lot of data. But, I think what matters is the comparison to other software.  I keep thinking it can't have that much more on a percentage basis than a game.  Games have a lot of code that simply describes the visual landscape (so similar to fonts).  Would anyone argue that a game isn't a program?  

  • Fonts are run-time applications, functionally not much different than a DLL (Windows, dunno what they are on a Mac, et al).

    Run-time applications do not execute code on their own. They are passed instructions, queries, et al and return a result, action, instruction (and more). Then the requesting application does something with that result.

    A for instance is an OS language display. A user chooses a language in the OS or an application. That is passed on (typically) to a DLL that then does whatever (translation tables typically) and a DLL returns the result to the requesting OS or application that then displays the result.

    A font gets a request from an application or OS. It then returns a result after which the OS or application displays the result. (Or prints, etc, the result).

    Computer applications that are copyrighted have that CR (can have that CR) applied to its run-time "applications" as well as core application code.
  • Games typically consist of multiple files. The actual executable component of a game is typically a small percentage of the total size, with the remainder being data files. So, for example, Civilization V on the Macintosh is 7.22 GB but the actual executable is only 41 MB (and I suspect much of that is actually embedded data as well) (I'm not much of a gamer -- that's the only game I have installed so I don't know how typical this ratio is).

    All of the visual landscape will be described by various data files which will often be standard graphic formats such as .png, .jpeg, etc. The code will be responsible for composing the various bits into a coherent whole, but it rarely generates the actual graphics themselves.

    I suspect the graphics themselves are subject to copyright because they fail to fall under the category of "useful objects" but rather are treated as art.

    Personally, I think the most productive route for those in the type industry to pursue would be to convince the copyright office that fonts, unlike alphabets (and abugidas, syllabaries, etc.) are not purely functional objects but rather constitute artistic interpretations of the alphabet, i.e. that fonts themselves are something worthy of protection.

    Convincing them that fonts are in fact code seems to me like a stop-gap measure; if the copyright office isn't convinced that fonts are worthy of protection, they can always find ad-hoc ways of excluding fonts from being software even if a non-XML-based text-only format for representing them is developed.
  • Kent LewKent Lew Posts: 867
    the most productive route for those in the type industry to pursue would be to convince the copyright office that fonts, unlike alphabets [...] are not purely functional objects but rather constitute artistic interpretations of the alphabet, i.e. that fonts themselves are something worthy of protection.
    Unfortunately, André, this is exactly the route that has not proven productive throughout the history of the U.S. Copyright Office and fonts.

    There is long-standing precedent against it. So, there’s quite a bit of history to overcome first — bureaucracy being what it is and all.

    I recall some back and forth with Frank Martinez, while I was on the board of SOTA, about a renewed attempt to make this argument, and an outline for legislative language that would start to grapple with the parameters of such a protection. I don’t recall if Frank was ever able to push that effort all the way to the legislature or not at that time.

  • Thomas PhinneyThomas Phinney Posts: 1,472
    edited August 2018
    I'm not sure anything other than an OpenType font is worth discussing - you could fairly easily convert an OpenType font to a MetaFont program (sort of like what happened to the Zapf math symbols font, the students gave up trying to write mf code that replicated Hermanns design and just used it to plot outline contours) but since no one would use your MetaFont nor your application bundle, I don't see how doing so is useful to you. 

    Although this makes superficial sense to me, nobody has been submitting the final-form compiled thing anyway. Font folks were mostly submitting a ttx file before, which was *also* not the font binary. People copyrighting apps submit source code, not the compiled app. So...?
    I remain skeptical that copyright registration matters much in the first place, since unregistered copyrights exist, and I can't imagine that - as Joyce was saying earlier - when actually attempting a licence enforcement action, it really matters if it's registered or not.
    Well, it wouldn’t matter for the license enforcement *directly*, but I would think that the lawyers on both sides knowing that the licensor could, or could not, also be pursued for copyright violations could make a difference in how they approach the negotiations and what they are willing to do. And of course, copyright (or something else, other than licensing), is exceedingly helpful in that it makes piracy illegal. Without it, there is nothing illegal about hosting pirated fonts, nor about downloading them.
    Dave Crossland said:
    IThat being said, how does anyone register a contemporary application for copyright? Thomas? :)
    AFAIK, everybody submits source code. For the USA, see: https://www.copyright.gov/circs/circ61.pdf (pp. 3–4) “Submit the source code for the specific version of the computer program you want to register.”
  • I remain skeptical that copyright registration matters much in the first place, since unregistered copyrights exist, 
    ... copyright(or something else, other than licensing), is exceedingly helpful in that it makes piracy illegal. Without it, there is nothing illegal about hosting pirated fonts, nor about downloading them.
    But the lack of registration doesn't mean there is no copyright.
  • Mark SimonsonMark Simonson Posts: 1,105
    edited August 2018
    As I understand it, filing makes a difference if you sue in federal court, making it possible to seek statutory damages and attorney's fees.
  • @Mark Simonson As I understand it having done research but not being a lawyer:

    Basically yes, to elaborate... A work that is eligible for copyright has it by default upon publishing.  You have to file in order to seek redress in a court.  If you file before the violation you can seek statutory damages (if you file after you cannot).  

    Further, my understanding is the approval of font applications was always provisional.  That is that the copyright office didn't have time to make a proper decision so was letting us file and with the understanding that a court would make the final decision of validity if the need should arise.  Given that this recent action seems precipitous.  It would be nice to know more about it.  
  • John Hudson said: 
    by the USCO's own criteria, computer programs are performative — they do things —, and it seems to me that treating them as literary works, as if the writing of the source code is the medium of the work rather than what the program does, is indicative of a bureaucracy that still has an essentially 19th Century paradigm for creative works.
    But fonts do not do things in the way programs do. Yes?
    But a python script is not doing anything, either. It is read, parsed and then some functions are executed that do things.
  • But the lack of registration doesn't mean there is no copyright.
    No. But registration allows (in the USA) the copyright holder to collect:
    - statutory damages, where infringement can get compensation regardless of actual loss
    - treble damages, where you get triple the actual loss

    Without registration, the copyright holder is limited to actual damages.
Sign In or Register to comment.