Google Fonts: Your Questions, Answered

Khaled Hosny

No idea why my comment is shown 4 times, but I only posted it once…

Deleted Account

John, I'm definitely not talking about Microsoft's mostly accretive versioning of excruciatingly carefully designed, engineered, quality assured and tested font software with excruciatingly carefully designed, engineered, quality assured and tested font software updates. There would surely be some overlap between MS versioning and Google font licensed versioning, but I cannot go there.

And

But on the Web, document reflow isn't a bug: it's the expected behaviour of the medium.

....this is a gross generality that I respect 1/2 of. Reflow is expected behavior for some text, but reflow of the same text in the same environment in the same font is not. I understand "the web" as a collection of documents made with and without concern for presentation, that is that, and it is forever.

Vernon:

I think there could be a difference between how type designers see type, and how type gets used in the wild.

Really?

I'm pretty certain that making a libre webfont is a completely different thing to, eg, making a font for a corporate client.

Nonsense to me. This "in the wild" thing is out of hand. Typography! is the result of fonts, not fonts in the wild. How some type designers see some typography, is maybe never. How other type designers see fonts is that they never see anything but the typography. This is not dependent on client type, or font licensing liberty.

I've had enough conversations with people who i consider "don't get" libre webfonts at all...

Yeah, but talking yourself into a greasy baloney sandwich, (that is, with no bread)... who's surprised!

i think there's a 'threshold' [...] when you design and create something and then allow it to go free, and go beyond your control.

...like "beyond our control" hasn't been the state of font design for, conservatively, 3o years.

I want to bring up the teachings of Ba DaBing and Doodis-Nawa, but I'm constrained until the release of their book. For now, it's safe to say from the view meter-to-font boo-boo ratio in google fonts, there is a deep pile of typographic boo-boo-doo-doo (type in the wild), and someone is going to have to clean it up.

Deleted Account

Khaled Hosny

No idea why my comment is shown 4 times, but I only posted it once…

So I would read it 4 times, go to the git hub, see what you are talking about, come back, and say: "See, if one developer takes the open source font and makes bigger accents for smaller use, and another developer takes the rest of the glyphs and makes another them smaller for larger use, two fonts of the same family name return to google world as each developer prefers it and two happy users go on. Meanwhile... 1,000's of others don't know to, much less how to, separate the "we" from "the shaft".

James Puckett

No idea why my comment is shown 4 times, but I only posted it once…

You posted a link in a comment right after registering and the spam filter blocked it, which sometime causes double posts. The double post showed up three times. I deleted it.

Scott-Martin Kosofsky

With this 13-page thread becoming well-near incomprehensible, I thought that, in the spirit of the season, we might reframe it as a simple question.

Google Fonts. Naughty or Nice?

Jan Schmoeger

The spirit of the season: faith needs no reason. Peace be with you all.

SiDaniels

You better watch out
You better not cry
Better not pout
I'm telling you why
Google fonts are coming to town
Raph's making a list
And checking it twice;
Gonna find fonts both naughty and nice
Google fonts are coming to town
He sees you when you've been cloning
He knows when fonts are fake
He checks if fonts are bad or good
So be good for goodness sake!
O! You better watch out!
You better not cry
Better not pout
I'm telling you why
Google fonts are coming to town
Google fonts are coming to town

Chris Lozos

Someone must be Sleepless in Seattle ;-)

SiDaniels

No sleep 'til we get a first round bye! ;-)

Chris Lozos

Last week was a big chance at home-field and first round, why did they say bye to it?

Dave Crossland

@John Hudson

This can be to some extent ameliorated through use of the Reserved Font Name mechanism in OFL, which requires that derivatives or forks of the original font be release under different names. ... there are some user's whose openness needs extend to not having a Reserved Font Name clause (presumably because they actually want behaviour variations under the same name).

http://scripts.sil.org/OFL_web_fonts_and_RFNs is required reading on this topic.

I personally consider the RFN inappropriate for web fonts, because it requires everyone hosting the fonts who converts to EOT, subsets them for themselves, or otherwise modifies them, to either (1) rename the font, or (2) get permission from each copyright holder to do so, or (3) convince the author to re-license the font without RFNs

Personally, I haven't reached a conclusion on which of these 3 options I prefer.

SiDaniels

Dave, do you understand Victor's objection to EOT? Like WOFF, EOT is just a wrapper, everything about the base font is preserved, and the spec is public.

John Hudson

I personally consider the RFN inappropriate for web fonts, because it requires everyone hosting the fonts who converts to EOT, subsets them for themselves, or otherwise modifies them, to either (1) rename the font, or (2) get permission from each copyright holder to do so, or (3) convince the author to re-license the font without RFNs

It seems to me that (4) if a font were licensed with two names: one Reserved and one Open, that would satisfy both the aims of the RFN in protecting the trademark of the original and facilitating web use without requiring every subset or otherwise modified font to have its name changed. So, for example, I might release a font with the RFN identified as 'Albatross' but the actual name table entry could be e.g. 'Albatross Web' which would not be separately identified as an RFN. This would preserve the Albatross name as a trademark and enable it to be used for non-open licensing if desired, but would not force the user to change the font name for every minor modification or format change.

Ideally, this would be something explicit in the licensing, and it seems to me that the whole issue as discussed at length by Victor could be resolved by adding an explicitly non-reserved name to the OFL license model.

John Hudson

I see that something like this is already done for some fonts in the Google Fonts directory. The fonts included there as 'Alegreya Sans' have the reserved font name 'Exo'.

Deleted Account

Ya, but it's like the little-French-town-crash-scene in the Pink Panther movie. It's okay if one knucklehead in costume is driving madly around town, but three or four knuckleheads in costume driving madly around town ends in AlbatrossSemiboldItalicWhiteinlineRedglyphBlackdropshadowBanner Perspective10Pitch30Yaw100RotateLatinKanjiGreekWebTTF.woff, doesn't it?

Chris Lozos

"Woould you like a phallic symbol?"

Ben Blom

In an old town, there were ten bakeries. On a good day, one of the bakeries started to give away all its bread for free — not temporarily, but permanently. Also, the supply of free bread would be without limit. It is not known why this bakery started to give away its bread for free, and how this bakery could afford to do so.

Then two wise men came to town. The people of the town asked the wise men what would happen with the nine bakeries that do not give away their bread for free. Surprisingly, the wise men disagreed. One of them said that the nine bakeries will sell more bread in the future, so they should hire new staff. The other wise man said that the nine bakeries will sell less bread in the future, so they should lay off staff.

Thomas Phinney

The big questions are: (1) in the eyes of the previously existing market, how good a substitute is the free bread for the bread produced by the other bakeries? (2) Does bakery location matter in the story?

IFF the market perceives them as completely interchangeable, and location is not an issue, then the nine bakeries will go out of business.

This is the same question with fonts. In the eyes of the people who would otherwise pay for fonts, how good a substitute are the libre fonts for the proprietary retail fonts?

Nick Shinn

So much for the loaves. What about the fishes?

Chris Lozos

The fishes started to smell after 3 days, so they were given away FREE!!!

Dave Crossland

Simon Daniels wrote,

Dave, do you understand Victor's objection to EOT? Like WOFF, EOT is just a wrapper, everything about the base font is preserved, and the spec is public.

Actually, Victor and Nicolas don't say that. Their document says, emphasis mine:

EOT was one of the earliest web font formats and did not initially support all of what is needed to preserve FE. However, more recent changes in both the format and supporting development tools may provide what is needed. Deeper technical analysis of this is needed before a judgment can be made about this format.

- http://scripts.sil.org/OFL_web_fonts_and_RFNs

I believe that is based on conversations I have had with Victor, where I put it to him that EOT could be like WOFF, but the eot command line tools that MS provided on Windows had a default behaviour that made them produce non-FE modifications. Without me asserting this, I expect he would have said EOT was still non-FE compliant.

I wrote,

I personally consider the RFN inappropriate for web fonts, because it requires everyone hosting the fonts who converts to EOT, subsets them for themselves, or otherwise modifies them, to either (1) rename the font, or (2) get permission from each copyright holder to do so, or (3) convince the author to re-license the font without RFNs

John Hudson replied,

It seems to me that (4) if a font were licensed with two names: one Reserved and one Open, that would satisfy both the aims of the RFN in protecting the trademark of the original and facilitating web use without requiring every subset or otherwise modified font to have its name changed. So, for example, I might release a font with the RFN identified as 'Albatross' but the actual name table entry could be e.g. 'Albatross Web' which would not be separately identified as an RFN. This would preserve the Albatross name as a trademark and enable it to be used for non-open licensing if desired, but would not force the user to change the font name for every minor modification or format change.

I'm sad to say that the intent of the OFL authors is not to allow that. The OFL FAQ says, emphasis mine:

Question: 5.4 Am I not allowed to use any part of the Reserved Font Names?

Answer: You may not use individual words from the Reserved Font Names ...

5.5 So what should I, as an author, identify as Reserved Font Names?

Answer: Original authors are encouraged to name their fonts using clear, distinct names, and only declare the unique parts of the name as Reserved Font Names. For example, the author of a font called "Foobar Sans" would declare "Foobar" as a Reserved Font Name, but not "Sans", as that is a common typographical term, and may be a useful word to use in a derivative font name. Reserved Font Names should also be single words for simplicity and legibility. A font called "Flowing River" should have Reserved Font Names "Flowing" and "River", not "Flowing River".

So, if you release a font with the RFN identified as 'Albatross' and actual name table entry is 'Albatross Web' then I will still have to rename it 'Zebra Web' or get permission or persuade you to drop the RFN.

Ideally, this would be something explicit in the licensing, and it seems to me that the whole issue as discussed at length by Victor could be resolved by adding an explicitly non-reserved name to the OFL license model.

Actually, as long as the non-reserved name isn't reserved, and is entirely distinct from the reserved name, that will do fine - and this is indeed (1) as I listed.

In software projects it is very common to develop a codebase with a name different to the trademarked, 'released' name - eg Canary and Aurora/ are codenames for Chrome and Firefox - so I can imagine a future where libre fonts are developed under code names on Github, but the big web APIs distribute them under trademarked names associated with principle designers who have done quality assurance work on them.

I see that something like this is already done for some fonts in the Google Fonts directory. The fonts included there as 'Alegreya Sans' have the reserved font name 'Exo'.

(Ah, that's a mistake; Alegreya Sans is by Juan Pablo del Peral and has the RFN 'Alegreya Sans' - https://code.google.com/p/googlefontdirectory/source/browse/ofl/alegreyasans/OFL.txt - but I see in the font files' NAME tables that they have the copyright notice from Exo. It should be corrected next week

John Hudson

I see in the font files' NAME tables that they have the copyright notice from Exo.

How'd that happen?
_____

Thanks for looking into the RFN issue in more detail. It is a pity that the license writers interpret

Dave Crossland

How'd that happen?

User error when copying NAME table strings.

John Hudson

What I'd tried to write in that previous post:

Thanks for looking into the RFN issue in more detail. It is a pity that the license writers decided that RFN should be inclusive rather than exclusive. It would be more flexible as a mechanism if RFNs were precisely excluded combinations of terms, rather than collections of terms any one of which would make a derivative name unusable.

What I'd really like to be able to do is to define specifically both terms and combinations of terms that are reserved and also uses of those terms in combinations that are explicitly not reserved. What we're talking about here is essentially trademark licensing, and there should be a range of options between 'You cannot use any of these terms in any way' and 'You can use all of these terms in any any way'.

Dave Crossland

trademark licensing

Not using RFNs and using Trademarks instead works for me personally, but some RFN using designers I spoke to about it were not so keen.

SiDaniels

Thanks Dave for the clarification on EOT. The explanation makes some sense, as the Windows implementation of EOT did include a form of obfuscation, which prevented decompressed fonts pulled from the cache on Windows 9x from being installed. However, WOFF and TTF fonts are often and routinely obfuscated, so I still think this comment reveals a bias against EOT - obfuscation is not part of the EOT specification. Thanks for trying to 'splain this to Victor. :-)

vernon adams

David Berlow said;

Nonsense to me.... Yeah, but talking yourself into a greasy baloney sandwich, (that is, with no bread)... who's surprised! ...like "beyond our control" hasn't been the state of font design for, conservatively, 3o years.
I want to bring up the teachings of Ba DaBing and Doodis-Nawa, but I'm constrained until the release of their book. For now, it's safe to say from the view meter-to-font boo-boo ratio in google fonts, there is a deep pile of typographic boo-boo-doo-doo (type in the wild), and someone is going to have to clean it up.

Success and riches gained creating and building the Font Bureau empire, and sat now at the top of the pile. Why the obsession on what the little guys, way below, are doing? Why the anger?

Dave Crossland

a bias against EOT

Well, EOT is more like WOFF2, in that it won't roundtrip TTFs with identical checksums, since it packs the data into more compact syntax, while being semantically identical.

As I understand it, this is a large part of the idea of 'Functional Equivalence.'

SiDaniels

Well that's really an artefact of the microtype express implementation, not the EOT format itself. But, if that's really the show stopper that would be a more serious issue for WOFF 2.0? Thanks again, Si.

Deleted Account

"Why the obsession on what the little guys, way below, are doing? Why the anger?"

Certainly not angry or obsessed. And certainly not sitting on top of a pile. That's the same height as standing on the ground. I said above, there is overlap between me and the indy "way below", I may kern with golden thimbles, but the kerns remain the same.

I hope we answered your question of the way the license and the name table interact, and we could go on to other tables, but maybe Google needs its own font format?

"As I understand it, this is a large part of the idea of 'Functional Equivalence.'"

The relatively hilarious thing about the FE issue in the real world is, we're encouraged as hell by our clients' linguistic requirements, OSs' scaling abilities, browsers and css of lesser functionality, to reduce the file size by means entirely other than compression, i.e. subsetting and whole table whacking, to functional relevance. FE is a joke for web fonts from commercial founders and probably libre founders as well.