 Okay I'm gonna talk really really fast because I didn't Time everything So I've actually given a talk with this title several times this year and I always have a little housekeeping thing Does anybody speak Polish? That might come up later I'm actually just happy that there's at least one person who knows Bengali here Maybe more than one because that will definitely come up later So can you hear me? Okay? Okay, so I like I was saying I've given this a talk with this title before but it's sort of a gimmick title So I hope that's okay that content can change a lot because that's just sort of a jumping off point asking a question So just to reiterate the question one thing I hear a lot These days is didn't we didn't we do fonts already? Isn't that isn't that done and especially you get that when you get outside of the creative graphics people in the in the FOS community What I really want to talk about though is this this gap like Web fonts were a huge success and still are a huge success story for free software but that means there's still a lot left to do and when I say web fonts I'm primarily talking about Google fonts and about open font library and you know Assuming that those are successful sort of the question is what else is there that we need to work on and For context You might not know me, but if you do I've been gone for a while I spent a long time being a journalist just writing about free software and then I left that glamorous high-paying world behind To go spend a year in England and study type design and then join the glamorous high-paying world of type That also means that I spent a lot of years talking primarily to software developers And then I spent a year just surrounded by graphic designers and document making people which was interesting but at the end of that year I spent this time thinking about what I wanted to do next and so I Tried to come up with the list of these other things that we should and can be working on related to type And so that's what this talk is about If you don't believe me I can spend one minute illustrating the fact that the massive success of open web fonts doesn't solve everything for you And I'll try and rush through this, but really I think it's pretty clear especially to people who are voluntarily here Lots and lots of hits come from web fonts That's not my opinion. I looked this afternoon. It was 22 trillion more or less Which is a lot more than you get even of a really successful YouTube video It goes nuts and the projects Have put a lot of work into open font development but it is only Stuff that works on the web and that's sort of the jumping off point You know in theory a well-designed open font doesn't have to be just for the web Look at e-readers. They're very similar. E-Pub is very similar to HTML, but there's very little impact from open fonts there Okay, so the Kindle is from Amazon that makes sense that they would commission fonts to use They're sort of a prestige thing saying we care about this platform so much for making perfect fonts for it But other end of the spectrum is the Kobo, which is the most open e-reader. You can buy it's not bootlocked or anything like that And they also don't use open fonts a little further out like print on demand. You don't have a lot of impact from open fonts there If you have any control over the fonts you use you get something a generic like like Times New Roman Little further fuel the need to make it marketing things printed up like t-shirts or frisbees or whatever You don't really have open fonts having impact there That's freeware stuff of dubious quality that you see in these places but Ultimately the reason why this happens is that the web font services are tailored to doing web font things and That's what they're good at and that means if you want to be the person who promotes open fonts in some other area That's that's a separate task that you have to tackle So yeah, there's boundaries to what you do and if you want to look at a different area, then you have different questions to answer Yeah, and that's a good thing right because it means we have a lot of stuff to do and when I made this list last fall I figured I would have a little time maybe to work on one or two of the items This is the list I came up with I'm not going to go into depth on all of these things. I do want to explain what I mean by the entries Just because I'm not going to talk about all of them doesn't mean I think people shouldn't work on those because obviously you should I am Going to talk about font packaging though because that's the one that I've had a little little time to Maybe push some things around for but first Templates we've all seen presentations that look like this. This is the default template use in LibreOffice and I wouldn't want to have to stare at that for a long period of time like you have to in a presentation The other templates you get in LibreOffice Not all that much more exciting You get different colors of rectangle, but yeah, so There's really nice open fonts, but they don't make it into the document templates that we give people and the result of that is that The fonts don't make it into the documents that users create Sort of the fallback is something like deja vu which has wide Language coverage in these situations and there's a good reason for choosing that but it's sort of a safe choice And Related to that is like the documents that open projects produce on the web And there's a couple of projects like Gnome and Mozilla that have style guides For for how you what your documents should look like typographically been very very few so that's another thing I think that we could probably push forward on and Both of those kind of get into what are the default fonts that we give people and that ties into what ships with an application and describe us in LibreOffice have both discussed this other projects I know have not like Inkscape and Gimp Will very reasonably tell you these are installs the application and whatever the system fonts are or what they have and I understand that But I think it's probably fair to say that's why people create documents with Inkscape that use aerial and You know comic sans and things so maybe looking forward Especially when we look at new packaging formats to deliver something like snap and flat pack and app image Maybe we could give them better typographic choices So those are the other issues I may have skipped a couple in there I'm gonna come back around to some of them at the end I'm also not gonna get into things like shouldn't we be making more fonts because we should obviously there's lots of room to grow there, but there's a lot out there and I'm not gonna talk about variable fonts because you get to hear about that just a few minutes from now and I'm not gonna talk about the work that other people are doing on font selection because that's like in process I'm just gonna talk about the packaging thing because I had a little time to devote to this earlier this year but there's a lot of interesting stuff being done on improving font selection in the GTK world and I think that's gonna bear fruit in the next release cycle or two So packaging this is a quote from unnamed source Richard Hughes Who said I hate the fact that we we package fonts like their CLI apps? And I originally put this in my first iteration of the slides is sort of a punchline to a joke But the more I thought about it the more that I think it's a really core issue And I think it's something that we should look at Because the limits of a package have this ripple effect on other things like there's not a lot of great font managers for Linux and I think part of the reason is that the font package that we have don't make it easy and They also make work Happen for other people so I talked to an inkscape developer yesterday, and I did not ask if I could use that person's name but This person right up here in the front to have and we were talking about showing showing the user interface Choices you have for exposing open-type features and he was saying we have to parse the font binary to show you this stuff and Okay, that that makes sense But that work gets duplicated in every application that wants to show those things So I think maybe if we had a little more full features for the platform level for fonts And maybe there'd be less duplication of effort there And So Packaging I think the underlying issue with font packages is that fonts are this weird hybrid thing or their part software and their part content And there's dichotomy there and it's hard to support both of those things So yeah, they have executable instructions in them, but when I say their content I mean there are page elements whenever the user is doing something they're part of the content of the document that they make and That's a really different mindset If you look at the the font manager, this is GTK font manager, which is someone's personal project It's not like the official GNOME font manager, but let's just look at it. You can see It shows you some metadata that's pulled out of the font binary like proportional or the width and slant and weight Settings things like that It can show you a basic waterfall rendering of the alphabets And But I mean it doesn't have any any awareness of this content aspect that the font has for the for the designer so the data here is sort of context-free and I spent a lot of time last year like I said around these document people like book designers and information designers and so I got to observe the way they used fonts and the way they selected fonts when they're working on things and The the key takeaway I got from that was that be the context of a person choosing the font is Whatever the document is that they're working on and the document is more important to them than the font is the font has to Fill a role for that Which means you should think about font selection from a document first perspective, I guess and again, this is sort of the the package of the the font packages content not You know just an executable binary So if you look at other things that are sort of content content driven like music managers you get a real different experience than you see in the The font managers like this is rhythm box. This is no music in both cases Most of the screen real estate is metadata a lot of which is editable by the user. So there's Applications like easy tag. They'll let you edit the mp3 tags of a file that you have That's sort of been taboo to edit font metadata in the past because people are afraid of you removing Copyright notices and things like that that could be done in mp3 too, but people got over it So yeah, if we if we look at the way you manage content that's that's audio I think there's some things that we can learn from that Like the searching the sorting is probably more important than anything else than just showing Stuff it's also really easy to add music to your to your music library You can take your music from anywhere. It can come from online. It can come from a USB stick You link to things like Album arts that don't actually come from within the binary and I think there's maybe applications there to the way We manage fonts too And yeah, the the user gets to edit things and And so the specifics that I think you're maybe unsatisfactory about traditional font packages is they usually just contain the font and not much else and and some of the stuff that will Possibly be improvable might be what happens at install time I think there's better metadata I'm gonna come back to the install time thing definitely at the end because I think you have to maybe think about the others Before it makes any sense to care how you install something then there's like support material and documentation and then a Feedback mechanism like when you buy a proprietary retail font There's always this Contract as you've engaged in business and it's if you have problems with this call this number or send an email to this person I know that's kind of hard to do for free software in general So I don't think I've got an answer on that one, but it's it's something that needs addressing As a tangent we there's also a big consistency issue I dug into Debian font packages and a lot of like Path inconsistency and naming inconsistency and missing metadata like just look at the stuff that gets installed here This is where the Polish speaking person would really come in handy, but no tow is a family or like a mega family Unrelated scripts and things Lato or Lato That's I don't know how to pronounce that it's Polish is a more focused like regular traditional family with a couple of different styles in it But then you have low heat which has an installed to fonts That's a means in Bengali use the same writing system. So why they have two separate directories. I don't understand and then Fonts being extra that's a whole collection of fonts that are the same script by different people So it's a little bit chaotic And and the same thing happens in like the package names and that just makes it a little trickier to find things and I mean cross-core is a set that's supposed to be compatible with Microsoft fonts deja vu is a single family George Williams is a designer's name and there's a language tag anyway just it's the Wild West as far as packages go and I think maybe If we want to talk about having Linux or free operating systems to be really good environments we could get our own house in order a little bit better and Set better expectations maybe for how things are packaged and organized Particularly when you look at like font managers on Mac OS They have this method where they keep the library of inactive fonts somewhere and they just sort of copy them into the right place So keeping that library organized is pretty important and if we don't organize our own font packages That's not gonna be so easy Anyway to get into the specifics What kind of metadata do I think we could do a better job with at the package level most Linux packages these days use AppStream Which is sort of generic for any kind of software package and the font section doesn't offer a whole lot So it provides font names you can link to the license You can put a screenshot in and you can add a URL which is sort of meant to be like the home page of the project So I thought I talked to Richard Hughes and some other people and Proposed maybe some additions that would be useful for fonts and I tried to make like a minimal set of these like the designer name because a lot of times You know the name Baskerville Because that's the original designer even if it's a revival and that's the thing that you're thinking of when you're trying to choose Font and you might get familiar with a type foundry. You might know maybe that they do Scripts that work well together. So if you need Arabic and Latin in a document You want them to be designed together. So you don't have to mess with the sizes on every single line so there's things like this that come from the way people choose fonts and It wouldn't be hard to add them to AppStream if you actually visit that URL for the issue. You'll see no comments on it I Think I probably just have to push a little more there But I am hopeful that we can have a little bit richer way to express metadata In the package, there's also obviously lots of binary metadata that I'm sure plenty of our open fonts are just sort of missing There's a talk I heard just a couple weeks ago about the D-lang tag, which is the design language. It tells you what the intended first language is for a font. So again a lot of Non-Latin fonts will have a Latin alphabet because it's needed for compatibility reasons But if you don't know which it is how do you know what to show people in that little preview string? And then there's things like style information That's really hard to come up with a good answer on. There's Panos and some some features like that that are meant to be ways to categorize style and present that to the user I'm not super convinced any of them Really work across cultures and things like that, but it's certainly something we could we could work on a more Specific thing that came up just last week in a discussion with Matthias Classon who's the the GTK maintainer is what do you do to show the open type features in a font? Some of them are obvious like if it's the numbers and you have the old style numerals Or you have the tabular or non-tabular option, you know what to show there It's zero through nine, but when it's like the swash Maybe you need to say that the queue is meant to curl under a U here So you need to show a queue and a U together or else no one will understand what the feature is supposed to show you there And then the second one there is SS01 so you can have stylistic sets and those have some meaning, but you still need to show someone a good example of it and Like in this case, it's the the eng the end with a hook on it. You need to have a word that shows the difference between the two variants of that letter that you you provide in the font and Initially when I was discussing this with other people who were interested like some of the some of the Google font folks Say, yeah, this is an issue. We would like to have Solution to two and I Initially thought, you know, there's this thing called the meta table in a font where you can just sort of put whatever you want And we could just have one little tag for each feature and just show a string You can show and then somebody pointed out that you need to have different things for different languages So the ligatures that you show For Dutch you probably want to show that that ij ligature if you're using that because That's the relevant thing to show that is a feature of your font Whereas if it's in German you might need to show a different string entirely So that could be a little more complicated But I think there's there's interested maybe coming up with a solution there where you just add this to the font and it doesn't require Changing the format or the specification or anything like that Moving back a little bit to the the back of your metadata You might have noticed that reserve font name thing This is a feature of the open font license where you release a font you can declare that This part of the name is reserved and that means people can make derivative versions of the font But they have to change that part of the name if they change anything of the font now if you've made it so that The font can be rebuilt bit for bit compatible. They don't have to change it But if they're changing the font they have to change the name And that's a pretty important thing and currently RFNs aren't tracked very well, so you end up with fonts that have RFN violations accidentally this is I found several packages in Debian that are like this and it I decided that this probably needs to be supported in the License compliance stack and that's what spdx is spdx is this XML database of licenses for all kinds of free software And for like quasi free software because there's there's things that are not that aren't going to meet the FSF Definition or the open source definition, but you need the track if you're working on a project so that you don't Cross contaminate. I don't particularly like that word. So I started working on this and I think we were going to get The reserve font name thing added as an exception in the expression formula for spdx And it also just in the process of doing that decided there's some font licenses Used for free fonts that are not in spdx either So I'm working to like document all those and currently there's 19 Which is a lot more than I expected to find because there's a lot of older things that are Still distributed in Linux like fonts that were donated to the X consortium and each have a weird little license on them And there's the bit stream license the bit stream vera license and things like that But I I've done a little work on that and I think that'll come through I don't know what kind of timeframe to tell you on but the idea is that adding those things to spdx means we can have a better Appstream experience for fonts, which means we can maybe change the font packages. So there's a lot of Yak shaving chain reaction things involved in this, but I think ultimately it'll be worth it Moving on to the support material This is always the controversial one like you saw that the sample image of an alphabet in the GTK font manager example, but You can't rely on that like you get this is I think the Ubuntu software center You can get a preview of a font in almost any package manager, but it doesn't always work, right? and If you're just relying on what's shown that's automatically generated this is a table of Bengali and it doesn't show you a lot to see it not in the context of of language Because a lot of the letters have to combine together to be orthographically correct And so you need something a little more complicated than just the table And yeah, again, the automatic data lets you down there So that second item on the list is specimens and this is sort of the outside the font stuff This is the specimen booklet for Gentium, which is a an open font from SIL and what it shows you is the font in use and Gentium is a little bit academic in nature. It's designed to be used for serious stuff. So it's got a serious looking specimen and you get this as a PDF and Type foundries in the retail world like to hand these things out as promotional material and you frequently get a PDF when you buy Retail font too, but there's not really a way to link that into the package so that the user can Can look at it when they're trying to decide what font will look right in the document. They're making But I mean ultimately it's just about showing the font in the context of real real language and this is an image from my own Font specimen I had to make last year and that is Bengali where you can see letters that combine and get Repositioned as is needed to show whether or not the font is working correctly So it's not even about the look being good as does the font actually work You can't tell that just from a table, but specimens can get interesting looking. This is a web specimen. They don't have to be PDFs, but I mean between the two of those and clearly the one on the right is Gives you a better feel for what you get when you're making a document And that's the point is that someone designed this it's not automatically created There's been some work in the past some tools for making specimens. This is a short list of these I'm sure I'm missing a lot and You know I've Suggested to people like the Debian font task force. This would be something we could Hack on and I presented that at Fozdom in the open source design room And Just the list here is not gonna solve it It's gonna take a lot of work. You need people who are good designers who are willing to do this So if you're interested and have ideas about how we could tackle the big List of things that need to be shown in a nice specimen. Let me know Sort of related to this is documentation and there's a lot of overlap there, but this is bungee bungee is a color font it has multiple layers in it and Those layers are meant to be overlaid on top of each other if you find software that supports it Each individual layer there you couldn't figure out how to combine those if you didn't see them in use like that So that's a little more Important than just showing off something nice looking, but that applies to like open type features, too This is another s. I'll font that shows you like language specific things you can do if you know how to turn on and off certain features A lot of times you get this sort of material in a PDF to and I that I think we could probably do a better job of linking in and That might not be too readable, but this is just sort of showing Ligature options and there's always so many literature tags you can use and So yeah, that's It's easy to just say Include that documentation, but I mean a font package could put anything it wants in user share doc But you might as well not because no one's gonna see it I mean you might be surprised if you looked and user share doc how many quant documentation Documents already have so that leads me to like my first sort of big reflective question this week is I'd like to know what we could do better how to link like documentation into a place where the user can use it as I realize I don't use yelp a lot on GNOME, but a lot people do and I don't I don't know is it's is it something I know there's Maps like inkscape and Gimp and Scribes you can have extensions and you can have things that get hooked in pretty easily And is that the appropriate place to maybe when you install a font? Have it drop in some documentation that just in a way that you can pop up and just read okay That's that's the set I need to turn on for this feature. Is that the right place to do it or should it happen? At the OS level should be a man page I don't know I don't have the answer that so I'm hoping to discuss this up at the typography buff But there's a lot of options out there. It's just sort of a what do you find that actually works for people question? So looping back around I said it would come back to this install time assistant thing reality is that whatever we Figure out. Hey, how we can improve a distribution font package It's not gonna be the only fonts that a person has they'll come from Google fonts downloading a zip file from some random place package that's released just on github and And Yeah, we want to we want to make a good experience there So what do you do if you just have a zip file like it's easy to install something through your distribution But when you get outside of that? Not a lot happens. This is downloading something from the web Fonts dot zip so you have to know what to do with that you open it up And you put it put the font file in the right place and good luck and I mean ultimately I think the issue there is that That just makes the font available to the system and what you really want is to make it available to the person at the keyboard And maybe that means doing more like with some sort of handler Install time I think you know You can pre-fill some metadata if someone does get motivated to write a really great font management system You can maybe ask the user is this rights turn it on turn it off If you see that there's PDFs in the same folder you can say are these specimens do you want to load these in? Room to explore there certainly and at the very least I think you could definitely generate that that sample image that otherwise If you don't have it you've got nothing to look at There's also interesting work on this installation problem happening outside. This is a project by Yanoni Have you pronounce that and he just proposed this on like a type discussion forum Month ago probably and it's meant to be like a self-hosted client server thing So you can deliver fonts to people who buy them from you, but it's all it's free So it's not limited to people selling them, but it lets you like Push out an update notification so people always get the latest version We at least need to Participate in the discussions happening there I don't know where that's gonna go But I also wanted to end by saying that it's easy to say why isn't there a good font manager But I don't think that working on these packaging things is Just leading to there being good font managers. There's we have a good rich Platform for fonts other things become possible, and I just jotted these down off the top of my head, but Yeah, if it's easy to look at the metadata in a font, maybe that helps translators Maybe it helps people who are interested in learning a language And maybe it just helps people debugging fonts to be able to say this is missing in this font file So now I can do something about it. So it's not just about having that rhythm box looking looking font manager About out of time there if you have any other I can answer your questions as long as they'll let me stand up here but Also come to the type off on Monday and talk at length about whatever you'd like. So thank you Okay. Oh, yes I'm actually not sure if it's The main problem is getting more funds onto your system or getting more Information about the funds installed the main problem to me at least for my from my perception is how do I manage the funds? installed on my system, and I think there is Also problems which can't be solved by just one font manager because For example, we just recently forgive the question dropped up How can I hide fonts from this application is like kind of Application-specific block lists because I know there are a ton of useless fonts But I don't dare to uninstall them because I don't really know where they are needed and I Totally agree. I guess I didn't really want to head down that road because it's broad and has no ending inside commercial and Non-commercial font managers make a big deal out of being able to activate and deactivate the sets of fonts because You don't want to uninstall the package from your system You just don't want it polluting your screen as you scroll through a thousand things and the app Specific block list thing. I don't think I had heard or thought of before so that's interesting But yeah, I agree that managing information overload is a big issue, too I Comment in case it didn't get on camera was Related issue is automatically activated a font if it's in a document that you open That shouldn't be impossible like this Yes For the talk For me, I'm teaching students and for them. It's They they do websites, so they are interaction designer for them. It's even like super hard to figure out What is supported in the font because in the CSS they have to type in four letters But they have no idea so it's trying their or like searching for documentation or open up in a font editor So it's super messy. So I really like Your Drop the font onto the web page things that help out there, but there's not really a good full-time solution Okay. All right. Thank you