Public Sans – A strong, neutral typeface for text or display
public-sans.digital.gov
public-sans.digital.gov
We do everything in the open, so feel free to follow our work here: https://github.com/18F/federalist
We also implement the USWDS, which you can follow here: https://github.com/18F/federalist-uswds-jekyll
Let me know if you have any questions!
In my country, all tax-funded government works are copyrighted. It's an ongoing battle to make each department release their works under a free license, and massive bodies of government-owned digitized cultural works sadly remain accessible only to those who pay for them.
Depends on the terms and conditions of the contract. Keep trying to get people to ensure they secure rights for the government by default but it's a constant struggle to get people to sweat the details :p
If all someone hears is complaints, I think they are probably more likely to burn out. I don’t have any source to back this claim, but it’s how I think it is.
To reject and fight FOIA requests takes _effort_, more so than simply complying with them. So no, I will not applaud shit. Our government needs to get past “basically good” before I praise them for frosting like this.
But feel free to embrace your feel good powwow approach. It’s done us so well so far.
Th EU has http://ec.europa.eu/ipg/services/analytics/ but I recently discovered that I cannot access any data as a citizen. Legally, I'm convinced this data should be public domain.
If this is meant for web sites, I have a question and I ask this on HN a lot. Why does the government need to provide ANY fonts? I have a real problem with websites pushing their own fonts on users. I don't think it's their place. Users should select their preferred font in their own browser. Why would a US government site feel the need to use particular fonts?
If this is for publishing in .pdf format I can understand. If it's for web, I just don't get it.
Does this negatively affect your experience? Can't you override fonts if you want to?
Edit: note, I'm not the person the question was directed at, and have no connection to the federalist project at all.
* Improper selection of default font for constrained display eg tabular
* Or the more popular "Let's create a font because we can. It's branding."
I can't recall any instance of "oh, that website with a great custom font is so appealing I'm going regularly and voraciously consume its content". The opposite of that is true though.
This is similar in effect to users liking a site more when it is faster, but attributing the improvement to any number of other things that haven't actually changed.
For millennia, of course, the written word would appear in the particular style of an individual scribe, and might take on an entirely different look when copied by another. Gutenberg’s movable type introduced the concept of a text appearing exactly the same across multiple copies, something that has persisted for centuries through lead type, phototypesetting, and into PDFs. Only with the advent of information technology in the 20th century did the idea arise of text appearing in formats other than those chosen by its publisher, advanced by technologies such as TeX and HTML.
In the latter’s case, however, the original idea of a platonically structured document to be interpreted per the user’s preferences has long since been superseded by a return to the concept of the publisher defining the presentation. Client CSS never caught on, while server CSS took over.
Enforcing the same font and size across all pages is ergonomic and does lower the cognitive overhead of recalibrating your sight-reading for every new page you visit. Just my two cents...
We've talked about 18F with Hillary Hartley and Aidan Feldman in the past here ~> https://changelog.com/podcast/230
Get in touch at editors@changelog.com if you're interested!
Isn't every business or organization concerned about various economic forces affecting their customer budgets?
Some of those are the worst.
This is a big part of why the USWDS project was created. 18F is tiny, and there's no way it could address even 1% of .gov by itself. Fortunately, there are many people all across government who want to improve UX on sites and apps. USWDS is a good starter kit that significantly reduces the cost of such projects for everyone, not just 18F.
Asking because I think this approach rocks but have a hard time selling it internally.
I haven’t noticed this strong of a branding in other cities I’ve lived, but would love to see examples of others.
Boston even maintains its code on github! [2] Their digital department also has a roadmap of what their initiatives are, and it’s communicated in what I think is such an easily digestible manner [3]
Bravo to their digital team!
[2] https://github.com/CityOfBoston
[3] https://www.boston.gov/departments/digital-team/digital-team...
https://www.gov.uk/government/organisations/government-digit...
And even if I were okay with both things, I'd rather fill out the SF-86 to get a real clearance, and work for one of the military-industrial-congressional complex "Beltway Baron" companies for higher pay, more job openings, and no 4-year term limit.
Hit me up if you're interested in getting into this space!
You can learn more about that project here, including the open source code: https://18f.gsa.gov/what-we-deliver/forest-service/ - and a few blog posts about it: https://18f.gsa.gov/tags/forest-service/
Both that Forest Service project and Federalist run on cloud.gov, which is a Platform as a Service operated by 18F.
(I work for 18F but don't officially represent it here.)
That leaves only those who have voluntarily disabled JavaScript (<1% of users), but fortunately those users are typically aware of how to resolve the issue of their own creation.
I've worked on public facing government websites (not 18F). We simply don't support this edge case, and our legal department supports our legal right to do so (in particular in relation to ADA requirements).
You mean under 1% of users that intentionally broke their browser?
Seems unreasonable to dedicate resources to that, that could be better spent on 99% of our users. Why should the 1% get special treatment? And what other parts of their browser can they disable that we need to support, perhaps no CSS? Maybe IE5? Maybe they only render XHTML? Etc.
I'd say: configured their browser to work like a browser instead of like a platform to run arbitrary code from the Internet.
Ideally we should be able to trust most of the web sites we visit. The last few years have shown us this is a bad idea, here are my two top reasons:
- security: while I'm personally less concerned with reasonable ads there are a number of problems with ad technology, like infectious ads and creepy tracking.
- a bigger problem for now IMO: poorly written web apps that makes the machine noticably slower.
Anyway, for public facing sites and apps, you may already be doing most of the necessary work for SEO purposes. Letting humans access the version that you're showing to search engine spiders shouldn't be a huge burden.
No, for the people who can't afford to upgrade to the latest technology.
Strongly disagree. Having your page completely break instead of degrading gracefully puts up a barrier to those who cannot run JavaScript (for example, users with older, weaker computers).
It is unlikely that there exists a subset of users with a modern enough computer to even hit our web servers that is under-powered to the point of not handling JavaScript. Our analytics definitely don't show this.
Does your analytics correctly register hits from clients that doesn't support Javascript (I'm thinking about survivorship bias).
Since we keep revisiting this, you might say that you're ahead of the times.
I think we should broaden the definition of accessibility to make websites aren't unneccesarily annoying or invasive for normal users either ;-)
(And yes, making web applications is part of my job.)
This is wrong, of course, because it requires that people buy expensive hardware to use the latest accessibility technology, which is not generally available. It's like demanding that people buy electric-powered wheelchairs instead of making your building accessible to normal wheelchairs.
> We simply don't support this edge case, and our legal department supports our legal right to do so (in particular in relation to ADA requirements).
And I'm sure that ignoring poor people enables you to sleep very well at night.
There are legitimate uses of javascript, but I'm not going to enable it just to look at a document.
Second, it's fine with me if some site wants to forgo my patronage, or provide "a deprecated experience", but not the government.
I certainly do expect my federal, state and local government to follow best practices and provide working websites that I can use without running JS.
I can't think of a single function of government that requires JS, thank God.
Best practices is a moving target. What made sense in 1999 doesn't make sense in 2019. The web simply requires JS, CSS, and HTML today. If you disable any one you aren't compatible.
There's no actual argument for why websites should spend significantly to support a tiny subset of users that intentionally break compatibility for ideological reasons. It is unfair to our other >99% of users who we'd have more time for.
The old arguments such as accessibility aren't correct any longer: accessibility devices specifically support JavaScript (text contrast, HTML organization/order, aria tags, video subtitles, etc remain highly important).
If you really insist on a JavaScript free world you are of course welcome to visit a government office in person, pick up, and mail back a paper form. The website is merely a convenience we offer to you.
Otherwise you'll need an IE 10 or newer browser, on an Operating System that supports TLS 1.1 (Windows Vista or newer), JavaScript, CSS, and HTML.
Sure, but it still makes sense to use JS sparingly. Running untrusted remote code in your browser is a huge nest of attack vectors. I don't think that, in 2019, running JS from the open internet willy-nilly can be described as "best practices", despite the prevalence of it. We're not there yet. If three hundred million people jump off a bridge I'm still not going to do it too.
> The web simply requires JS, CSS, and HTML today. If you disable any one you aren't compatible.
That's, like, your opinion, man.
You're trying to insist that your concept of the Internet is the concept of the Internet. It's a self-fulfilling prophecy. But it's not quite true yet, eh?
> There's no actual argument for why websites should spend significantly to support [non-JS users]
Right, they shouldn't spend more because the tech they use should provide for non-JS users out-of-the box without additional overhead. If devs have chosen NOT to use tech like that then they are at fault, not the user, eh?
> for ideological reasons.
What about for security reasons?
> It is unfair to our other >99% of users who we'd have more time for.
But the reason you have to "spend significantly" to do the right thing is that you chose to use and deploy crappy JS frameworks, not that some people refuse to run your crappy frameworks. This is classic "blame the user".
Now, this is your prerogative if you're doing your own site/app, but the government doesn't get to exclude some people from service just because they don't run JS. Speaking as a techno-elitist, that's techno-elitist BS.
> If you really insist on a JavaScript free world you are of course welcome to visit a government office in person, pick up, and mail back a paper form.
AH-whaaaa? Rather than fallback to plain HTML+CSS you're content to let the user fallback to hard copies and physically transporting their meat-puppet? To save costs? On web development? Where's the sense in that?
> The website is merely a convenience we offer to you.
Well, no. It's an INconvenience you offer me. If you're offering convenience to most people but deliberately excluding some that seems to me to go against the egalitarian spirit of our American government, no? "Unfair"?
> Otherwise you'll need an IE 10 or newer browser, on an Operating System that supports TLS 1.1 (Windows Vista or newer), JavaScript, CSS, and HTML.
I run Dillo. A government website that doesn't look decent and work right when accessed with the Dillo browser is just broken and sad in 2019.
If you do web for a living, you should also know that fallbacks and graceful degradation and server-side rendering all come with a cost. Both monetarily and in terms of complexity.
Yes, I'll Grant you that if you've designed an interactive, multipage form then it is costly to rebuild it to degrade. However, I'd argue that the form was probably unnecessary technically complex and that starting with having a degraded option in the first place isn't additionally costly (and better protects the agency from ada lawsuits).
In the very rare cases that we're talking about a true webapp (i.e. Google docs, or I'll even Grant you GIS (mapping)visualizations, even if it's possible to degrade those), then yes, decisions need to be made on what minimal technical requirements are required. E.g. a government agency offering an application that only works in chrome would be a non starter.
Yeah, iSnow made the same point, but it doesn't scan. "Boss, I can make it work w/o JS but it will cost more..." Huh? Doesn't that sound just like what an unprofessional developer would say?
The fact that so many popular JS frameworks don't do the right thing is part of the JS abuse in my opinion (same goes for accessibility.) Lazy developers wrote half-assed frameworks and other lazy developers chose to use them and then people start to believe that adding JS somehow makes it hard or expensive to do without JS when really they are just doing it wrong in the first place.
> when it's such a small minority of users.
The population of the US is just under 330M, so if, say, 0.5% can't or won't run JS to interact with taxpayer-funded government services that's about 1.5M people. Those folks (of whom I am one) should not be disenfranchised, so to speak, because the gov hired unprofessional developers. The government shouldn't do that, and they certainly shouldn't try to tell me that I'm some out-of-date digital neanderthal for caring enough about web insecurity to disable JS, eh? Lousy devs (as demonstrated by the fact that they can't provide a non-JS web experience/fallbacks at an affordable rate) are precisely the ones the JS code of whom I have no wish to run, and I certainly don't want my tax dollars going to pay them to screw me out of access to the service also paid for by my tax dollars.
> choose between spending their limited departmental resources on an extremely small minority, or the greater userbase as a whole.
Or they could use tech that works for everybody automatically for the same cost, eh?
The "modern web" is dramatically worse (slower, more compute intensive, less consistent, less secure) than the web used to be. This is mostly because of the tendency to force javascript into places where it was never necessary.
At this point, there is no reason to support <IE11 style JS, but JavaScript-less browsers are still A Thing and always will be. Search engines, Opera Mini, various accessibility thing, people with extremely low bandwidth, and weirdos who choose to disable JS are all factors.
Yes, if you're making a webapp, there's no good way to do it without JS and you can just be upfront about that. Also, people who can't use ES6+ are going to zero over time, so it's fine to just write ES6. But you should also have a no-JS fallback version of an information page (anything that's not a webapp) because that's an important usecase that won't drop to zero over time.
and for more as to why: https://gds.blog.gov.uk/2013/10/21/how-many-people-are-missi...
They're now building sites to meet WCAG 2.1 which does support the usage of JavaScript for accessible users.
And it doesn't even work with Javascript.
For example, executive order numbers have both "O" and "0" appearing in them. But I'm sure there are stronger examples in various gov't depts. where those characters can occur in sequence and create errors. Not to mention any code snippets that already appear in government web sites.
I suppose the extra ink could be a concern, but we should be moving away from paper printouts anyway. And as we do that, extinguishing a small dot in this great nation's zeros would be an important symbolic first step in combating climate change.
Because it isn’t a coding/monospaced font?
Most text/display typefaces don’t have dotted or slashed zeros.
The expectation is that if you have ID numbers etc where you’d need them, you would use a different typeface.
The US Web Design System 2.0 (of which Public Sans is a part) specifies Roboto Mono (which does have a slashed zero) for such purposes. See https://v2.designsystem.digital.gov/components/typography/
article: https://digital.gov/2019/04/08/introducing-uswds-2/
Though if you try such legal tricks you may as well use the CC0 [1], it's designed to emulate public domain as closely as possible in a way that's valid everywhere.
1: https://creativecommons.org/share-your-work/public-domain/cc...
How does this work? Can you 'dual license' with public domain?
Note that this doesn't apply to non-USG-made derivative works, though.
By developing their own font, they don't set preference for any organization, group, or business.
Is Sans Serif Fonts the department where the US government can really make a difference, or is this just a designer who happened to get a job with the US government and designed a font because they always wanted to and now they got the opportunity?
Isn't this a pretty basic "appeal to worse problems" fallacy?
The way I understand the "appeal to worse problems" fallacy is: "Y is worse than X, so why work on X if we haven't solved Y yet?"
In this case, I see X as a solved problem (perfectly appropriate Sans Serif typefaces exist) and Y as unsolved, whereas the fallacy treats both X and Y as unsolved.
I don't think it's in good faith to assume that because parent said "perfectly appropriate Sans Serif typefaces exist" that the implication is that solvedness is a binary; the implication could just has easily been that the problem is sufficiently solved so as to not matter, which is a disagreement about where to draw a line, not an abuse of logic.
There are a plethora of established sans serif typefaces available under various licenses, as seen on this page: https://en.wikipedia.org/wiki/List_of_sans_serif_typefaces
These have existed well before someone said "let's create Public Sans!". It would be nice to know what the reasoning was. The Github page only goes into the details of how Public Sans is different from Libre Franklin.
* https://github.com/uswds/uswds-for-designers#fonts
* https://designsystem.digital.gov/components/typography/
* https://github.com/adobe-fonts/source-sans-pro
* https://github.com/SorkinType/Merriweather/blob/master/READM...
According to the Public Sans's Github page, "Source Sans Pro" was the USWDS default, which is a SIL font (so no licensing issues) and is also not in the OS/browser font stack.
I agree that there are other typefaces they could use! I was mostly commenting on how a list of faces that includes Avenir, Univers, and FF Meta was probably not the best summary of the available options.
Q: Who would be prepared to do their taxes with Comic Sans if taxes were lower for those who did?
Generally just using the system font is fine for most UI interfaces and the only good reasons to change it are:
1. You have a specific branding style you need to achieve (this is becoming rare) 2. Your interface has pixel perfect user created content where the difference in font to font usage between OSes / Browsers would cause actual breaks in layouts from user to user.
I run into the later often, and so we end up picking a neutral style font like this because it's the closest thing to a generic "system font" like Roboto or San Francisco that has a friendlier license.
Personally I prefer Inter UI https://rsms.me/inter/
For print sure serifs are good but again it depends. There are many highly readable sans serifs (that have big x-height).
I would imagine tech geeks and artists are generally the people who are downloading new fonts, and those poeple are probably spending most of their time using sans serif fonts or monospace.
There are multiple needs here and it seems the only action comes from the "sexy" side of things.
* https://github.com/uswds/uswds-for-designers#fonts
* https://designsystem.digital.gov/components/typography/
* https://github.com/SorkinType/Merriweather/blob/master/READM...
A typeface expresses many things, you can imagine it as different sliders along, say, 100 different dimensions, similar to songs expressing combinations of feelings.
An off-the-shelf typeface will express things that are close to exactly what you're looking for, but almost never exactly 100%. Commissioning a font gives you exactly the visual identity you're looking for, zero compromises.
Additionally, distinctiveness/uniqueness has its own value too -- corporations commission typefaces so no other brand will share the same identity. For a large nation-state, that carries the same value.
(Although, unlike a company like Microsoft or IBM, I'm not sure if the US Government can prevent any private company from using it?)
I could totally see a local government org being put through the ringer for every bit and bob of a website. If there is already an established font choice, that's one less widget they are being charged for.
The site is hosted via Federalist (https://federalist.18f.gov/) a SaaS which runs on top of sibling Platform as a Service cloud.gov (https://cloud.gov), which is based on Cloud Foundry. (Think of cloud.gov as "government-compliant and -operated Heroku" and you're not too far off.)
One of the benefits of using cloud.gov is that it automatically brokers and renews certs from Let's Encrypt... Math is math, and there's no sense spending taxpayer money where we don't have to! Here's the page about it, including a link to the broker source at the bottom: https://cloud.gov/docs/services/cdn-route/
Both are operated by 18F (https://18f.gsa.gov)... We're all feds! If anyone is interested in joining us check out https://18f.gsa.gov/join/
Different strokes (literally). I quite dislike it. In my mind this font lacks symmetry and consistency needed to be easily readable. Especially how the curved base appears to drop below the flat base! It hurts my eyes.
For example, this is a "strong" face, but it's not "neutral".
https://eyeondesign.aiga.org/this-brutalist-inspired-stencil...
Also strong, and not neutral.
https://eyeondesign.aiga.org/use-this-brick-of-a-typeface-wh...
Very readable, and it screams "America" which makes sense given who made it.
This, on the other hand, screams America: https://www.dafont.com/american-captain.font
Still, it only has to be done once! I'm a huge fan of what the USWDS has done so far. I love the model of an internal tech resource for the government.
It does make me wonder though if the usage rights of these projects should be restricted to use by US taxpayers though since that is who is ultimately paying for this work.
I don't think that 18F and US taxpayers lose out on anything by open sourcing this. It builds a lot of good will and they are probably using systems that have been contributed to by others as well.
Things generally don't have to be open source until you get to your beta assessment, as the alpha assessment is really just checking that you have a clear plan for getting to beta and they agree with your approach, so you don't fail for not being open source at this stage. The project I'm working on currently (an app to let companies wishing to perform road works see all road works being performed across the country, hopefully preventing things like two different companies digging up and resurfacing the road in a short space of time, etc.) is about to hit beta so will be open-sourced shortly
I just wish there was a monospaced variant/equivalent so I can run it in my terminal/Emacs windows.
1lO05S and other commonly-confused characters are best clearly distinguished.
And then there is the military.
Removing ambiguity and uncertainty should be the point.
My idea is being hashed out on their GitHub, though sadly I suspect they’ll need lawyers to advise them rather than HN commenters such as I: https://github.com/uswds/public-sans/issues/30
(I know this work is not GPL, but similar idea applies)
[1] 17 USC 105: https://www.law.cornell.edu/uscode/text/17/105
In the US “public domain” means “without copyright”, while in some other countries it means “without authors rights” (authors rights are a superset of copyright). The problem here is that some authors rights (both in the US & in those countries) are “inalienable”, meaning they can't be given or taken away (although there might be exceptions listed in the law).
|Il nonsense is unacceptable, but I have to deal with it all the time.
1. Publicly-funded font(s) created
2. Use on govt websites grows
3. Publicly-funded CDN used for said govt site assets
4. Other, non-govt sites use said font(s)
5. Govt becomes asset host for web basics across thousands of non-govt sites
This may be another step in CDN lifecycle of startup --> megacorp --> commoditization --> govt service.
I don't see why my taxes should fund a public CDN, though.
In that case, it's also possible (probable, even) that some particularly-lazily-developed smaller sites would simply point to those same CDNs instead of hosting things themselves (much like how quite a few sites point toward CDN-hosted versions of jQuery and FontAwesome and such).
The only remaining step would then be for that taxpayer-funded CDN to cache any and all taxpayer-funded static assets.
Fonts without that manual hinting look like crap on Windows because Windows does not force through its own rendering/hinting/kerning overrides.
And all that is probably besides the point: if you're referring to how the fonts are rendered within a webpage, all the major browsers use their own custom rendering via one (finely tuned) set of parameters and interfaces for macOS/Linux/etc and another (rudimentary) for Windows. And most webfonts do not include any sort of hinting instructions in the payload, so even if the original typeface had manual hinting and would render well on a proper type engine normally, when served over the web even to a browser doing things right it'll still appear horrible.
In case you didn't know, Microsoft is really big on typography, and has commissioned a number of incredibly beautiful typefaces that render perfectly under Windows (and other operating systems).
Maybe I just haven't viewed enough fonts on recent Windows versions to have encountered the better examples, though; are there any fonts (to your knowledge) that do look especially good on Windows (and/or better than they do on macOS or a FreeType-using Unix-like OS)?
But after reading this http://www.antigrain.com/research/font_rasterization/index.h... you can only conclude that font rendering could be a lot better... I bet it isn't because of 'legacy'
you mean that big:
http://www.unics.uni-hannover.de/nhtcapri/quotation-marks.ht...
not even able to provide correct quotation marks!
In my mind this font lacks symmetry and consistency needed to be easily readable. It hurts my eyes.
[1] https://www.youtube.com/watch?v=soYKFWqVVzg#t=34m53s
"I disagree, I like sans [for long prose] better" comments miss the point; this is not an anecdata competition, it's about how many people prefer each one (and also it would be worth considering how strongly they hold that preference, since mildly pissing off even a slight majority of your readers/customers is better than really pissing off a smaller percentage). Even if you're skeptical of the studies Mayer looked at, Google (literata) and Amazon (bookerly) seem to be in the serif camp. Do you think Amazon made a mistake, too?