Why are tech companies making custom typefaces?
arun.is
arun.is
I have to imagine a lot of these commenters would say the same in regards to any sort of subjective artistic choice that isn't purely optimized for efficiency.
I would like to leave you with a quote by Briano Eno from a really good lecture he gave several years ago that I hope can provide a jumping off point for alternate ways to think about style, and why I think y'all are asking the wrong questions.
"So the first question is, why is any of that important? Why do we do it? And notice it’s not only us relatively wealthy people, in terms of global wealth, who are doing it - it’s everybody that we know of. Every human group we know of is spending a lot of their time – in fact almost all of their surplus time and energy – is spent in the act of stylising things and enjoying other people’s stylisations of things. So my question is, what is it for? In fact, my friend Danny Hillus, who’s a scientist, was asked by a well-known science website, along with about 300 other scientists, he was asked what is the most interesting scientific question at the moment? A lot of the other people replied with things about the cosmological constant and Ryman’s Hypothesis and all these very complicated things. And his question was very simple: he said why do we like music? And if you start thinking about that, that is really one of the most mysterious things you can imagine. Why do we even have an interest in music? Why do we have preferences? Why do we like this song better than that one? Why do we like this Beethoven sonata better than that Beethoven sonata? Why do we like this performance of that same sonata better than that other performance? We had very fine distinctions about things that we prefer, aesthetic things. And, yet, none of it seems to have much to do with functionalism, with staying alive and certainly not with industries I would say." - Brian Eno
https://www.bbc.co.uk/mediacentre/speeches/2015/bbc-music-jo...
Unfortunately, "subjective" is often used as a dirty word especially around here where some folks even claim to prefer html stripped of all styling. As you pointed out, there is a lot of thought and sweat that goes into the creation of a typeface. The artistic aspects deserve respect and attention not just from the practitioners but also from their users.
I think it's actually a good thing that there's a proliferation of typefaces. If a large company, wanting to avoid millions of dollars in fees to license a typeface, can pay a designer to create a new one that's a good thing. Even if the typeface is created for reasons of brand vanity, it puts someone to work and helps to further the development of typography in general.
As a developer and sysadmin: Let me know how that works for you. End users will rarely if ever recognize your contributions; will only really notice your contribution when it doesn't work.
Sorry.
However let me offer a counter anecdote: I love the BSD man pages. Do I think BSD is objectively better than other systems? That’s debatable. But I sure do appreciate the meticulous care with which the BSD manual is maintained. I don’t remember names off the top of my head but I always read the authors section of a BSD man page.
If you’re really seeking end user appreciation I’d suggest you try your hand in a relatively more creative context, like writing a user interface. The amount that people care matters enough that a program with fewer features, or that is less stable, or inferior on an “objective” technical level, but with a snazzier, more refined, UI will often be preferred to a technically superior one by users.
> Sorry.
Typefaces, like other graphic elements, might not be felt to the point of the reader explicitly "recognizing" these contributions. Their impact is real albeit subtle. Designers are aware of that but still work really hard anyway.
It might very well be subjective and difficult to put into words, but even if one is a strict positivist, that's acceptable.
If you have more resources and need to ask questions A/B testing doesn't handle well, you can also do it with good old-fashioned surveying. Take (say) a few hundred people sampled as representatively as possible from your target market. Randomly show half of them a version of the design that uses one typeface, and the other half a version that uses another. Ask them to rate how well they liked the design. See if there's a difference. Or ask them to recall things they saw, and see if there's a difference.
I received immediate, positive feedback on my initial contributions. Applying basic design to otherwise "undesigned" interfaces, made a big difference to our users.
While that may be an extreme example, it does highlight the need for design. Whether or not someone can "respect" it, as the OP suggests, is debatable. But, the effectiveness of something designed well, isn't.
IMHO, designers just don’t know when to stop.
As a developer, and maybe more relevant here, a user of software, it's kind of surprising how development efforts often continue well past the point of shipping something useful. Now, I'm not arguing that this applies to every software product, or that people should stop at the bare minimum like "welp, we shipped v1, let's pack it up everyone."
But this underlying belief that things must be continually improved is very pervasive in software and even moreso in open source. ("Last commit was 6 months ago? This project is obviously dead!")
I don't know that anything can or should be done to "fix" this, but it's an interesting observation. Think about it next time there's an "upgrade" that breaks something or changes a workflow you liked. Why did that happen?
Conversely, designers can not in isolation build a new product and thus are stuck (again, in my opinion) reinventing the wheel for what exists.
And yes, I agree with you that there is a lot of change for its own sake and I wish it would stop regardless of the source. See the recent UI overhaul of gmail. It was enough to force me to use their basic HTML UI and now I’m looking for an alternative all together.
When someone complains about how superfluous "design" is, I always assume they mean the latter definition. But I could be wrong.
Because so many other things designers come up with are less readable and less usable in general, even as they protest that their designs have manifold improvements over the previous status quo and that we're just too philistine to see them.
I appreciate good styling, I merely disagree that gray text on a gray background with 80% of the screen eaten up by margins qualifies as such.
I don't think his/her frustration was with the article, rather with some of the HN comments on the article.
"The scientist does not study nature because it is useful; he studies it because he delights in it, and he delights in it because it is beautiful."
I definitely see beauty in music, science, typefaces and many other aspects of the universe.
See this tweet and the responses: https://twitter.com/kodform/status/996447044100386816?s=19
Digital fonts changed the economics of typeface production. There's a glut of independent typeface designers now, creating fonts that I suppose count as valid self-expression. Far more typefaces are being created than are actually needed. It's as if creating a typeface is like performing music; it was one thing for many performers each to play live to a small audience, it's something totally different when all their recordings are available in the same marketplace. The best performances will obviously get a large share of the market, and recordings that take a lot of resources to produce (like a corporate typeface) will exist in a different category to the homemade ones.
As I see it, most of these corporate typefaces lack authenticity. They're typically not a meaningful part of corporate design strategy. It is a cosmetic activity. The exception is those which cover larger that unusual chunks of Unicode.
*And often functionally inferior to the standard fonts they're aesthetically derivative of, because they were hinted poorly to save development costs.
Except that quote has nothing to do with business decision making around aesthetics. It's about arts & entertainment and an existential pondering about human thirst for creativity. For profit software (which is the topic we're talking about here) is about providing a solution to a real world problem.
That being said there is some research around aesthetics from a usability perspective: https://www.nngroup.com/articles/aesthetic-usability-effect/
Summary: Users are more tolerant of minor usability issues when they find an interface visually appealing. This aesthetic-usability effect can mask UI problems and can prevent issue discovery during usability testing. Identify instances of the aesthetic-usability effect in your user research by watching what your users do, as well as listening to what they say.Why we like music is pretty well studied.
https://www.nationalgeographic.com/science/phenomena/2013/04...
Also, https://en.wikipedia.org/wiki/Sexual_selection_in_humans#Geo...
That's exactly what the quote says. You're being very uncharitable with your interpretation.
If you disagree, can you articulate the point in your own words so that others can see the subtlety it's missing?
The point of clear communication is so that someone can understand your argument without having to squint and struggle.
Edit: To which I would add, the point of quoting someone else's thoughts is because they ostensibly communicate it better than you yourself can. I think this quote fails that.
The Wall and The Downward Spiral have similar elements of challenge, now that I think of it.
But if they do just want to avoid licensing..and? If it's less expensive to go through the design and construction and acceptance and branding, then the typeface is probably a tad overpriced. A grotesque sans-serif is a grotesque sans-serif, not a rip off of Helvetica. It doesn't own that entire concept.
While I find this piece interesting and informative, it seems biased in what it considers a useful outcome and what it doesn't. YouTube cutting off "random" pieces, for instance, yields a playful and unique typeface.
Besides, I don't think it's open source.
Though the sans is humanist and not neo-grotesk like Helvetica.
Hell, Noto Sans is available in more weights than Helvetica Neue LT Pro. I've actually found myself wishing that Helvetica Neue LT Pro had a real demi instead of jumping from medium to bold. Noto solves that problem.
Unless you use Chrome, you can also go to the reader mode to avoid all the CSS. Plenty of options for you!
Glyph wise, I would love to have both tabular and proportional numbers. I also want great hinting so it looks good on Windows devices. I also prefer something that isn't strongly tied to another brand like Roboto is to Android.
Once you take those into account, there isn't actually a ton of options left. And if you want something that compares to the quality you get Hoefler & Co, you end up with close to zero.
"When purchasing your digital ad license, you specify the number of impressions your campaign will require. If you’re uncertain how many impressions you will need, you can purchase a small allotment and true up at the end of the month."
- They track use by embedding a link to the third party font provider, which does a 302 redirect to a file on your server that requires revalidation every load. So while you host the actual file delivered to the client, you can't optimize away the time added by that needless redirect on every single pageload[1].
- On some clients, invalidating that original link will also invalidate the cache for the destination of the 302 redirect, so the user is forced to truly redownload the font each pageload.
- Automated bot traffic, whether benign, malicious, search engines, or anything at all, becomes very expensive. A flood of scrapers hitting your site shifts from a minor nuisance to a very big deal. Something that your infrastructure can absorb without a hiccup can easily rack up hundreds of thousands of dollars worth of font usage costs. You can put security measures in place to prevent some of this, but at the risk of adding friction to legitimate traffic. And don't wan't to stop legitimate bots like social media crawlers and search engines. So you can either leave the risk open ended or "fail open" where, rather than blocking a visit, you selectively decide when to include the custom font and when not to. Cloudflare's SSE[2] comes in handy for this, combined with dynamic inclusion in the server side code to exclude it from being served to legitimate bots like Googlebot which don't need the custom font. I've also experimented recently with using Cloudflare Workers for handling that process and abstracting it completely from the original site. While Cloudflare Workers also have variable pricing based on usage, it's over an order of magnitude cheaper than the usage-based pricing of the font itself, and well worth the cost.
- Maintaining that added complexity, purely to cap the risk of the pay-per-view pricing model of the font, adds a needless amount of friction to the website management and development process.
[1] https://www.smugmug.com/ is a site that shows this workflow. They load a font file from cloud.typography.com which 302 redirects to the real font file hosted on Smugmug's CDN. While the response is usually fairly quick, I've seen inconsistent response times from the third party providers. And the fact that in some scenarios the must-revalidate flag from the 302 response causes the cache setting from the destination font file to be ignored as well, making it get downloaded every single pageview.
[2] https://support.cloudflare.com/hc/en-us/articles/200170036-W...
GDPR, here we come!
> Automated bot traffic, whether benign, malicious, search engines, or anything at all, becomes very expensive. A flood of scrapers hitting your site shifts from a minor nuisance to a very big deal.
As they used to say on the chans: "Oh, exploitable!"
This would be a wonderful way to financially DDoS some organization you just don't like very much.
I wish I had that card to play! For that specific provider in the example, their ToS explicitly says that, other than your login credentials, they're allowed to do and distribute any and everything they want with any data they collect from the usage of their service[1]. I'd be really curious to see exactly what usage is hiding behind that clause.
> This would be a wonderful way to financially DDoS some organization you just don't like very much.
Right?! It was mind boggling to learn how easily abused of a system it is. Their ToS even states that it's your problem, not theirs, and any call whatsoever to the "CSS key" (the nonexistent font file on their server that redirects to the real file on your server) is and will be charged as a view. You don't even need to hit the site it's embedded in - just hook up to TOR tunnel and throw a curl command to the "CSS Key" URL into an endless loop. And for optional flourish, add fake user-agent, origin, and referrer headers to the curl command. Let it run for a couple hours every week. Since you didn't hit their site, their site analytics never triggered, and they won't know what hit 'em until months later when it reaches a level that warrants finance asking why you blew so far past your budget.
I managed the digital analytics for that particular client, and an overly aggressive WAF[1] was in place to block scraping attempts because they'd been burned by font fees in the past. The WAF's browser challenge was poorly designed and as a side effect it destroyed all referral (and therefore channel attribution) data.
I had to work my way through a lot of layers within a Fortune 500 company, from my marketing client to their global infrastructure team, to get buy in to test out (and eventually switch) to Cloudflare as a much more robust solution to their font concerns[2]. Which also just so happened to correct the referral/channel attribution issue I was seeing[3]. Along the way I picked up a pretty comprehensive understanding of licensed fonts, as I had to use completely different approaches for each department and set of red tape I needed to cut through. Marketing, media buying, procurement, IT all have different ways you have to frame the risks and opportunities, even if they're all the same thing said different ways.
[1] https://en.wikipedia.org/wiki/Web_application_firewall
[2] The font issue was irrelevant to me. But my referral data was irrelevant to IT, and they put the WAF causing my problems in place due to the font issue. So I able to get procurement and IT's blessing by enticing them with the additional cost savings and protection from the Cloudflare-based solution.
[3] I was only able to get to procurement and IT in the first place by having the backing of marketing and media buying, who cared not at all about fonts but greatly about website performance, conversion attribution, and being able to analyze visitors by traffic source.
The fonts have some differences, but I'd have to agree with you the changes are so subtle it's hard to believe they resulted from an attempt to make a _radically_ new typeface.
[1] https://www.arun.is/blog/custom-typefaces/unique_typefaces_m...
Also many of these look like they just adjusted the weight ever so slightly. With variable fonts, I'm curious how this come into play when considering copyrights.
https://www.bitlaw.com/source/37cfr/202_1.html
Note that although the design of the typeface is not copyrightable, the computer programs which generate typefaces (‘fonts’) contain elements which are:
When discussing copyright "font" doesn't refer to the shapes of the letters (which comprise a "typeface") but the programmatic implementation of that typeface, which has been held to involve independent creativity. So if you take an existing vector font file, and alter the control points, you've created a derived work.
The situation is different in many other countries, which do allow for a copyright in a typeface. Generally the term of copyright is much less for a typeface than for many other works.
For example, I never really liked when Apple switched from Lucida Grande as the main UI Font to Helvetica, and I was very relieved when they replaced Helvetica with San Francisco.
I think San Francisco is a decent UI font -- it's much better than Helvetica. I don't particularly like it, but at least it doesn't annoy me.
The differences are subtle. If you show me short strings I probably wouldn't know them apart, but changing the font for the whole UI has a big effect.
https://en.wikipedia.org/wiki/Intellectual_property_protecti...
There is an interesting history of font copying going back to the beginning of the USA. Where the US printers wanted to use European fonts (typefaces?), but didn't want to be hampered by their copyright laws. So there was no copyright protections of fonts in the US.
IANAL but Adobe lost some court case on this recently where they tried to accuse someone of "stealing/copying" a font. But since you can't copyright the alphabet, you can only copyright the font's actual data.
In theory you could trace any font you want in a font design program, and it is now legally yours.
Which makes sense if you are a business that uses fonts (ie a software company) as a primary part of your product instead of a permanent fee to a font provider.
Because they hate me. I disable font smoothing on my machines. Custom typefaces usually have poor hinting, which means they look terrible without smoothing (e.g. inconstant line thickness and inconsistent spacing).
The fonts that ship with Windows and Mac OS are wonderfully hinted, and appear crisp and clean without font smoothing. I curse the day web fonts were invented and gave the industry bad excuses not to use them.
Google is the absolute worst here. They use custom fonts heavily now, and somehow they're not even blockable with the Font Blocker extension (https://chrome.google.com/webstore/detail/font-blocker/knpga...).
No, I have some nice, large 16:10 monitors.
However, that brings up a good point: if you're designing your stuff to only look good on a MacBook's retina display, you're doing it wrong.
I recently upgraded from a 26" 16:10 1920×1200 display to a similarly-sized 16:9 4K display. I ran them side-by-side for a few days to compare. The difference in text quality was night and day. After using the 4K for a week the quality of the old display looked terrible: blurry and unreadable. Like going back to 800×600! I sold it; I saw no value in it as a secondary display. If I do ever need a second, it will be another 4K. Or 8K when it's available and affordable. If you're developing, writing or reading all day long, the improved legibility makes such a difference for ease of reading, eyestrain and headaches that it makes little sense to retain the old.
We clearly need to support low DPI displays for some time to come; the installed base is huge. However, high DPI displays are the future, and it's going to be increasingly a requirement that they are also properly catered for.
I'm not a particularly big fan of the trend by Microsoft (for example) to use very thin fonts with Windows 10. Just because you can to show off the technology, doesn't mean you should. I'd prefer bolder, more easily legible text even with a hi-DPI display, even if the thin ones are fashionable for some reason.
On the other hand, it's sufficiently big that I don't generally need such a large space. I could have gone with a physically smaller 4K display with a higher DPI and still been very happy.
Tilting them on their side doesn't help, because they're far too narrow that way.
The right font matters a lot to me. I like thin sans serif...always looking for a font that resembles what I saw on an old Apollo workstation (if I recall correctly, and I probably don’t).
Source Code Pro seems to be the best I’ve found, but I can’t get it to look right in putty.
I blame Microsoft. For far too long Windows programs haven't been able to scale gracefully, making odd resolutions painful to work with.
Why?
I also personally like the look of well-hinted screen fonts.
I've tried that, but unfortunately it breaks websites that use custom fonts for interface icons.
Similarly, you can tell the era of computer software by what strange thing they decide they need to customize. We're in the era of "slight variants of Helvetica". (Nobody's asking for a Zapfino clone, strangely.) For whatever reason, fonts and logotypes are the axis on which everybody seems to need to compete today. And gratuitous 2D animation.
We already went through the "custom sound" phase, and it (mostly) went out of style, thankfully. Splash screens are also on the way out, since we no longer need a way to hide long loading times.
I wonder what's next. Gratuitous 3D animation? Custom smells? Plaids?
To be clear, I have no problem with nice fonts on my computer, but I think it's funny that everyone is so focused on showing off their creativity on this one very specific axis. You really can't think of any other possible ways to have functional style?
If, say, physics-engine-based layouts became popular for general user interfaces, it would take over the entire development of the product. It would affect every part of the budget and schedule and architecture. You couldn't feasibly A/B test it.
[1] https://www.bbc.co.uk/gel/articles/introducing-bbc-reith
It's also an opportunity to brand something as uniquely theirs, unlike a Google web font that anybody can use.
We do have magazines and merchandise for some franchises, but as far as I know, those were secondary when it came to BBC Reith/Reith Sans. TV & Online are big, and especially when it comes to legibility across resolutions and encoding systems (PAL/NTSC/SD/Pixel Ratios/HD/4K/Mobile/High-DPI..)
I know that the current licensing is as open as it can be, but if Google or the actual legal owner changes its mind, will the current users need to go through a clearing of everything they have that uses the font ?
I would guess there’s no guarantee, and wouldn’t take that risk if I was in the BBC.
I love designer talk.
I think designers are on to something when they say to use a custom font to promote the brand. It's something that's both low-key and high-impact.
We spent a lot of time ensuring our build came in as small as possible, in order to reduce server costs and page load times, for millions of potential users.
Suddenly, the creative side of the business wanted to add our own font to the application. The bespoke font files (woff2&1, thanks lord) we were given were almost 5x the size of the initial build package.
All the worth of optimisation, right?
>custom typeface simply because that’s what everyone else is
>doing. This cargo cult mentality that is so prevalent in
>design is at best wasteful and at worst illegal.
Illegal?!? I'm going to need to see some work on this part. I couldn't find anything in the article that seemed to directly support the assertion.
on edit: formatting
Why? Because a long time ago, the US government was worried that somebody would exploit a loophole and try to copyright the alphabet. So they set up protections to absolutely make sure nobody can do this.
Interestingly, this means that text-only logos like the Coca-Cola logo can't be copyrighted; they're only protected by trademark.
If you are using a typeface to print a metric ton of publications, I'm pretty sure you would want it to be engineered to be fit for that purpose. And that the typeface definitions retains its' quality. I would not trust an open source process to maintain a standard of quality. Hence, you need an institution to maintain the quality. Which needs funding.
What is the "correct" price for the font foundries services is of course a different matter altogether.
On what basis?
It seems to me that a font, suitable for printing, is a niche product with a very old history of proprietary providers and craftsmen.
It's this crafts aspect which makes the open sourcing a bit more difficult, IMO. I.e you need specialists to maintain a certain quality.
Mickey Mouse is not a very good analog to a typeface. The analog to Mickey Mouse would be a letter in an alphabet - an abstract concept, that, never the less, is immediately recognizable.
A typeface, then, is a specific set of instructions to a specific implementation of this abstract idea. I.e. with Mickey it might be a specification to a specific implementation of this character including the type of voice actor to use, how the character behaves specifically, how it moves, etc etc. I.e. a much more specific than just "It's mickey, it's ours, fuck off" type of obstruction of creativity.
Like I said, I have no idea if foundries and their license fees are necessary, but at least their clients know how to source a specific high quality implementation of their typefaces.
Question is whether or not anyone cares.
Not sure what classes as major jurisdictions for you but I thought that lack of protection for typefaces was a peculiarity of USA-ian copyright?
Certainly it appears in the UK that there is typeface protection [CDPA S.55 limits it to <26 years though] and that you can't trace to work around it?
Re: UK
- Helvetica (to choose one) is over 50 years old
- It is not copyright infringement in the UK to use a font
Why? Because a pirate can theoretically extract the TTF/OTF from a PDF, so the licensor needs to be paid extra money to compensate for the piracy that PDF usage will bring.
This is such a bullshit 20th-century excuse. It's 2018. Any pirate worth their salt can find any foundry's complete collection on a number of torrent sites. Or for that matter, they can just Google the name of any font and one of the top Google results will be a website illegally offering the font as a free direct download. They don't need to extract a file from a PDF to get their hands on it. There are more efficient ways.
I was doing some work with Nokia around the time they released Nokia Pure, and I can certainly say that multiple language support was, at least for us, one of the primary drivers for using it, and it made our lives a lot easier at the time.
That doesn't really hold up. The cost of developing a custom typeface can easily run from hundreds of thousands into millions of dollars.
It can take a huge amount of time and energy from within the company to commission and direct a large creative project.
Typeface licensing costs are a very small proportion of what most companies will spend on brand marketing.
And saving money on a small line-item isn't exactly a recipe for getting promoted.
Announcing that your shiny new typeface will save the company lots of money is, more likely, a post-purchase rationalisation that helps make everyone feel good about the investment.
If the driver was primarily to reduce licensing costs, there are some great alternatives:
1. Use one of the many terrific free and open source typefaces that will cover all the languages and use cases you really need (like Noto or Open Sans). Or even better, system fonts.
Jeremiah Shoaf has a terrific curated list here: https://www.typewolf.com/google-fonts
2. Threaten to use a free alternative and then negotiate a better deal on licensing
3. Go ahead and use your favourite typeface without declaring the full usage so it is unlicensed or under-licensed. This saves a lot of money and is actually pretty common.
In the past I've interviewed designers and creative teams about how they choose and decide to license typefaces. They talk about things like:
* wanting to create exactly the right aesthetic for their brand (i.e. I can only be satisfied by something that doesn't yet exist)
* finding usage-based licensing complex as it creates non-financial costs in terms of understanding, tracking and justifying the licensing costs. That all gets much easier if you own the typeface.
* reducing the risk of inadvertent copyright infringement and subsequent reputational damage
* feeling in control by owning the IP (and therefore not dependent on any third party in future)
No-one's arguing the up front investment in creating the font, but instead focusing on the "rent-seeking behaviour" that sees the recurring costs come in anywhere up to $1M a year. Some foundries charge you per pageview and per mobile app developer. There is a company/audience size at which keeping track of pageviews and seat count and paying hundreds of thousands of dollars or perhaps millions per year is significantly more trouble than the initial hurt of commissioning a typeface. At the end of the exercise you can wrap it up as a marketing or branding opportunity and generate a few hundred thousand pageviews and call it a day.
An ultra-high-quality, visually attractive font family that includes a wide array of weights and even a decent selection of widths on top of what is quite possibly the best Unicode support in the world.
I'm at the point where I'm not going to look a gift horse in the mouth and just be thankful I have Noto for my typesetting projects.
I wasn't aware that these common fonts charged royalties...
I was really sad when Google switched Android's default font from Droid Sans, a humanist sans (and the direct predecessor of Noto Sans), to Roboto, a neo-grotesk.
Or when Apple went from Myriad to Helevetica Neue and San Francisco.
I just really hate everything about the way neo-grotesk fonts look, and mobile UIs have looked awful to me ever since.
What am I missing?
That said, open source fonts are definitely on the rise. Many people don't have a problem using a font that isn't quite as refined as the classics, and some enlightened companies are bankrolling development.
It is absolutely perfect. And you know what? I don't think I've gotten a single complaint from anyone about using a serif font for dialogue, which is something that's really uncommon in the fansubbing community (there is a tendency to stick to old taboos, even though the reasons why serif fonts were avoided are no longer valid now that almost everyone, including me, only releases softsubs). It's just that gorgeous.
The widths and weights Noto is available in is also massively useful to me. There have been times I've tried to set some text, observed that the regular is too light and the bold is too heavy... and then I remember that Noto has a demi. Or when it started to irk me that the width of the font didn't harmonize well with the width of the Japanese text and I started to play with \fscx until I realized that I could just use Noto Sans SemiCondensed and get something that's optically sound.
And the best part is that it's all open-source.
When you select Noto Sans in Word, you end up with Arial. If you change the font color you can get the correct letter shapes but the spacing is still set for Arial, which looks incredibly bad.
I don't know how big it was when it made the font though.
I work closely with design and I am consistently saddened by the state of the tech industry when design discussions pop up on HN. They almost always devolve into a cesspool of blanket criticism and rejection. "Those designers..." an HN commenter will inevitably say, "...all they care about is making things look pretty." Well, sure. Yes.
The facts are plain, and actually boring. Customers like pretty products. Well-designed experiences and objects, compared to purely utilitarian products of similar functional capabilities, will always sell better. They create more passionate users and make more money. This is understood by businesses across the world, almost forever. To read an HN thread on design with no previous convictions is to see a world where designers are scam artists and their contributions to the tech industry nothing but snake oil.
Of course there is excess. That's also obvious. Tarsnap and Pinboard are HN darlings and they clearly place all their emphasis on functional design rather than visual design, and are very successful businesses. Craigslist as well is one of these types of businesses. But these companies do not appeal to average consumers. Pinboard doesn't compete with Pinterest for users – nor could it ever, nor does it wish to. Would I, as a proponent for modern design sensibilities, argue that Craigslist doesn't have good design? No! They have fantastic design... for them. It suits their brand. Would Netflix be functional if it had a "Web 1.0" Craigslist-style interface? Yes! But people would hate it. It wouldn't suit Netflix. It would look terrible in a living room on a TV, for one, but Netflix is a platform for cinematic experiences and the branding and interface should allow designers the kind of control that let's them create a suitably cinematic interface – however they wish to do that. Craigslist works, by comparison, because it is essentially a glorified newsletter (hot take!).
The price of a custom typeface, depending on who you commission to make it, can probably range between $100,000 and $500,000 or something like that. For a mature business like Airbnb or Netflix or Google this is not an investment in efficiency or in growth. This probably makes the HN reader's head explode. It is an exercise in fashion and style and distinction. There is a reason that the console-dwelling set is often stereotyped as the guy in poorly-fitting jeans and neon sneakers and a hoodie with a shark high-fiving a sasquatch on the front... they often don't waste time on considering things like fashion or style or distinction.
One mentor told me that "Design is anything that makes people more successful". Sometimes this is making one button red and one button green, for example. Sometimes it is about aligning critical actions to the edges of the experience, a la Fitts' Law. But I think this is a narrow understanding of what design can be. Design is also about creating a visual atmosphere, a well-defined world for the narrative of a product to become more than it's functions.
Software is allowed to have an emotional quality to it, it is good to delight users with little illustrations and round buttons. An interface should have feeling, and mood, and a vibe to it. Personality and soul is always, at some level, designed – visually and functionally in equally significant parts.
So you go on a rant about 'us' stereotyping designers..by stereotyping us?
Bold.
In the end, as many are saying, it’s probably about licensing fees.
Just don’t make a major faux pas. Don’t wear a paisley tie, don’t use comic sans nor lobster unless you have a good reason.
Good typography can make a massive difference to the presentation of text. To some people that makes it worth the effort