Removing Font Setting


#1

Currently, in Account Settings > Display there is a setting to control which font the site uses.

However, offering a variety of fonts:

  1. Has a negative impact on our first page load time
  2. Creates issues as some fonts don’t work well with all the languages we support

Because of this, we will be removing the option to choose what font to use as part of our v4.0 release (July 16, 2017).

If you have feedback or concerns, please comment and let us know :slight_smile:


#2

#3

If you disable “Anti Aliasing” for font rendering on Ubuntu (like one of my colleagues has) the default font is not readable any more. Please add at least one font that does not require anti-aliasing (e.g. ubuntu or arial font).


#4

Hi @soccertrash - would you be able to open a feature request for adding a font that doesn’t require anti-aliasing?


#5

On Windows my helvetica and siblings fonts don’t render properly, so I was changing the font to Lato so everything was more readable.

For font changing, I understand the faster load times being a good reason, but shouldn’t you just make it so the single font that loads is one defined by the theme and then let themes change it as needed?


#6

Not a fan of this change. Open Sans is not a nice looking font. Roboto is way better.


#7

Not a fan of the change either, a refresh seems affordable when switching fonts. A simple Stylish theme appears to be enough to fix it if you are not using Electron, edit as necessary:

@import url('https://fonts.googleapis.com/css?family=Exo+2');
* {
    font-family: 'Exo 2';
}

#8

I definitely prefer the previous ability to change fonts, please bring it back!

I never really noticed the first page load time (in fact, I tried side by side on two mattermost server instances before and after the change and the difference seems pretty small to me). Being able to customize the font in order to improve readability on different platforms/settings I think outweighs the negative of some fonts not being compatible with some locales— in those cases the user still has the option to switch to a different font without those issues.


#9

I strongly agree with the points quoted. But, rather than removing the options to use custom font, giving dropdown options to choose open source fonts would have been great. (I mean tested fonts which are CDN backed and does not break)


#11

I think its important from a org point of view to be able to choose our own font and avoid using a generic look and feel. Font color by itself is not enough to make the instance feel ‘ours’.


#12

My default font is quite illegible… would be great to customize fonts again.


#13

Where do you put this change?


#14

Offering custom fonts shouldn’t greatly increase load time - only the chosen font need be loaded. And in general the negative UX outcomes of locking down the font would, imo, massively outweigh the benefits of the freedom to choose fonts.

Speaking to point 1, I think that if viewed as a technical restriction (users must be able to choose their own fonts), it won’t be hard, certainly not impossible, to solve the load time problem without restricting to a single font.

Speaking to point 2, It’s not hard to find fonts that support many languages. I would support a UX change that encouraged people to choose those fonts, rather than forcing them.


#15

Dear mattermost team,

Mattermost is about human interactions, in channels on the web. And somehow, type has become the way to add humanity to those computer tool. Look at Comic sans like the biggest proof of that. Designer keep explaining that this is a bad font, but when it was on the screen, it was close to some human writing. This feeling is only made thanks to a way to design letters as more than just a reading system. This is where, in my opinion, typography matters, specifically in discussion applications. On a branding strategy side: going back to helvetica/arial/san-franciso wont let mattermost stand out in the crowd, as type will just be like anywhere else. Unless you ask a designer to design a specific font for you, with branding in mind (i know a lot of type designer who loves to do that, if needed), but that not something that would be more helpful than keeping the font possibility changes for the following reasons.

Legibility and readability is the result of the habits of the user: scientist can read long lines of text, while i’m more used to a classic 55-65 character per line. Same thing go for type: some people have a simpler reading experience with a sans-serif, other with a monospace (hi developers!). So, removing this option is the same as creating a unique user experience for such different peoples.

As a designer, i know how the feeling of speed is important in any applications, specially since i’m not always near a huge 4G network when i’m on the move or when my internet is down because my neighbour is having a binge netflix moment (which happen quite a lot these days, but who does not like Seinfeld? :D)
But i’m curious about the analytics you have about connections and terminal: how many user aren’t running mattermost full time in the background for example. Or how many connexions are made from low bandwith connexions? If those numbers show that the low bandwith is not happening that much, do we really need to win those 5 seconds before the app open?

And the most important: the awesome tool you keep improving is my daily connection with my all-around-the-world team. We talk about work, share videos and sounds, link to interesting discoveries, and so much other things that i need to say that i would not be able to live without mattermost since i started using it.

And i think that this is important to me: mattermost is my office.

It’s made of all those discussions and smiles and all those human things that mails kept away from me before. My office and my coffee spot where all the human interactions happen.And i want to be able to design my office: i would love to change the painting, add frames on the wall and the only way to have a glimpse of that in mattermost is to change the font. I have two instances of mattermost, each has its own font choicem which i regularly change, depending on the screen i’m readin on, or the OS i use, since rendering is different on windows Clear type technology or on OSX or on linux.

So here are a few ideas about how we could keep this option while adressing the speed issue:

  • on unicode. foreign language and special characters: my team speak many languages, which cover an important range of unicode characters, and i think that we could even put kanji in here for fun. So i’m well aware that it’s not that easy to have a font which contain all that languages (not even talking about latin/arab fonts). It’s impossible. But, and that the best thing about unicode: if a character is not in a font, it will be rendered by the browser with a standard one. (have a look in there https://unicode-table.com/). So, unless i missed something, foreign language may not be of an issue, or is it? If that is the case i’d be curious to see what happen, because it would mean that the font is badly programmed.

  • on legibility and screen rendering: looking at mattermost channel on different browsers, on different operating systems with different DPI on the screen, make different UX. That’s ok, i dont need the same experience everywhere, so i’d be happy to use different fonts in different locations/experience to address this DPI issue.

  • on font loading: there are a lot of different ways to implement webfonts in the browser: https://www.zachleat.com/web/comprehensive-webfonts/ . Look at the FOUT with a Class part, which can be used with localStorage for quicker access to the app on different devices. Or, there is a possibility to use mediaqueries or device or network quality detection system to use standard fonts on devices with low bandwith.

  • Variable fonts will be an option in the future, as you’ll only need 1 file to have different font-weight and font style. We’re not there yet, as the implementation in browsers is being built. But if google support it, so much, it’s because of performance first.

  • Also, the app build for the mobile ecosystem: ios and android (and maybe Libre OS or ubuntu touch later) could embbed those fonts and win time on the page load.

  • The list of font you choose may also impact how this: some fonts handle latin extended, other greek, some have specific character and let you write vietnamese. Font squirrel made a small filtering by language tool that could help as you would describe different @font-face for different unicode range, so you could reduce the bandwith as chinese font or bengali ones would only be loaded if some text ask for it.

  • The fontlist is the most complicated thing as people are all different regarding taste in type. I trust your designer to come with the best font selection, but i think that it could be a nice place to discuss it with the community and see which one are the best for the job, on a developer side: size of the font in kb, subset of character and so on. I don’t ask for a vote, it’s a designer job to select and assign fonts, but a discussion that we would all learn from.

Phew, that was a long post to write, but it’s about my way of saying how mattermost is important to my team, how we enjoy using your tool, and why this special type thing is more than just taste or painting, it’s about making this small discution room the office we want to have.

Thanks a lot for the hard work you put in Mattermost.

I cant post the link to the font squirrel tool as i’m new user and can only post two links (weird number as 1 link is enough for a spam to be annoying though :smiley: ) The Font Squirrel tool is on the front page on the right column, under the languages button.