Commons:Requests for comment/Hosting of free fonts in Commons

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search

Recently I've suggested an ability to upload fonts in Wikimedia and specifically Wikimedia Commons in the Wikimedia's bug tracker, phab:T367644.

In nutshell the idea is to make Commons also a free repository of fonts, just like Open Font Library https://openfontlibrary.org but using capabilities of a Wiki to organize files, provide metadata, patrolling and so.

The uploaded fonts may also be used for:

  • To display historic text of hieroglyphs, cuneiforms or other less known scripts using special fonts.
  • To create an image of a text using a font by some wiki syntax so in order to create this page for example we won't need to upload each letter of different free fonts separately, like this category and its subcategories.
  • As webfonts for parts of pages that need specific styles, phab:T166138

This is similar to the idea of addition of 3D models to Commons, STL files that had support as a community wishlist in 2015 and happened on the following years, but this time for fonts.

The reason I'm bringing this to RfC in Commons is however this comment phab:T166138#7223790 that is uncertain whether Commons community is willing to support yet another format of files. I think we as a community should keep ourselves relevant in the web by allowing new ways of contributions and I hope you feel the same way also. Thanks −Ebrahimtalk 12:32, 18 July 2024 (UTC)[reply]

@Ebrahim: Interesting idea. What would be the potential downsides aside from it just being another file format? --Adamant1 (talk) 13:07, 18 July 2024 (UTC)[reply]
Adamant1: Thanks for the interest! Community wise, what Xover has described in phab:T166138#7223790 here, new policies, patrolling, increase of Commons scope and so and technical wise there are challenges to resolve but they are much smaller than let's say STL support of Wikimedia Commons and this RfC is about the former one. −Ebrahimtalk 13:36, 18 July 2024 (UTC)[reply]
I found this RfC description very technical but the use cases at the start of phab:T367644 are worth reading. Commander Keane (talk) 13:29, 18 July 2024 (UTC)[reply]
Commander Keane: I'm very sorry about that, somehow I had to make sure concerns of technical aspects of such support are answered when I'm asking a wider audience to make sure the discussion about community support of this wouldn't be derailed because of them, I perhaps should've used a different approach. −Ebrahimtalk 13:36, 18 July 2024 (UTC)[reply]
Commander Keane: I think I've now simplified this now. −Ebrahimtalk 06:22, 19 July 2024 (UTC)[reply]
As Commons is a free media repository, I am convinced that any free (legal and potientially useful) media should be able to find its place here. Basically, font files are only containers of glyphs that are usually vectorized, but they can hold many glyphs, which make it more efficient to store. There are many typographical styles that need to be visualized (Fraktur, Grotesque, Antiqua, Ornamental, LCD, bitmap, etc.), and we can generate a table as preview of font files. I hope we can get back to adding new file formats, because there are several that are not covered yet, such as LIDAR files (point clouds), textured or animated, rigged meshes, etc. --PantheraLeo1359531 😺 (talk) 07:29, 19 July 2024 (UTC)[reply]
User:PantheraLeo1359531: Thanks for the insightful comment and the support 🙏 I like just to add as the rise of variable fonts, fonts can now have more than simple outlines and can't be simply mapped to a finite number of vector files, moving sliders on https://fonts.google.com/noto/specimen/Noto+Sans/tester shows what I mean. −Ebrahimtalk 08:00, 19 July 2024 (UTC)[reply]
The Noto font family is a good example, and fonts are a very important aspect to typographers. I am sure there are many free fonts in the world wide web that are worth to be archived and preserved :) --PantheraLeo1359531 😺 (talk) 08:07, 19 July 2024 (UTC)[reply]
  • Honestly, this sounds like a great idea. But something I wonder is if this could make it easier to display alternative fonts on Wikipedia? For example, a title is in fraktur, would it then be possible to use the Commonswiki-hosted "Franktur font" file and create a title like "{{Font|Fraktur|EXAMPLE TITLE}}" where the "EXAMPLE TITLE" is displayed in fraktur? I work a lot with ancient Chinese seal script which also has countless of different fonts and it would be quite handy if we could illustrate this on Wikipedia without needing a specific image, and that's only if such an image is available. I'm sure that someone like user "BabelStone" can convince a couple of people to donate ancient fonts. --Donald Trung 『徵國單』 (No Fake News 💬) (WikiProject Numismatics 💴) (Articles 📚) 17:48, 19 July 2024 (UTC)[reply]
User:Donald Trung: Thanks for the support 😊 I even have a proof of concept MediaWiki extension I showcased in phab:T367644#9910748 that has a syntax similar to yours which is like [[File:NotoSansOldPersian-Regular.ttf|text=𐎡𐎢𐎣𐎤𐎥𐎦𐎧𐎨𐎩]] and that MediaWiki extension actually works in a fairly regular Linux setup but sure is far from anything that can be deployed in short term in WMF. −Ebrahimtalk 20:35, 19 July 2024 (UTC)[reply]
I believe the templateStyles extension already has support for loading fonts from commons, it would just need to be turned on. Bawolff (talk) 17:42, 21 July 2024 (UTC)[reply]
Also proposed as a Community Wishlist on m:Community Wishlist/Wishes/Hosting of free fonts in Commons 😊 −Ebrahimtalk 17:06, 21 July 2024 (UTC)[reply]
  • Hi, I tend to look after the font articles on Wikipedia. Because fonts are software, they can often have updates fixing glitches or adding new characters, see the github pages for Noto and other actively maintained open-source fonts like the Adobe Source project. Also, there are lots of random self-published fonts out there, dafont claims to have 86 000 currently, that would be a lot of work to manage if people started uploading all their own projects.
    For this reason, I think it would be simplest not to do this and instead have relevant Wikipedia articles link to creators' websites or repositories like GitHub, the Open Font Library or Google Fonts where up-to-date versions can be downloaded (for example, the article on Linear B cites specific fonts you can use for it).
    Commons doesn't act as a general repository for open-source software that I know of, and although I don't know if there have been specific reasons why not, my guess is similar concerns apply there (project scope, updating, managing problems like buggy versions or forks, complexity). (Or of self-published creative projects like self-published fiction, if you feel that's a better analogy to self-published fonts.)
    So at the moment instinctively I'm leaning no but am willing to consider arguments to the contrary e.g. if there's reason to believe current font repositories are untrustworthy or likely to close down. There may also be security concerns as fonts can have quite complicated code inside them, that can at least potentially be used as a vector for viruses.
    Lastly, I should say that Commons acting as a repository for fonts is a different issue to Wikipedia pages serving webfonts to display text, that would be best taken up with engineering teams. Blythwood (talk) 01:22, 23 July 2024 (UTC)[reply]
    I feel like fonts are software the same way an SVG file is software. Bawolff (talk) 10:22, 29 July 2024 (UTC)[reply]
    Exactly, this definition of software is too broad that perhaps can contain fonts in PDF that already exists in Commons or multi lingual SVGs. Wikimedia Commons is even more suitable than GitHub to host multi megabyte files and just imagine how great it can be to manage and categorize them using MediaWiki provided facilities and that's for sure can go even better than sites you mentioned, categories, languages, collaborative way of filling metadata. Re complicated code, there are software to check and sanitize font files, such as OTS and we can put them at the first front just like Chrome and Firefox but likewise PDF and SVG files also can contain JavaScript codes in theory, or there were always news about different picture formats vulnerabilities, sure fonts that are served to billions of users everyday as web fonts aren't that different from other types of media formats. −Ebrahimtalk 19:35, 31 July 2024 (UTC)[reply]
We already have so much half supported content formats ? This seems like a thing with pretty limited use, someone can just build a Toolforge tool that hosts a font library ? We already have https://fontcdn.toolforge.org/ for instance. —TheDJ (talkcontribs) 13:54, 24 July 2024 (UTC)[reply]
It's a bit weird you feel that way but I understand, just to note all that facilities to manage types of fonts, managing and filling metadata collaboratively that are unavailable on the alternative approach you are suggesting. A third party website won't be able for example to generate icons from multi mega byte Emoji fonts like the use Emoji that each Emoji is uploaded separately but an extension can facilitate that, I believe. I say let's be inclusive about such ideas 😊 −Ebrahimtalk 19:35, 31 July 2024 (UTC)[reply]
I'm not saying we shouldn't support fonts, but I was just thinking about that with Category:Tabular data. Really if anything the amount of formats supported on Commons should be decreased. I'm not a big of extending the meaning of "media" to anything outside pictures, video, and maybe ebooks either. otherwise you get into some really niche file formats that are questionably educational and better supported by other projects or websites anyway. --Adamant1 (talk) 21:18, 24 July 2024 (UTC)[reply]
I feel like tabular data has been a success also, at least in my local wiki we use it's updated data, weird it doesn't seem that way to you. I believe if glyphs extracted from a font are educational and are uploaded to Wikimedia Commons, the fonts they are extracted from can be educational also and that way we don't need just to extract all those glyphs each time and manage their versioning with difficulty. −Ebrahimtalk 19:35, 31 July 2024 (UTC)[reply]
I think this is a good idea and clearly has usecases. One practical point however is the detection of copyright violations. What's stopping a person from claiming Arial as their own work? We'd need a metadata scanning system to check for copyvios as well. —Matrix(!) {user - talk? - uselesscontributions} 15:45, 31 July 2024 (UTC)[reply]
Fortunately fonts do have metadata that aren't editable as easy as EXIF metadata for regular users 😊 and the same concern of course applies to say PDFs or STL files, let's don't underestimate the power of community. −Ebrahimtalk 19:35, 31 July 2024 (UTC)[reply]