(en)(blog) graphical representation of languages?

locale/en/blog/graphical-languages.md

@@ -0,0 +1,90 @@
+# Graphical representation of languages?
+
+<small>2024-08-17 | [[@andrea](/@andrea)</small>
+
+![A collage of a bunch of methods mentioned later in the article (list of languages with graphemes, a map, Duolingo's flags, combined flags…)](/img-local/blog/graphical-languages/graphical-languages.png)
+
+Placing graphical elements next to text labels is a cool design trick – it helps your brain skip a few steps when
+navigating a user interface: instead of reading the label it can immediately know what a button is doing just by glancing at the icon.
+
+That's the reason why most buttons on our website have a corresponding icon – and you can only imagine how much easier
+it makes the work on the project! I can easily navigate *any* language version without understanding a single word of it,
+just because I remember which icons are marking which button.
+
+It's also a reason why many user interfaces put flags next to language names. Is it really a good approach, though?
+After all, flags represent countries, not languages – any if the ideological and political complications
+of mixing up those two concepts don't convince you,
+the practical issues hopefully will: languages simply cannot be mapped one-to-one with countries.
+
+First example that comes to mind: which flag should represent English?
+It originated in 🏴󠁧󠁢󠁥󠁮󠁧󠁿 England, but it's also spoken / official  in the rest of 🇬🇧 United Kingdom,
+in 🇺🇸 United States, 🇦🇺 Australia, 🇨🇦 Canada, 🇮🇳 India, and many others.
+Why would we pick one over another? If we go with place of origin, would people even recognise the flag of England?
+If we go by population size, wouldn't the flag of India make people think of Hindi rather than English?
+Similar issues arise with other languages. Which flag should represent Spanish? 🇪🇸 Spain, 🇲🇽 Mexico, 🇦🇷 Argentina, 🇨🇴 Colombia?
+Which flag should represent Portuguese? 🇵🇹 Portugal, 🇧🇷 Brazil? Which flag should represent Arabic?
+🇸🇦 Saudi Arabia, 🇪🇬 Egypt, 🇮🇶 Iraq, 🇮🇷 Iran?
+
+Many languages aren't recognised as official languages of any country. Some groups of speakers, like Silesians or Basques,
+have a flag for their region, but others, like Ladino speakers, don't – they don't even occupy a single region.
+Some constructed languages might have a flag, like Esperanto does, but others simply do not.
+And even if a good candidate for a flag exists, it might not be universally accepted by the speakers of the language,
+and it might be harder to use – for example, I was able to easily add country flag emojis to this text,
+but embedding an Esperanto or Basque flag in a text would require way more effort.
+
+People have been designing flags that would represent languages better than country flags – for example
+by combining elements of flags of countries where the language is spoken. Here's a fun video exploring a bunch of ideas:
+
+{embed=//youtube.com/embed/2I3UJkDmMco=What are "Language Flags?" (And Should They Exist?)}
+
+As the number of language versions of our project grows, we've been discussing ideas to make it easier to find the one people are looking for 
+– without using country flags.
+I've created a proof-of-concept for an approach in which we pick a single grapheme from a language that can represent it –
+for example `ß` for German, `ñ` for Spanish or `ą` for Polish – and we put it on a background colour that represents the language family –
+like Germanic, Romance or Slavic, respectively.
+
+![Other language versions: ß Deutsch, ñ Español, õ Eesti keel, , ç Français, ny Ladino (Djudezmo), ij Nederlands, æ Norsk (Bokmål), ą Polski, ã Português, ă Română, å Svenska, ớ Tiếng Việt, ش العربية (الفصحى), й Русский, ї Українська, の 日本語, 人 中文](/img-local/blog/graphical-languages/graphemes.png)
+
+This approach is not ideal either. It's still a very arbitrary choice: why not use `ł` for Polish? Should Swedish get `å`, or should Norwegian?
+The English alphabet is quite boring, should we just pick a random Latin letter, or maybe take a risk of not being understood and use historical `þ`?
+Those graphemes are admittedly less recognisable than the flags we're used to seeing representing languages –
+while it might be a good idea in a long run, it doesn't really help much short-term.
+
+Another approach we've been looking into was using a map. While speakers of any language might be located literally anywhere,
+it is generally possible to mark a general area where the language is commonly spoken.
+Unfortunately, finding an open source dataset of those areas is not easy,
+and even if we had it, we'd still need to deal with issues like overlapping territories,
+or some areas being too small to click on while others would span across continents.
+Instead, we tried using a dataset of simple lat/long coordinates for each language from [The World Atlas of Language Structures](https://wals.info/),
+each language being represented by a single dot on a map, in a general “centre of gravity” as picked by the linguists who created that dataset.
+But we're still not really getting rid of the main problem with mixing up languages with countries –
+while it might help Polish speakers quickly find and click on the dot in the middle of Poland,
+for users from the US it probably wouldn't be obvious to look around England.
+And where would we even put Esperanto or Toki Pona?
+
+![Map of the world with a bunch of circles representing general locations of languages](/img-local/blog/graphical-languages/map.png)
+
+This map is not supposed to be an accurate mapping of languages, the idea was not to draw out where exactly each language is spoken – 
+that's a very difficult thing to do (and also not our job). The map was supposed to be just an extra tool to help with the overwhelm when visiting
+the page and seeing an ever-growing list of languages without any graphical helpers (like flags).
+It would allow some users to immediately find the language they're looking for by just glancing on a map – 
+while others could still fall back to the simple list of languages.
+But ultimately, we decided that this approach is too far from good to actually put on the website.
+
+What we ended up implementing on the main homepage at [pronouns.page](https://pronouns.page),
+is simply adding a few helpers to the existing “raw” list of languages.
+We can use the list of preferred languages as specified by the user in their browser settings – and show a few suggestions on top;
+this way it's very likely that they'll see what they're looking for right away.
+Underneath there's a full list without any extra graphical elements, but now it gets a filter –
+one can just start typing the name of the language they're looking for
+(either [endonym or exonym](https://en.wikipedia.org/wiki/Endonym_and_exonym)) or its ISO code – and they'll find it right away.
+
+![Screenshot of the language selector: on the top a few suggestions based on browser headers, lower the full list with a filter field](/img-local/blog/graphical-languages/filtering.png)
+
+The list isn't perfect either. How do you even sort languages alphabetically if they use different scripts?
+We've settled on a pretty eurocentric approach, due to the composition of the list of languages currently present,
+but there's no ideal answer…
+
+So, what do y'all think about this issue? Do you have better ideas?
+Is there an approach that would solve all the issues and become a useful standard for listing languages in user interfaces?
+We're very open to experiments!