Aaharoni-WMF

Hint

On Benutzer Diskussion:Kritzolina#Ton I wrote something (in German) about your Kurier article. Maybe you are interested, maybe not. I only want to inform you.--Mautpreller (Diskussion) 21:41, 14. Okt. 2017 (CEST)

Mautpreller, of course I'm interested. Thank you.

I'm also writing other replies to you, they are just a bit long :) --Amir E. Aharoni (WMF) (Diskussion) 12:29, 16. Okt. 2017 (CEST)

It's a bit difficult of a discussion, I would gladly switch to English, because that IS the lingua franca of the world, but then again, some German speakers don't seem to be comfortable to switch. Let me summarize my points here: The division you chose just sucks. *Nobody*, not a linguist, not a sociologist, not even a person who did not think about language at all, sorts languages after continent. And even if I would do such a strange thing, why would Ladino pop up on the Americas?!? GMFB! And why is Volapük, a language maybe 20 people understand, a "world language"? VOLAPÜK IS ABSOLUTELY USELESS. Nobody wants it, get rid of it.

Why am I clustered into two German dialects all the time, even though, one of them I don't even speak? Why do you think, that, when I am reading a Croatian article, I may know Serbian, bt not Bosnian or Serbo-Croat? Okay, I'll leave it here, I'm already hyperventilating enough. Mautpreller's argument, that you make it difficult to get additional information is another important argument and User:Hund96 made very good points as well. ATB, fossa net ?! 13:33, 16. Okt. 2017 (CEST)

Fossa, let me start with something very serious and personal: I'd be the last person in the world to tell you not to speak your language. I actually already had read your comment at the Kurier talk page with Google Translate and I meant to reply there, but thank you for writing it here in English.

I explained the division into continents in one of the replies. It's not a matter of linguistics or sociology. It's a matter of design. It's easier to understand long lists when they are divided into parts, and dividing into continents was the simplest and most balanced to divide. If you have a different suggestion, I'll be happy to hear it.

But the details of the division are not important, because we analyzed the usage and we saw that most users don't actually scroll through the long geographical list, and just use the search box.

Ladino appears under America because it is spoken by some people in America. Not a lot of people, but definitely a few; I know at least one in person. The Wikipedia article says so, too.

I agree that the location of Volapük and some other artificial languages (like Novial and Interlingue) is imperfect. We are thinking of a better solution, that will work well with the design and be fair to the people who care about these languages. Again, however, the really important part is that users are able to find their languages, and having a couple of more languages under "World" doesn't actually get in the way. In a short list most people simply ignore the languages that they don't need.

About Serbo-Croat and its variants, that's an interesting point, and the response has several parts:

First, the technical reasons: The Compact Language Links feature chooses the languages by the user, not by the project. Core MediaWiki and Wikidata have an additional feature, which allows prioritizing some languages in a particular wiki. This feature has existed long before Compact Language Links, and it still exists. For example, the Hungarian Wikipedia asked to prioritize English, and Norwegian Bokmål asked to prioritize Nynorsk, Danish, Swedish, Faroese and Icelandic. This is up to the editor communities. If the Croatian Wikipedia editors community thinks that it's useful to prioritize Bosnian and Serbian, they can request it and it will be done.

Now let's take a look at actual usage. Most readers of the Croatian Wikipedia are in these countries:

1. Croatia - 11 million pageviews per month
2. US - 5 million pageviews per month
3. Bosnia and Herzegovina - 3.4 million pageviews per month
4. Germany - 1.1 million pageviews per month
5. Serbia - 1 million pageviews per month
6. France - 0.5 million pageviews per month

The curious point is that these numbers are almost the same in February 2016 and in February 2017, that is before and after the deployment of Compact Language Links as the default view to anonymous readers.

Now let's take look at the most popular projects that are read from the country of Croatia, this is the data for February 2016, after the deployment:

1. English Wikipedia - 17.2 million pageviews per month
2. Croatian Wikipedia - 11.4 million pageviews per month
3. Bosnian Wikipedia - 1.1 million pageviews per month
4. Serbo-Croatian Wikipedia - 0.62 million pageviews per month
5. German Wikipedia - 0.45 million pageviews per month
6. Italian Wikipedia - 0.29 million pageviews per month
7. Serbian Wikipedia - 0.19 million pageviews per month
8. ... some other languages
9. Russian Wikipedia - 0.042 million pageviews per month

In February 2017, after the deployment, these numbers were also almost identical, with only a few changes:

1. English Wikipedia - 16.3 million pageviews per month (a bit less)
2. Croatian Wikipedia - 11.2 million pageviews per month (a bit less)
3. Bosnian Wikipedia - 1.1 million pageviews per month (same)
4. Serbo-Croatian Wikipedia - 0.64 million pageviews per month (a bit more)
5. Russian Wikipedia - 0.42 million pageviews per month (significantly more)
6. German Wikipedia - 0.40 million pageviews per month (a bit less)
7. Italian Wikipedia - 0.28 million pageviews per month (a bit less)
8. Serbian Wikipedia - 0.24 million pageviews per month (a bit more)

So you can see that the only considerable change was in the growth of Russian, and a little bit in Serbian. The Compact Language Links feature had very little effect here, and if it did anything at all, then it made finding Russian and Serbian easier for people who need it, and maybe they changed for other reasons.

To sum up the Serbo-Croation topic: While I understand why do you expect to see Serbian, Bosnian, and Serbo-Croatian from the Croatian Wikipedia, it's up to the community to change this, and even if they change it, I'm not sure that it will affect the actual clicking and pageviews. I'll be very happy to talk to people who are well-familiar with these countries and languages, and see how we can improve their experience. (My guess is that the Serbian Wikipedia is less popular in Croatia because people in Croatia mostly use search engines in the Latin alphabet, and the Serbian Wikipedia's default alphabet is Cyrillic. Perhaps we should expose its Latin version better to search engines, but this is completely unrelated to the interlanguage links, which is what we're discussing here.) --Amir E. Aharoni (WMF) (Diskussion) 14:21, 16. Okt. 2017 (CEST)

@user:Mautpreller: A „hint“ is not an innocent „Hinweis“, it implies a serious information asymmetry, you'll find hints often in textbooks under the heading of "exercises". „FYI“ would IMNSHO fit better, or „PSA“ as a quip. NB would also do, methinks. fossa net ?! 13:56, 16. Okt. 2017 (CEST)

Okay. FYI would have been the best ...--Mautpreller (Diskussion) 14:15, 16. Okt. 2017 (CEST)

I think that I understood everything :) --Amir E. Aharoni (WMF) (Diskussion) 14:21, 16. Okt. 2017 (CEST)

There are a lot of problems, some of which I told you resp. User:Pginer-WMF on MediaWiki. Concerning the problem that Fossa adressed: There will be hardly any user who will profit from your "initial selection". Bavarian ("Boarisch") is not really a language but a dialect. It is not even reasonable to assume that Germans usually understand Bavarian (I do, because I come from there); if there is a movie in German television in dialect dialogues will often be subtitled. The number of persons who are interested to read anything in this dialect is probably extremely small. Much smaller than the number of persons who read Portuguese, Swedish etc. "Plattdütsch" is a slightly different case: there is at least written literature in Plattdütsch and there are (few) professors for this idiom and there is a number of people who are interested and able to read it (not me, because I don't come from Northern Germany). There is controversy whether this is a language or a dialect. However, still the number of people that are able to understand this language/dialect is relatively small and the number of people who are interested to read articles in this idiom is even smaller. So the selection is not useful for the task to make language selection easier. But it is possible to give another reason: one might include them in the initial selection exactly because they are small idioms (you could compare it to diversity conservation). It is obvious that both reasons are directly conflicting.

Languages versions like Boarisch and Plattdütsch but also language versions like, say, azerbaidshanic language will only be clicked if one knows that an article in these languages exists at all. In order to know this you must see them all. For my part: I wrote a "featured article" about Requiem (Mozart) in German. Seeing the full list of languages in the side bar, I discover that there is also a featured article about this subject in azerbaidshanic language. I never even knew that there is a Wikipedia in this language, let alone that they have a featured article about Mozart's Requiem. I can't read azerbaidshanic. However, I clicked upon it and saw that it is a translation of my article. Interesting thing. I shouldn't have known this if your tool had been activated for me.

The "continent" selection isn't useful at all. If you want an English article you won't look under "Middle East" etc. If you're really looking for a regional language this presentation is next to useless. No one will use it. So why don't you simply activate the new Search function (which is definitely a fine thing) and leave the language list as it is? And why don't you provide a complete and searchable list of language versions so one can immediately see which language versions exist? --Mautpreller (Diskussion) 16:24, 16. Okt. 2017 (CEST)