Community Call: Managing Multilingual Communities

General Q&A
1

Our next community call, on Friday June 28, 2019, 2PM PST, will be on Managing Multilingual Communities.

rOpenSci’s community is increasingly international and multilingual. While we have operated primarily in English, we now receive submissions of packages from authors whose primary language is not. As we expand our community in this way, we want to learn from the experience of other organizations. How can we manage our peer-review process and open-source projects to be welcoming to non-native English speakers?

Our guest speakers will include:

The call will be moderated by Melina Vidoni, an associate editor for rOpenSci software peer review. (@melvidoni)

We want to hear from you! What has been useful or not useful in building mutli-lingual communities and projects you have been involved in? Add your comments below and we’ll organize our panel questions and topics around your contributions and questions.

5 Likes
2

As a non-native English speaker I think it will help to expand the current documentation about having several languages supported by the same package. This way we can ease the adoption of a package without restricting (much) ourselfs to a language.

I think it has something to do about the /po folder of a package but I couldn’t understand how to use it/fill it. It is only briefly commented on advanced R (http://r-pkgs.had.co.nz/misc.html) and on the R manual itself (https://cran.rstudio.com/doc/manuals/r-devel/R-exts.html#Internationalization 1.8.3 Preparing translations)

3 Likes
3

As someone from Sub saharan Africa, i would say a huge percentage of devs/analysts use English, French and a few Portuguese speaking countries like Angola, Cape verde and Mozambique. We’ve seen past organizations like Ushahidi working with crowdsourcing any localization content using tools like POEditor as an initial start to building the communities. I’ll be happy to join and share my perspectives from E.Africa.

3 Likes
4

Hi everybody, I’m very much looking forward to the discussion.

I’d like to hear from people about the different ways in which they think language is a barrier to their becoming more involved with organizations like ROpenSci.

For instance, as Editor I have heard everything from “All I need is for people to speak a bit more slowly” to “I am worried that people won’t take my review seriously because it might have some grammatical mistakes” to “more nuanced aspects of reviews are often easier to communicate in my native language, so working with Editors that understand that language can help”.

Any thoughts?

3 Likes
5

re: @embruna’s example

more nuanced aspects of reviews are often easier to communicate in my native language, so working with Editors that understand that language can help

:thinking: Among rOpenSci Editors, we have English, French, Spanish, Greek native speakers and likely other languages. Acknowledging that it might be extra work for some, but would we give an option for a package author to declare their “comfort” with a specific language when English is not their first language?

We need to keep reviews transparent so if some nuances were discussed in another language, I imagine the essence would have to be translated to English.

2 Likes
6

I can share my own perspective. Even little things, such as thinking of a package title or parameters description, can be an obstacle to create and share R packages. In these cases, two things could be helpful: (a) creating some examples how to write them (e.g., good practices or FAQ), (b) encourage people to submit their work even if they aware of the language issues and then provide some support.

2 Likes
7

Hi Jakub @Nowosad. For package naming guidance, we have a basic start in our dev guide. That looks like a good place to add more. Have you seen any good examples of this kind of guidance in context of native language that we could link to?

8

Piece in Nature Careers section apropos of our upcoming call: https://www.nature.com/articles/d41586-019-01797-0?WT.ec_id=NATURE-20190613

2 Likes
9

Looking forward to participate to this call!

I’m from Democratic Republic of Congo the first french speaker country. English is a barrier for many R learners. To overcome this when you teach R you must first explain English simple words it is not easy!

We need a inclusive community where everyone despite is language can be happy to work in.

2 Likes
10

Hi, I am super sorry, it’s very likely I won’t make it to this call. In absence, my 2 cents about this topic.

I also believe that English is a barrier to non-native speakers. I am privileged enough to feel comfortable with English and navigate the English-speaking world naturally, but I know of tons of highly talented people that don’t and are left out of the conversation when it is in English. I have experienced this first hand in the R-Ladies Global organizer’s Slack and have been able to contrast participation of the same people in that English-speaking Slack versus their participation in a Spanish-speaking Slack. So many wonderful ideas that are not present in the first Slack because of the English barrier! :frowning_face:

I also had students asking me for books in Spanish when it comes to learn intro applied stats. It is a lot more difficult to learn a topic when you have the additional English barrier. That is why I add my two cents to translation projects such as that of the R for Data Science book or translating materials from the Carpentries. I am a firm believer that timely material translation helps a lot.

I thought perhaps the results of a little survey I run a couple of months ago in R-Ladies Buenos Aires and R en Buenos Aires (Buenos Aires general RUG) Slacks may be of interest in the discussion:

I surveyed through a Slack poll (unblind) these two RUGs. R-Ladies has a more novice profile than the general RUG. Responders are by no means representative of the wider Spanish-speaking community but this is way better than my opinion only. Most of the responders hold PhDs/MS in various disciplines, that is, they know how to go around looking for help. They are also quite connected to the more general R community.

I asked how they search for R help when they have a question - some people answered more than 1 option. RL: R-Ladies answers. RiBA: R in Buenos Aires answers.

I search in English (even if English may not be 100% comfy for me):

  • RiBA = 20/23
  • RL = 12/14

I use translated materials (cheatsheets and R4DS in Spanish):

  • RiBA = 3/23
  • RL = 2/14

I use google translate to translate materials in English to Spanish:

  • RiBA = 2/23
  • RL = 4/14

I use Stack Overflow in Spanish:

  • RiBA = 0/23
  • RL = 0/14

I use community networks in Spanish like Slacks, Telegram groups, #rstatsES in twitter, etc:

  • RiBA = 10/23
  • RL = 3/14

I also asked “Would you use something like https://community.rstudio.com/ in Spanish?”

  • RiBA = Yes 7/20, No 11/20, Other 2/20, if it had better and more content than the English version, if the answers were fast, I would use both
  • RL = Yes 11/11

An interesting comment: a community in Spanish could introduce topics specific to the language like text mining questions or UTF-8/ASCII problems due to special characters like ñ.

I hope this helps. Looking forward to the reading more comments and a summary of the discussion during the call.
Happy call!
Best,
Laura

1 Like
11

No, I haven’t seen any good examples…

Another idea I can think of is to encourage non-native speakers to share their materials (e.g., blog posts, slides, etc.) and ideas in English. I learned a lot (both about the topics, but also about the English language) by creating those materials - and even when they have not been in perfect English, I was never criticized for that.

1 Like
12

Yes, we have one example of a package where an author made use of the /po folder to have messages in multiple languages - Rpolyhedra by Alejandro Baranek et al: https://github.com/ropensci/Rpolyhedra .

I think @thosjleeper wrote or was involved in a package to make setting up these translations easier, but my Google-Fu is failing me.

1 Like
13

Huh, Edgar Ruiz has also made a framework for translating documentation and data sets: https://github.com/edgararuiz/datalang, most prominently used for https://github.com/cienciadedatos/datos.

1 Like
14

Thanks @noamross for the link. I’ve seen a couple of packages with said folder, but I never found how or which messages are translated. I probably would need more time to learn how it works… but having a list of packages which show support for several languages could be helpful.

I won’t be able to attend the community call, but hope to read the resulting blog post with the summary.

1 Like
15

Congrats on the initiative! Sadly, I won’t be able to participate but would love to read the outcome. Still, I would like to highlight one issue that, at least in Mexico, may be hindering progress: the lack of a consolidated community devoted to discuss this (an other) issue(s). Most efforts are individual ones and on a case-by-case basis. Of course, there are some efforts but still minor (to my view; and usually at large universities) and they certainly don’t deal with language-barrier issues (as far as I know). Perhaps reaching out to those who are leading some of these efforts can be a way forward.

I agree with @llrs in having several languages supported by the same package and with @lauracion on the specific comment about language-specific topics regarding special characters (I have always found that very frustrating). Finally, examples like the one @noamross provided, seem like a potentially simple solution (provided someone(s) would do so for as many languages as possible).

Thanks!

2 Likes
16

The web page for this Community Call on Involving Multilingual Communities now has the 1-hr video, along with speakers’ slides and collaborative notes doc with answers to questions from the Q & A.