User:Contraexemplo/Outreachy/Daily notes/February 2018

February 8

I finally have time to write down my thoughts. It's been a while, right?

The last few weeks were very exhausting and hasty. For instance, I barely had time to sit down and write my bi-weekly reports — I was constantly focused on reaching out and talking to people. I felt a great amount of guilt if I tried to do something else, fearing I was losing time I could dedicate to the outreach. Fortunately, things are starting to get back to normal now.

Wikimedia blog post

This task certainly stalled for some time, but I finally ended reviewing what I wrote and covering every single important point. Since others review might take some time, it may suffer some changes and I am two bi-weekly reports behind, I decided to publish my version on my blog.

• English version: Bringing documentation to light
• Brazilian Portuguese version: Colocando a documentação em foco

It is frustrating that I didn't submitted it this way some weeks ago, but it also has some advantages:

• I found the EF EPI index, which was really useful to build some hypotheses and understand which native speakers of some languages need special attention.
• Displaying data of the Help:Contents translation rate and pageviews side by side was also really insightful. I am especially impressed with how much people who presumably speak Arabic look for user guides, even though the translation rate is low. It makes me wonder how much of the pageviews of the English version of Help:Contents are from people whose native language isn't English.
• I wanted to record video guides but editing videos and having good audio equipment is difficult. I ended up writing a shorter version of the Translate extension documentation with some complementary visual resources. I tried to do what I wished I had when beginning to contribute to MediaWiki: a straight-forward guide that let me know the most important aspects of the tools I am about to use. I want to improve it later, but it was certainly uplifting to see it ready.

Outreach

Things happen in a very slow pace and this is absolutely frustrating but expected. Most of universities are starting their academic years right now, so I didn't receive reponses in the first days. I decided to contact students directly, and most of them pointed out that Facebook is the main mean of contact. I tried to create an account there (didn't have one for a long time!), but that attempt was unsuccessful due to Facebook policies on disclosing the user identity. Providing the information they asked didn't work, and soon they deactivated my account.

My next step was using Twitter to find students of various universities and ask for their help. People from different universities (UFSC, UnB, UFOP, USP) offered me help. I ended up creating some promotional material though I am no designer, and sending them instructions on how to present the initiative.

Published this tweet to test out a flyer I made exploring the relationship of Wikipedia with MediaWiki.

Since my attempts to contact people from my own university weren't working, I went there in person a couple of times. First, in the department responsible for all the means of communication -- social networks, newspaper, website. I had a talk with the director of the department explaining the nature of my project and they said they are willing to help. I sent them the material and instructions, but I had problems with their email firewall and had to send it again a couple of times using my university's email address. I also sent it out to my own undergraduate colleagues and all tech-related undergraduates.

I also had a meeting with the coordinator of the language school of my university. She thinks it will be more effective to meet students in their classrooms and talk about the project, presenting it as an opportunity to make volunteer work. However, she expressed some concerns: that, naturally, knowing a foreign language doesn't make you a good translator; and this may disturb the learning process for some because they advise against building bridges between two languages -- they want students to associate words with ideas.

During this period, I couldn't help but question: are we really ready to welcome new translators? Or people that know other languages, but never were a translator? Those people need proper training -- nothing fancy, something that will help them understand the process behind translating. After all, bad translations can be really as harmful as no translations at all.

February 12

Looking back

I updated the Outreach strategies page with more information about the initial outreach. I will add more information tomorrow exploring solutions to the problems in the translation process I mentioned. Today, I focused on evaluating my efforts and asking myself what went wrong, what did work and what I can't really be certain of (and why).

I was feeling really hopeless when I had the meeting with Benoît and Johan today. Sometimes it feels like this is a task beyond my limits. I think about how it could work if I had more time, or I had more technical power to solve some things and get pretty pessimistic. But I am trying to focus on thinking that I explored a lead they wanted to for long a time. If this didn't happen, they would never know things like proper proceedings and level of difficulty to make this work. So this was, to some extent, a useful experiment.

Changing focus

I will still see if I get any response from the initial outreach, but now that we know university students as a whole are difficult to reach and build a consistent strategy, I will focus on only those who are studying to become translators. While anyone can be a translator with a bit of dedication and training, those people have a special motivator to engage in this kind of activity: professional experience.

• Johan suggested that I could try to create a translation team to test this kind of experience. This is actually a wonderful idea! Translation efforts in the Wikimedia movement are usually anarchic, lone and independent. This isn't bad per se, but having teams is useful to coordinate efforts and establish conventions.
  • So, first, I will look for students with the intent of building a team. This is way more direct than simply promoting the role of the technical translator. I will reach them making explicit that my purpose is to create a force task to improve localization practices on MediaWiki. This is going to happen this week, after the evaluation period is over. If needed, I will add one to two days to this outreach period.
  • Then, after gathering two from four people, I will ask them about their level of experience and what they think they will need to perform this task (like documentation) and how many hours they can dedicate to this task. I will make a training and take notes on what they find sufficient and insufficient or confused and well explained.
  • Together, we'll define a couple of pages to translate in the next four days and create a shared document (or page -- it would be awesome if I could encourage them to use wiki tools only to make the process public). If there is something, like a word, which meaning is not clear, we write it down in the document and have a little discussion about it.
  • After the four day period, I have a meeting with them and discuss what they think about the translation process, if they intend to keep contributing, what might keep them engaged and what could be improved.

Now, we were talking about having a way to reward them. What if I gave them a certificate of participation after this experiment for their help? This would be beneficial for us all. So here's a task for tomorrow: search for how I could make this happen.

February 13

Arrangements for next steps

Continuous outreach

I am taking a look at the requirements for extracurricular activities for language students in my university. They have a 20 hour limit for online activities. On the other hand, translation students at Federal University of Uberlândia have a 80 hour limit for activities related to learning useful techniques for translators. I'm going to contact both, and hopefully I can get a really diverse team.

Meanwhile, I need to make some preparations to receive them:

• Make a standalone page here explaining the project. It should include:
  • what is this project
  • its scope
  • who to contact.
• Improve the Translation quick guide.
  • Include helpful resources such as Manual:Glossary and Wikipedia:Glossary to present them to the terminology in Wikimedia projects.
  • Review the current content and update the writing style.
  • Pending: Maybe move it to Project:PD help.

Final report

While I designated a final evaluation on March 5, I also have a final report as a deliverable and it's unrealistic to think I could write it in one day. It's time to review my notes and begin to write in depth about what I learned throughout the internship and my recommendations for the future.

February 14

Restarting the outreach

Carnaval is finally over in Brazil, so I sent emails and messages to everyone that I didn't heard back until the end of the initial outreach and first evaluation to see if there's a chance of building more bridges.

I also thought about the need of having a better page to introduce my soon-to-be teammates to the most basic concepts behind this project. Having that in mind, while waiting for responses, I am writing a Beginner's Guide in Brazilian Portuguese (and when it's finished, I will translate it to English as well). I am using Write the Docs and SUSE documentation style guide as references to achieve a better text quality.

February 15 & 16

Current progress

So far, I was contacted by three people on Twitter willing to participate of my experiment. One of them, in particular, said it was great to find this opportunity because they were looking for this kind of activity. This is a good sign. I think I will have the definitive number of teammates on Monday.

Bi-weekly reports

Meanwhile, I wrote two bi-weekly reports...

... one about commercial social media, which is more of a vent than anything else...

• English: Some loose thoughts about commercial social networks
• Brazilian Portuguese: Alguns pensamentos soltos sobre redes sociais comerciais

... and another about my feelings about this internship final moments, and my latest strategy...

• English: Switching strategies and recalibrating expectations
• Brazilian Portuguese: Mudando de estratégia e recalibrando expectativas

... which means I am finally up to date with my bi-weekly reports! Yay!

MediaWiki exploration

I mentioned that three people showed their interest, but a fourth person has already begun to do their work: User:QtK6z. I had the opportunity to watch as they tried to understand the path to become a contributor, from a Duck Duck Go search for "wikimedia" to their first interactions with the Translate extension. Their thoughts:

• The contribution process isn't much clear. Beginning on Special:LanguageStats is quite strange; starting on Special:Translate would be better.
• The Translate extension should warn Tor users their contributions won't be submitted 'before' they try to contribute. The way it warns users their translations won't be saved is too subtle.
• A glossary is a must, as well as translation conventions. "I was unsure about if I should translate certain technical terms or not", they said.
• "This is actually fun, and really similar to the way Duolingo was designed. In fact, I would rather do this. It feels more useful."

February 19

So, these are the pages Johan suggested:

Page	Messages	Untranslated	Completion
Help:Images	204	171	16%
Help:Links	132	132	0%
Help:User contributions	96	75	21%
Help:Tables	412	371	9%
Help:Cite	94	94	0%
Help:Structured Discussions/Quick tour	83	41	50%
Help:Range blocks	93	74	20%
Help:Log	12	12	0%
Total	1,126	970	24%

I see a problem, though. Let's be optimistic and say all four people are going to be engaged in this activity. We have 8 pages, and but some are too large and some are too small. I need that at least every person has a page so they can understand the process (under, let's say, 50 untranslated messages), a larger page to translate completely (under 100 untranslated messages) and an even larger one that they should translate significantly (above 100 untranslated messages). So the ideal is to give each person a set of three pages and give them the option to explore others if they feel like it.

I sent an email for those who showed interest in participate with the first steps: creating an account and exploring the wiki. Let's see how it goes.

Update: At 5 PM (local time, UTC -3h), one person confirmed they will participate, one confirmed they can't. One person hasn't answered the email yet. With one in and one out, and one unsure, I guess we can work with those pages and see what we'll get.

February 20 & 21

We have, officially, a team of four people (including me). I spent the last days helping them get familiar with the work environment, presenting them concepts like wiki markup and pages like Manual:Glossary. At first, I'm letting them choose which page they want to translate. I'm giving them comfort and freedom to schedule their work when it seems fit. If I don't see much results, I will change my approach.

I'm encouraging them to take notes on everything they find difficult or any doubts they have throughout the translation process. This data will be useful when discussing what kind of documentation a localization team needs.

Meanwhile, I've been thinking about how I'm going to structure my final report. Yesterday, Benoît talked to me about Wikimedia certificates, how some users have asked for them and how issuing them would be scalable if we determine some thresholds (a X number of edits and thanks, period of contribution, etc). This is, as we discussed in the meeting on Monday, a really great idea. But it needs validation from the Wikimedia community.

February 26 & 27

This is the last week of my internship. I can't believe it's almost over.

I'm focusing on producing my final report this week. I'm re-reading my daily notes, trying to write down what I think it's important to mention, organizing my ideas and interviewing my fellow translators. I'm really satisfied my last strategy worked and now we have two more Brazilian Portuguese translators helping us out!

If I have any time left, I plan on talking to Haytham abulela about Arabic translations. Arabic was one of the languages that stood out for me because of the great amount of documentation pageviews and low translation rate.

I need to find some time to write my second-to-last bi-weekly report. Not sure about what I'm going to talk about, though (maybe about why I think that, systematically, technical translations need a more organized workflow than, for instance, editing Wikipedia pages?).

Lastly, I decided to leave this account untouched after the internship ends as a way to preserve its contents. I'll be on Anna e só after that, as I plan on continue in the movement as a volunteer in the Brazilian Portuguese translation team and will probably go to a great number of events this year in my country to talk about my internship project (and introduce people to the Wikimedia movement!).

February 28

Answers from my teammates

I made some questions to my teammates about their experience with MediaWiki and the Translate extension in the last few weeks. I'm publishing it here in Brazilian Portuguese so I can find it easily later. I do plan on translating it to English, though.

• Pergunta: Quais foram as suas primeiras impressões a respeito do MediaWiki? O que é claro e o que é confuso?

• Respostas:
  • QtK6z: Talvez o site seja um pouco confuso ou sobrecarregado para a maioria das pessoas. Eu por exemplo tive que procurar um pouco até achar a página que me permitiu traduzir. Também não faz sentido para mim permitir que um usuário abra a página de tradução, escreva a tradução e só na hora de salvar seja avisado que não pode editar usando um nó de saída do Tor, depois de todo esse trabalho. Em circunstâncias normais fecharia a página e nunca mais pensaria em contribuir. A página para a qual o usuário é levado quando clica em traduzir não é muito útil, não só por não ser intuitiva(o que é ok) mas também por não ter explicação nenhuma sobre o que fazer naquela página. Também não sei quais páginas são consideradas mais prioritárias e devem ser traduzidas primeiro, então meu critério é meio aleatório. Fora isso, não lembro de nenhum outro problema até o momento que começo a traduzir.
  • Viviannesh: A primeira impressão que tive sobre o MediaWiki é que é uma ferramenta/projeto para auxiliar editores e usuários da wiki, porém ainda é confuso como se utiliza, e principalmente o que exatamente é a MediaWiki.

• Pergunta: Foi fácil seguir o guia de tradução? Como foi a sua experiência com a extensão Translate?

• Respostas:
  • QtK6z: Foi fácil, na verdade não precisei do tutorial. O uso das variáveis às vezes não é claro e a sugestão automática mostrou um texto com as variáveis erradas, que achei que era certo. Isso gera mais confusão e dúvidas do que ajuda, seria melhor não ter uma sugestão do que ter uma errada. Algumas traduções não foram salvas e nenhum código de erro ou mensagem útil foi mostrada para que eu entendesse o que aconteceu e como corrigir o problema. Apesar desses problemas, acho a interface muito intuitiva e agradável.

• • Viviannesh: No começo, naturalmente, houveram algumas dúvidas, porém o guia foi bem informativo (incentivador até). O vídeo facilitou muito também. Sinto que não explorei todas as ferramentas que a extensão tem a oferecer, e que se tivesse o feito a experiência de tradução seria melhor (não apenas a minha quanto a de outros usuários).

• Pergunta: O que te pareceu complicado no processo de tradução? Você teve alguma dificuldade? Se sim, essa dificuldade persiste ou acabou após alguma prática?

• Respostas:
  • QtK6z: Não saber quais convenções adotar é um problema, tenho que procurar outras ocorrências na mesma página antes de fazer uma tradução. Se não existir outra ocorrência na mesma página não sei em qual outra posso procurar. Nesse caso só escrevo o que acho correto assumindo o risco de divergir com outra tradução e talvez causar confusão no usuário. Outro problema é não saber se devo traduzir links para outras páginas que só existem em inglês, se devo só colocar o nome em português ou um aviso que o link leva para uma página em inglês. Todas as dificuldades persistem, apesar de ter decidido adotar soluções sem saber se são consenso na comunidade ou se vão prejudicar algum usuário.

• • Viviannesh: O posicionamento do texto. Ao ir traduzindo mensagem por mensagem, você não enxerga a página como um todo. Em legendas de imagem por exemplo, não fica informado que a mensagem se trata de uma legenda, e ao selecionar o texto acaba se traduzindo "às escuras". Essa dificuldade cessa numa revisão, porém acredito ser algo que pode ser reavaliado.

• Pergunta: Quais são as suas motivações para fazer tradução voluntária? O certificado influenciou a sua escolha em se candidatar?

• Respostas:
  • QtK6z: Sempre tive vontade de contribuir de alguma forma, não só com software livre mas para a humanidade. Até então não sabia de nenhuma forma suficientemente conveniente e dentro das minhas habilidades. Ao descobrir a facilidade de raduzir essas páginas da mediawiki decidi finalmente colocar esse plano em prática, considerando que sou usuário frequente dos produtos da wikimedia e concordo com os valores e a missão da organização. O certificado não influenciou minha escolha, mas se precisasse dele por alguma motivo teria influenciado fortemente.
  • Viviannesh: Eu já possuía interesse em tradução, e vejo o voluntariado como uma oportunidade de adquirir experiência. O certificado é um bônus.

• Pergunta: Em que aspectos o processo de tradução técnica pode melhorar? E que aspectos são dignos de elogios (e por quê)?
  • QtK6z: É necessário diminuir a fricção atenuando os problemas citados antes. A impressão passada pela interface do site é que é necessário ser um membro ativo na comunidade por algum tempo para fazer qualquer contribuição, por menor que seja pois existem muitas definições sem explicação e muitos passos para efetuar essa contribuição. A interface da extensão Translate é muito boa e me permite concentrar na tradução na maior parte do tempo, as sugestões automáticas também já foram úteis muitas vezes, seja para me poupar o esforço de realizar uma tradução, seja para me informar de outras traduções que frequentemente usadas se tornam convenções.
  • Viviannesh: Acredito que apenas a questão de posicionamento do texto. No geral achei bem intuitivo, o que é bom ao se tratar de algo para um público amplo. Traduções equivalentes, recorrentes e a "documentação" (embora não ache essa nomenclatura muito informativa) são ferramentas de grande apoio, principalmente considerando que a wiki é feita por diversos colaboradores. Com essas ferramentas o tradutor já tem referências para se guiar e tem a oportunidade de guiar outros usuários também.

• Comentários adicionais:
  • QtK6z: Domínio técnico, contexto, ler página, comunidade. Um problema quase inevitável é ter que saber o que uma expressão significa em um determinado contexto, o que faz que sempre tenha que estar com o artigo que estou traduzindo aberto, além de ter que ler partes do mesmo. De maneira geral possuo conhecimento sobre os temas tratados nos artigos, mas tenho certeza que uma parcela enorme da população que poderia contribuir não possui, o que pode dificultar a tradução pois pode ser difícil presumir o significado de um termo ou como ele é usualmente traduzido pelos usuários técnicos da sua língua nativa ou se sequer é traduzido ao invés de usado como no idioma tradicional. Também pode ser difícil para um tradutor sem conhecimento técnico notar que a sintaxe de uma linguagem de programação ou marcação deve ser escrita da mesma forma sempre e não traduzida. Não sei se existe uma solução para esses problemas, mas são obstáculos presentes na busca de novo voluntários e geração de traduções adequadas. Outro ponto, apesar de não ter procurado ativamente, não vi links claros que me mostram como encontrar ou interagir com a comunidade, nem me senti incentivado a procurar. A experiência de contribuição parece muito solitária e enquanto usuários como eu podem apreciar isso, saber quem são as pessoas que você está ajudando, onde pedir ajuda e receber feedback de membros mais experientes ou simplesmente compartilhar opiniões sobre o projeto ou jogar conversa fora pode fazer o processo de contribuição mais humano e significativo, ao invés de só parecer que você está fazendo trabalho repetitivo para um máquina gigante sem rosto ou personalidade.

Loose thoughts

As I do some research about the Wikimedia movement and read the meta:Strategy page, I'm thinking about two things that I experienced constantly in this internship:

• Information is so fragmented I have to go down the rabbit hole and spend some hours looking for relevant information because most of the time, searching for important content don't follow a predictable pattern and sometimes it's difficult to say if a page is still relevant regardless of the last time it was edited.
• You have so much going on in the movement sometimes you end up not knowing about a relevant initiative. Most recent example: MediaWiki Documentation Day 2017. Thank you, person that mentioned it on my Wikimedia blog post. It's definitely interesting to me to come across efforts to improve documentation here. I'll note that on my final report.