Questions about providing translations/foreign language editions of DH projects.

Digital Humanities Questions & Answers » Interfaces, Design & Usability

Questions about providing translations/foreign language editions of DH projects.

(5 posts) (4 voices)

Asked 4 years ago by Karin Dalziel
Latest answer from Arno Bosse

Tags:

Karin Dalziel
Member
This is mostly just a request for anyone with experience providing translations of DH projects to share what they learned.

Some general questions:

-How do the translations affect the workflow? i.e. if a change comes in for page (usually in English, in our case), what's the process to roll those changes out to the translated pages, especially if the original translator is no longer available?

-How do you choose the languages to translate into?

-What content do you translate? For instance, it is pretty easy to provide translations for the introductory pages, but if there is a lot of content it may be harder to provide translations for all of the content immediately. Is it worth rolling out translations for introductory pages before the translations for the content are complete?

I'm just starting research on this topic (beginning with the W3C's Internationalization Best Practices: Specifying Language in XHTML & HTML Content) but would love to hear other's experiences. This isn't a question as much about the technical implementation, though I'd love to hear about that too if you want to share. :)
Tweet this question
Posted 4 years ago Permalink
Stéfan Sinclair
Admin

This topic seems a bit like walking on eggshells, especially since their are some very adamant opinions out there. My experience is that translation is somewhat of a losing battle, but may still be well worth the trouble, especially when trying to reach certain communities. In an ideal scenario translation responsibilities fall back on people within the community (rather than assuming responsibility centrally in a project).

Way back when the digitalhumanities.org site was launched we created translations in French and Italian – both were done with modest stipends. The site then was powered by a wiki and there was a mechanism to notify editors when a change was made to a translation page, but over time the translated content fell out of sync. That was using TWiki – I'd be curious to see if MediaWiki has more sophisticated support, but I suspect not (several Wikipedia pages are not translations at all, but entirely autonomous articles).

For other projects I've built custom localization modules, partly because I find it easier to keep all translations together in a common resource (as opposed to one file per set of language strings). But that approach doesn't scale and only works in a small development team.

I guess for most of the multilingual projects I've been involved in the translations seem like a gesture more than a necessity, and frankly I'm not entirely convinced that the gesture has been worth the trouble. It's like the DH conference call for proposals that exists in several languages, which is a worthy nod to multiculturalism, but ultimately it's unlikely that someone would propose something without having functional reading competencies in English.

It may not be a choice in your case, but a generic piece of advice would be to do translation on a concrete need basis only, not on an anticipated need basis (this might be akin to agile programming where a client expresses a need rather than the developer anticipating a use). If translation is in the cards, then hopefully there's an engaged group of people who can sustain it and avoid linguistic atrophy over time.

<annotation>hope I'm not sounding cynical</annotation>

<annotation>this probably didn't even answer your question</annotation>

Posted 4 years ago Permalink
Karin Dalziel
Member

Replying to @Stéfan Sinclair's post:

Thanks for your response! I'm at the beginning part of implementing these translations, but am already starting to see problems:

- People don't by default give me new text in foreign languages.
- When changes are sent for a page, I have to ask for corresponding changes for the foreign language pages. Sometimes I can make an educated guess, but that way lies madness (and laughably bad translations.)

My usability alarms go off when I am asked to make a "foreign language" edition of a site where only the intro text is in a foreign language, but I can't find any articles that have actually tested this scenario or done informational interviews with users of a site outside the primary language.

I would love to hear if any foreign language users find this kind of limited translation useful, or, if not, what would be useful.

My co-worker proposed an alternative for sites that only have limited content in a foreign language: make a section of the site that explicitly details what resources *are* available in each language. There are questions of how to direct users to that area of the site, but that approach seems much more usable and sustainable.

Posted 4 years ago Permalink
amandafrenchphd@gmail.com
Member

For what it's worth, I made thatcamp.org multilingual (though I'm sure it's inelegant) by installing a WordPress translation plugin. I think it's this one: http://wordpress.org/extend/plugins/global-translator/ -- I can check if you like. It makes a "Translate" link appear on every page.

No, we haven't tested it, really, though a couple of our friends in Italy and France were appreciative on Twitter. And of course it only works on WordPress sites, though I'm sure there's something similar for Drupal.

Posted 4 years ago Permalink
Arno Bosse
Member

I've just finished reviewing the user's manual for a German text editor/annotation tool. Something I realized immediately was that it's fairly counterproductive to have non-native English speakers (say, the project's developers) create an initial "literal" translation and then request a native speaker to add some final "polish" (no, not Polish!) to the final draft. Untangling the resulting mangle of embedded idioms is extremely time consuming and fraught with compromises. You'd be much better off re-writing the manual from scratch. The better approach, I feel, is to develop a well thought out style manual and high-level concept for the translation, ask the developers to deliver a combination of terse, bullet-point style lists and longer explanations where necessary describing the functions and then have the translation written by a native speaker. And based at least on my experience it would be wise to create a variant of the style manual for each of your target languages.

Posted 4 years ago Permalink

RSS feed for this topic

Reply

You must log in to post.