Archive for June, 2011

New Feature: Manual File Upload/Download

June 8th, 2011 by Dario Solera | No Comments | Filed in Amanuens

Stand-alone File Repository

Starting from today, Amanuens allows you to manually upload and download files without the need to connect to your source code repository.

Using the Stand-alone File Repository, as we call it, is very easy. Just launch the new Project Creation Wizard from your account Dashboard, and select “No” when it asks if you want to connect to a source code repository.

Project Wizard

The new Stand-alone File Repository allows you to:

  • upload and download individual resource files
  • upload and download resource files in ZIP archives
  • email a ZIP archive containing all the files in the repository to a recipient of your choice (or to yourself)
  • easily manage the repository structure, ensuring that all files are always named correctly (file naming and directory structure are enforced).

Apart from being managed by Amanuens, the repository works 100% like a regular source code repository. Actually, it’s a plain Subversion 1.6 repo, so we might even make more versioning and diffing features available in the future.

We think that the Stand-alone File Repository is very useful when you want to try Amanuens without having to configure a connection to your repository, but still getting 100% of the platform functionality and workflow management. We’re still convinced that connecting to your repo saves you lots of overhead, but after all zipping all your resource files and uploading them isn’t that complex either.

What do you think about this new feature?

What Will Change After Google Translate API Shutdown

June 1st, 2011 by Dario Solera | 3 Comments | Filed in Amanuens

A few days ago Google announced that the Translate API will be shutdown this December. All hell broke loose.

My take on this decision is that it will change exactly nothing, and the reason is that Google Translate is quite poor in terms of translation quality. Since when we integrated Amanuens with the Translate API, I started getting quite upset at the quality of translations. I’ll make an example. The source text is:

Warning: this Page is being edited by another user

Google Translate comes up with (in Italian):

Attenzione: questa pagina viene modificato da un altro utente

Not only the translation sounds weird in Italian, but it’s even wrong as “modificato” is masculine, while “pagina” (page) is feminine.

Sure, Google Translate is great if you have a piece of text in a language you don’t understand and you need to get a grasp of its meaning, but using it for translating text for production is just plain dumb. I also believe that using Google Translate to get a first draft of a translation is not a good idea because, oftentimes, it’s harder to fix an existing translation rather than writing it from scratch. Taking the example above, the translator should just fix one word, assuming she spots the error, (“modificato” -> “modificata”) – yet the sentence still sounds weird and should be rewritten from scratch.

So, for very short sentences, Google Translate might work well (e.g. “Save document”), but for those we already have pretty powerful translation memory technologies. So, Translate becomes useless even in this case.

For all these reasons, although shutting the API down is a bit lame, I think that that decision will not affect the translation and localization industry at all. Possibly, it will give us back the feeling that professional translators are extremely valuable, even for often-undervalued software localization jobs.