Dishwasher localization

Originally posted to Shawn Hargreaves Blog on MSDN, Friday, October 10, 2008

The latest post from James Silva discusses the drudgeful process of localizing Dishwasher.

This triggered a range of incoherent and rambling thoughts, which for some reason I am inspired to post here.

• It is amazing how much work can be involved in finishing a game, even one like Dishwasher that was already looking great a year ago!

• I'm reminded of the old joke that the first 90% of a game takes 90% of the development time, and then the remaining 10% takes the other 90% of the time.

• While we are on the subject of boring, unglamorous tasks that prepare a game for commercial release, check out our newly released Best Practices guide.

• I wonder if things will turn out any different for Community Games, where these best practices are optional, compared to XBLA, where James has to pass a much longer list of certification requirements before he can ship? ie. how many of the TCRs are really important for creating a polished game that people will want to buy, versus how many will turn out to be unnecessary?

• On the topic of localization, I recently finished a sample that shows how to read translated strings from .resx files, and how to build fonts including all those crazy European accented vowels, Japanese kanji, etc. Stay tuned...

• I feel James' pain when every little tweak to the game requires a laborious retranslation of the modified strings. Brings back memories of a few hours before the final build of MotoGP:

"Dang, this is a nasty bug!"

"Easy to fix though."

"But if we fix it, we'll have to change some text strings, and we have no time to get new translations back."

"Anyone know anyone who speaks German?"

"Nope."

"Hmm, I wonder..."

I'm embarrassed to admit, we actually shipped the game with a couple of error messages that were translated using Babelfish  :-)