I was listening to FLOSS 127 with Randal Schwartz and Randi Harper aka FreeBSD Girl. They had on Guillermo Amaral famous not only for his contributions to PC-BSD and KDE contributions but also his work in promoting Open Source in Mexico.

The discussion centered around a variety of topics related to promoting the use of Open Source Software in Mexico, including topics around Open Source vs Piracy, use of OSS in government and university, and lastly the slow adoption rate of OSS in Mexico because of the language barrier. There isn't enough support for Mexican Spanish in open source projects. At best they will get a translation in to "Spanish for Spaniards".

I think I have the solution for this, and it's crowd sourcing of the localization, and I'm ready to partner with a project to see if it works.

Piracy
Guillermo shared with us that the cost of pirated software in Mexico is so cheap that you can get a fully functional illegal copy of Windows for a few pesos (Say roughly $1 USD). Because of this, their is not such a clear cost save of using "Free Software" when you can get software developed by a company like Microsoft that makes the investment in localization for their commercial product line. Buy a few legitimate copies and you can leverage the support contract to get support when you have issues with your dozens of illegal copies. Support issues are compounded by language barriers. If you reach out to someone for support and you have communications issues you will be driven back to proprietary software. In this case I think we have to work with the community to make sure that we have localized ongoing support. Below you'll see that it's not the only issue related to communication and language barriers.
Corruption
When it comes to government and university another problem is corruption. Software vendors will come in and give steep discounts to get whatever money they can from a market that is so prone to theft. Corrupt officials would claim to pay higher fees, comparable to full price or even the equivalent price of purchasing it in America. I can't substantiate any of this, but Guillermo's arguments make a lot of sense and I could see it happening. I think there might be a place here for OSS that is commercially supported, but it seems like a horrible contortion of Open Source.
Language Barrier
The third issue is one that I feel we can do something about! The slow adoption of Open Source Software in Mexico due to the language barrier is a much more manageable project. There are actually a couple of issues here. Unlike some countries which really promote English in the school system, while Mexico does teach English in the schools, it's not pushed as strongly and definitely doesn't leave someone qualified to use anything beyond the most basic software. A real barrier to entry for the general population. The second issue is that when you do get a translation into Spanish, it is often targeted at people from Spain and not people from Mexico. Anyone English speaker will be familiar with at least the dialect gap between American English, British English, and Australian English.

So how do handle localization better
I have been a developer for a number of years and I know it's really a challenge to think about localization. It requires you to work with a platform that cleanly provides support for these problems. As hard as it is to believe, there are actually major platforms that only recently brought this to the table. Once you have that support you need to design your application to support multiple languages. The techniques you use can be as simple as an include file for each language, or a more specific technology driven solution like a "Resource Bundle".

Assuming you're a sophisticated project who understands the importance of factoring out the localization components, you have choices. At a minimum you can break the work out and divide it amongst multiple contributors. Each can have their own responsibilities and take ownership, have a stronger sense of contribution.

Next Level: Crowd Sourcing
I won't say it would be trivial to crowd source this, but I do have a solution modeled that I believe in. By crowd sourcing this, you could distribute the work not across a few people, but across any of the contributors. Beyond that you could source it across the entire community, even users. Anyone who has a vested use in the project and the capabilities of doing the translation could participate. By building a "recaptcha" style model you could even do verification. You could allow people to see three things and approve them and translate one. There are a lot of options here.

I believe this model will work so much that I am prepared to set up a system to "webify" the localization effort of an open source project* that wants to collaborate on this. The only requirement is that they have a mechanism to factor out the language configuration such that I can bridge the gap with a webpage. I will put together a system to display the contents, capture delta, and regenerate the configuration files. If a couple of these projects are successful, I'll turn it into an open source project.

If interested, contact me:
smtp:admin@osswatch.org
voice:(330) 474-9852
http://twitter.com/osswatch_org

*Giant Disclaimer: Since I'm paying for all the bandwidth out of my pocket, I may have to pull the plug if it's too successful, but I'll turn over any code generated as part of the project.