A Django site.
September 11, 2008
» Rapunga Google

One of the weaknesses I recognize in my own research (and that of others) is that I often find myself studying the artifacts of online communities rather than learning anything about the people who created and make up these communities. I have described this as taking an “Archaeological” approach rather than an “Anthropological” approach. By adopting a more anthropological approach perhaps we can stop producing interesting statistics and start to understand peoples language behavior and how technological artifacts affect this behavior.

TangataWhenua.com has a nice feature article giving some of the back story to the Google Maori project. It doesn’t really tell us anything about the technical or linguistic issues faced, but paints a nice picture of some of the people and processes (and problems!) involved.

July 24, 2008
» Te Wiki o te Reo Maori

Well it appears to be an interesting time for te reo Māori and the internet with the launch of Google Aotearoa as reported in BizReport and excitement around the possibility of reo Māori internet domain names according to a press release from the New Zealand Māori Internet Society reported in Scoop.

However as I seem to habitually see the glass as (less than) half full…

...I suspect that Google Aotearoa is just a skin and this concerns me. Now I know there are many people (and people’s whose opinions I respect greatly) who maintain that anything is better than nothing as far as minority language provision goes. With some reservations about quality, I generally agree with this position.

BUT, I can’t help feeling that search engines are different, and that a minority language skin over an English language search engine is fundamentally misleading minority language speaking users. To my mind (and perhaps I am wrong) if a naive user finds a Google skin in their language, it is natural for them to assume that Google will be searching in that language – making plurals, stemming, mutating and so on. OK the more sophisticated user may realise that Google only searches in some languages, but I suspect that the average user won’t. Given that Google may also be a naive users main (or only) way of searching for information on the internet, the fact that the results returned may only be a subset of the results that would have been returned if the search algorithm had been tailored to their language, may be significant. They may miss important relevant results, they may perceive the presence of their language on the internet as being smaller and perhaps less useful than it really is. This may actually reduce the status and use of the language, rather than enhance it.

If google.co.uk was actually using French language search algorithms, I would want to know about it. Perhaps users of Google’s language skins would also like to know.