Last updated on: May 21, 2015 by Musa Kurhula Baloyi
10,000 Dictionary Entries
It took me less than 2 weeks to clean the data that now forms part of the more than 10,000-word online dictionary. This is besides the work I had to do in scanning the hardcopy dictionary and then converting from PDF to text file. In comparison, the Second Edition of the Oxford English Dictionary contains around 231,000 words [OED.com , Oxford Dictionaries ], while the Urban Dictionary contains approximately 6 million definitions and 2.1 million unique words and phrases. It would be hard to know how many words in total exist in the Xitsonga language or for all African languages combined, but for an estimated 2,000 languages, there should be around a billion. In a way, 10,000 is an achievement, but still scratching the surface in terms of the possibilities that exist.
As I continued to populate the database with more data, I received another notification a little while later:
Data entry was made easier by writing a uniform SQL script and then running that against a hobbyist Heroku Postgres database. Each new word is a new entry. The same spelling gets one entry but will be formatted better on the display for cosmetic purposes. Different spellings for the same word are nonexistent in the standardised version of Xitsonga. But this will become a factor once we open the dictionary up for comments and the contributions start rolling in. Now I am sitting on an interesting problem: what is the most cost-effective and efficient way to host what will eventually become millions, perhaps billions, of word/meaning pairs and their metadata?
During this time, I learnt that
10,000 Facebook Likes
It took us 4 years (2010 - 2014) to to hit the 10,000 mark on our Facebook page, the First Online Tsonga Dictionary . The numbers have dwindled a little bit since Facebook has not been a major focus since the meetup group, and the move away from a Xitsonga-only drive. Those 4 years taught me a lot about how people interact with languages and a service that seeks to enhance conversations about African languages.
Some of the learnings are:
At this point, we need to have new languages added, but the focus should be on adding new functionality to improve the user experience. There is an existing copy of the Xilenge dictionary, from the incorporation of which we can gain some major insights on dealing with multiple languages. This is an important precursor for the continent.
The current challenge is to get users to engage just as they did when controversial stuff was posted on the Facebook page. I know not how. Up until now, the Tsonga vs Shangaan debate had taken over the Tsonga Internet to a point where we had to call time on it. We seek similar engagement on other topics. The users have so far helped bring out the core meaning of many words in a way that made them intuitive. We need them to do the same on this new website.
These two metrics (the number of Facebook likes and the number of words in the database) have been instrumental thus far. But, how do we measure ourselves moving forward? By 10,000 daily searches? $10,000 in funding? 10,000 new sign ups? 10,000 downloads? Whatever the next metric to evaluate ourselves, it has to have a bigger impact than the other two combined.