They have a nice little Java indexing tool that shows the page and lets me enter the data. I've transcribed 2501 records so far, which feels like a major accomplishment. One thing I would like is to be able to pick which pages I index; it would be a favour to me because I could find and index the pages I need for my own research, but it would also be advantageous to them, because I'm more familiar with the names in the areas I've been doing research in, and less likely to interpret the old handwriting incorrectly.
I like to think that I do a particularly good job at this because I've worked at deciphering old handwriting before, and because I use other resources to cross-check names. The rule is always to transcribe what's there, but there are many times when it's hard to see exactly what's there. An "n" can look a lot like a "u", and a double "r" can look like an "m" in these old records. Because the births I've been indexing have fallen between 1898 and 1902, I use the 1901 census heavily to search for the parents and make sure that the names I'm indexing make sense.
These Ontario birth records are key to genealogy researchers because they record the mother's maiden name. The female lines are always harder to trace, and that one last name can unlock a whole other family. Other interesting information that is not being indexed is the residence and occupation of the father. For rural families it's not helpful, but if they lived in town, the actual street address will be given.