I am in the process of updating my ranking program, so that there is less 'human' input. So, for a start, the start rating for each team will now be calculated by the program, instead of being a kind of educated guess. In order for a team to be ranked, it must play at least 2 matches against teams who already have rankings. This is why, for example, Christmas Island, Cocos Islands, Easter Island, etc have disappeared from the rankings. Also, most of the Indian states have disappeared because they mainly play among each other, but Sikkim play against the likes of Bhutan and Tibet, so they have managed to get themselves a ranking, and so any Indian state who then plays Sikkim twice can then enter the ranking (Chandigarh, for example, have done this, so now every team who plays Chandigarh twice from now on can enter the ranking) and so on. I think this is the only way to get a meaningfully comparative ranking - it is easy to get a comparative ranking AMONG Indian states, but this does not tell us where they really stand on a global level.
So, both parts of the ranking procedure are now almost fully automatic - the form-based side and the ELO-based (more historical) side. Start rankings are given only to the first two teams to play (England and Scotland) and all other rankings are calculated based on only this input. So, until a team plays England or Scotland, it cannot be ranked, meaning a lot of Austria v Hungary matches etc. have no effect on the early rankings.
I believe the end result will remove a lot of the subjectivity involved in ranking systems by making things a bit more automatic.
After I have fixed my archive, a lot of Australian states and perhaps other teams will enter the ranking, it's just I haven't had time to add these to the ranking program yet, so in effect, the ranking program is using data from the old archive.