As I understand sidechains there's a complicated proof mechanism that gets set up for transfer of bitcoin 'value tokens' between the bitcoin blockchain and alternative blockchains without creating any additional currency units, and a fall back to a much simpler exchange mechanism for actual practical day to day exchange.
But the whole thing about not creating additional currency units actually seems fairly arbitrary to me, given that this adds a whole bunch of otherwise unnecessary complexity.
The fallback exchange mechanism for sidechains is based on an atomic exchange algorithm that is well known for many years now.
This is something that could easily be standardised in the form of a relatively simple 'pay on reveal secret' transaction type to permit decentralised exchange between arbitrary pairs of blockchains, as I discussed in the following blog post: http://upcoder.com/11/atomic-cross-chain-exchange/
Neat proposal. It's worth mentioning that it can't be implemented in the current bitcoin implementation, yet the alpha sidechain introduced here happens to add exactly the bits it requires to work, namely time locked transactions and a more plausible malleability fix.
Which means if you find/build another altcoin with similar capabilities, you can start exchanging testnet coins with those altcoins using the very approach you describe today.
I don't believe that's a happy accident, but rather one of those core reasons for getting sidechains off the ground, as it provides a clear path toward finally building all the insane stuff people have been dreaming about for years.
It feels like there are much simpler ways to support pay on reveal secret directly on the bitcoin blockchain, if this is accepted as an important goal (e.g. with explicit conditional redemption conditions based on spending transaction blockcount).
But I guess a more general solution for time locking and malleability can give you other stuff as well (e.g. off chain payment microchannels).
But the whole thing about not creating additional currency units actually seems fairly arbitrary to me, given that this adds a whole bunch of otherwise unnecessary complexity.
The fallback exchange mechanism for sidechains is based on an atomic exchange algorithm that is well known for many years now.
This is something that could easily be standardised in the form of a relatively simple 'pay on reveal secret' transaction type to permit decentralised exchange between arbitrary pairs of blockchains, as I discussed in the following blog post: http://upcoder.com/11/atomic-cross-chain-exchange/