On Sat, Apr 8, 2017, at 20:27, Tom Harding via bitcoin-dev wrote:


On Apr 7, 2017 12:42, "Gregory Maxwell" <greg@xiph.org> wrote:
On Fri, Apr 7, 2017 at 6:52 PM, Tom Harding via bitcoin-dev
> A network in which many nodes maintain a transaction index also enables a
> class of light node applications that ask peers to prove existence and
> spentness of TXO's.

Only with the additional commitment structure such as those proposed
by Peter Todd in his stxo/txo commitment designs, e.g.
Light nodes are improved by detecting invalid transactions, even before they are mined.
_______________________________________________

I am not quite sure why you think this approach would help in this regard. I may be missing part of how Core works here, but Bitcrust's txindex is merely used to lookup transactions from hashes and currently, and seems to fulfil the same role as Core's -txindex mode. 

This can be pruned, and in the future auto-pruned as the "flat files" used as base for all data allow for concurrent pruning. But unlike Core, it is always needed as without UTXO index, it is needed to find outputs during base load validation.