Quote from: starik69 on September 06, 2014, 09:57:59 PM

Quote from: Iffy on September 06, 2014, 08:58:22 PM

You clearly do not understand the Block Tree design that Dan has explained numerous times over the past year+. For example "fork problems?"; Block Trees are designed to fork, it makes the system far more secure and will lead to a lighter wallet resource wise. Less data to process and store makes syncing less of an issue than you seem to think it is.
Yea tests have never been of any use to indicate real world performance for anything. o.0

Syncing was great problem in previous open betas just after they were "tested" on 25 nodes. This is fact. With all that Trees design. So your argument is irrelevant. Tongue

Btw, has Dan disprove cap theorem? Grin

I just emphasized the operative part of your post. Yes there "was" a problem to start, this has been resolved during testing and development.

Firstly no, Dan has not disproved CAP theorem. The use of a block tree means that guaranteeing Consistency (all nodes see the same data at the same time) is not required. So long as a portion of nodes see a portion of the same data at the same time all is good.

Secondly, block tree's allow HIGH Partition tolerance because a block tree is allowed to be partitioned, (or forked in lay terms).

Hence the only guarantee we need to provide as per CAP theorem is availability.

http://en.wikipedia.org/wiki/CAP_theorem