Methods for protecting Alt-chains from attack

[crazy_rabbit]

This is a general question about ways to protect Alt-chains from attack. I'm not worried about the implications of pseudo-central control or checkpoints or someone in control, trust, etc....

I'm more interested in practical and simple solutions that allow people to be able to experiment and learn on parallel crypto-currencies, (knowledge that they can reapply later on real Bitcoin projects) without beginners or novice programmers having to worry immedtialy about protection from being attacked by those who would do so just because they can.

Any ideas? I see Terracoin has checkpoints regularly added, but does this offer adequate protection? I'm talking about enough protection to pick up from an attack and keep on working, not necessarily about so much protection that the 'value' isn't affected, because I not yet thinking about Terracoin as a real live money exchange, but rather like trading marbles. So it doesn't have to be Bitcoin Level security. Just enough so that people don't feel like all their work is just wasted for nothing.

[deepceleron]

I tried to address that here:
https://bitcointalk.org/index.php?topic=121845.0

Get the blockchain merge-mined by a lot of existing Bitcoin pool hashpower before it becomes active or generates coins. Otherwise there will always be some entity able to attack the working coin.

One would need to convince a large percentage of exiting miners that the new blockchain is a worthwhile concept first.

[crazy_rabbit]

I tried to address that here:
https://bitcointalk.org/index.php?topic=121845.0

Get the blockchain merge-mined by a lot of existing Bitcoin pool hashpower before it becomes active or generates coins. Otherwise there will always be some entity able to attack the working coin.

One would need to convince a large percentage of exiting miners that the new blockchain is a worthwhile concept first.

That is a good idea, although seems like a hard sell if the new project is just a sandbox basically. :-)

[pyra-proxy]

One idea I have seen thrown about (and I should dig up that strange white-paper I saw that referenced it too but I'm too lazy atm) was an implementation of hashing which used 2 or more hashing algorithms.  I remember reading about it first related to litecoin to solve some problem they were having but it never got implemented.  The idea being that (in the case of say 2 hash algo's) the chain would mine odd blocks bitcoin style and even blocks litecoin style (although this might need to be beefed up a bit to bring the memory requirements back into cpu territory).  What this would seem to do is prevent a large gpu/fpga/asic pool from finding consecutive blocks and also prevent botnets from finding consecutive blocks, once the 2 hashing sources both had ample hashing power.

The implementation I would think would be relatively simple and needing mostly to coding around handling of 2+ difficulty values.  And while presumably the bitcoin style hashers could find blocks in even numbered blocks and litecoin style hashers could find blocks in the odd numbered blocks they would both be considerably more inefficient at it vs. the optimized correct block hashers and it would be an uphill battle to find the consecutive blocks necessary to perform 51% style attacks.  I don't know what the new % attack would be but my uneducated guess would be that it should quickly rise to needing 75 or 80%+ hash rate from a single source OR 51% bitcoin style hashing AND 51% litecoin style hashing simultaneously.

This would also possibly provide interesting merge mining opportunities in that in theory one could merge mine this chain with either bitcoin or litecoin and optimize their mining to only look at the more efficient block mining that they have merged with or try hammering through finding any block even the less optimized versions.

Seemed interesting enough at least... throw some PoS in there as well for an even more dynamic system which makes it that much more difficult to find subsequent blocks and 51% begins to become a very difficult to accomplish attack vector.

[mskwik]

I tried to address that here:
https://bitcointalk.org/index.php?topic=121845.0

Get the blockchain merge-mined by a lot of existing Bitcoin pool hashpower before it becomes active or generates coins. Otherwise there will always be some entity able to attack the working coin.

One would need to convince a large percentage of exiting miners that the new blockchain is a worthwhile concept first.

It is a much easier sell though if it can be merge-mined alongside bitcoin rather than taking hashing power away from somewhere else.  Just look at the current difficulties on Namecoin/Devcoin/ixcoin etc, despite being worth fairly minimal amounts they have a good amount of hash power pointed at them (making them that much harder to attack) just because the extra effort to merge-mine them is also minimal.