If I had to pick only one option, that would be by stopping making a mess such as tens of coins, merged mining confusion and such. Keep it simple and neat. In its current state, even to computer-aware people BTC looks fishy (not stating my opinion there, just reporting some opinions I got from people who were introduced to the software on some forums like half a year ago). I leave up to your imagination what they could think of R0xorCoin-SupaBrix merged mining...

I think you pretty much summed it all.

Even with that, it's hard to compare IMO: when it started, BTC didn't have some large cryptocurrency user base to pick in. And it didn't face a (kind of) well-established concurrent.

Cf post #25 above: https://bitcointalk.org/index.php?topic=55378.msg660413#msg660413
Even though actually we can have pretty much every combination such as (original value of A + original value of B) remains unchanged, it seems more likely that we'll reach one of the extreme, ie either both tending to the same value (the average) or one seriously dominating the other. It's not like fiat currencies, where every currency has a captive user base (the people who live in the currency's country), here people can pick whichever currency they want. So we're more likely to reach an equilibrium such as 50-50 or 0-100 than some other random combination such as (population of country 1 - population of country 2).

But anyway, we're starting with different (and even somewhat opposing) hypotheses so it's not surprising we come to different conclusions. And those are just theories and the only way to be sure is to actually implement the mayhem. I'm focusing on BTC anyway, so if you happen to screw LTC this way, well, be it... Maybe LTC will win by 100-0, after all

It's not a matter of your personal choice, it's a matter of what is done on the global scale. But as I explained, the coin values would tend to get averaged, so merged-mining BTC and NMC would rather decrease the value of BTC and increase the value of NMC. Although as I explained too the popularity difference was so huge to begin with (and the merged mining so marginally supported - ie notably not by the biggest BTC pool) that this effect remained negligible. Probably when the difference is that huge at the beginning we could even get the opposite effect: the weakest currency collapsing because most of the miners just sell it as they mine it to convert it into the other coin.
But anyway, the value from mining both didn't raise above the previous value of mining only BTC... except maybe if you consider the fact that the global hashing power diminished quite considerably since merged mining was introduced (which goes against the argument of merged mining = more hashing power).

Does it work if I only have 0.01BTC?

Meh, real men use Wuala. They even accept Bitcoin payments, btw.

(sorry otherwise, already got a Dropbox account, even though I don't use it much)

Ok, I see where we have different views then: you assume that the value is all in the ultimate protection against this overrated 50%+1 attack, while I assume the value is mainly in the use/demand for the currency (which would in turn likely increase mining rate and thus security, +/- proportionally to the interest of actually performing the attack).
I'd just like to point out that the mining power in TBX and FBX is marginal compared to the one of LTC, so unfortunately merged mining wouldn't increase LTC's security a lot.

Ok, I forgot to mention in my analogy that when I fill my 2 fuel tanks, I only pay to fill the first one while the second one if free. Please, stop assuming that all people criticizing merged mining don't understand how it works. It prevents you from getting the bigger picture, ie leaving aside the technical details to start thinking about what the PRACTICAL impact could be.

The practical impact is, if you generate one coin A AND one coin B instead of generating only one coin A OR one coin B (from the same effort), assuming that the global value you are generating is the same because it's a function of the cost to generate it (that seems like a fair assumption, although I'm aware it's not certain and probably not 100% accurate either), then the value of coin A and the value of coin B are likely to be averaged, ie to tend toward (original value of A + original value of B)/2. That's not a decrease yet, but when we add to this the fact that mining becomes, say, twice as difficult on average (in the case where there was an even mining power for A and B), the result of mining is (original value of A + original value of B)/2/2*2 (the last multiplier is because we now mine 2 coins instead of one)=
- the same as before, if coin A and coin B had the same value/diff ratio at the beginning,
- or lower for someone who was mining the coin with the best value/diff ratio
- or higher for someone who was mining the lowest value/diff ratio
So, the only winners in the process are those who mined the coin with a lower value/diff ratio, aka probably just a few gamblers, or a few people who didn't do the math.

Again, I'm not saying merged mining would directly decrease the global generated value. Just that it will split it over several ccurencies, rendering it less PRACTICAL to use, which in turn has a chance to actually decrease the generated value (because not practical = not wanted).
For BTC vs NMC, it seems like it didn't occur (yet, at least) because BTC has so much more support behind it (I don't mean customer support there , just users and even merchants) that the coin value was indeed unlikely get averaged. For LTC vs TBX/FBX, which all have like zero merchants, well, I wouldn't be so sure. Since the only ccurrencies/users that can really benefit from merged mining are those with sub-average value/diff ratios, well, you can guess who would make a good deal (and who wouldn't) by merging the mining for LTC-TBX-FBX.
Of course, this is a very over-simplistic economical model, but that's the risk. If I'm already on the best ccurrency, I have no reason to want it to be merged with junk ccurencies. But if it does get merged, of course my best option then is to follow the move and merged-mine.

Not sure... it just seems obvious, actually. Same global value for 3 currencies, requiring 3 clients, 3 balances to handle, etc. It's a bit as if your car ran on 2 different fuels, in 2 separate fuel tanks (that you could refuel at the same time, though, as in merged mining), and to turn right (like, to buy from merchant 1) you need to use fuel A but to turn left (like, to buy from merchant 2) you need to use fuel B. This doesn't look that terrible, but still it feels like added complexity with no added value.

Maybe by a few fractions of a percent of a percent, yes. Although I'm not sure people who use a CPU to mine LTC would use this CPU to mine BTC if no CPU chain was available.

I definitely don't think that 1 more = great. But since it's there I jump on the boat "just in case". I have no plan to ever accept any other ccurrency than BTC as a mean of payment, though.

Sure, if it's implemented I'll use it (unless there's no decent pool to do so like is the case for BTC and NMC). But this will likely be harmful to LTC because it will obviously dilute it's value. There's a bit of the same problem with NMC diluting BTC's value, but because BTC's lead is very clear it's not too much of an issue. I'm not sure LTC's lead is that clear over TBX and FBX. But I guess we'll see, if merged mining happens to get implemented...

Because it seems to be the only credible CPU chain at the moment. Clearly it has a lower priority to me than BTC mining (e.g., when there was still the Catalyst bug, I gave one full CPU core to each of my GPUs), but when the CPU isn't doing anything else, well, why not?

As I said, "from the number of people actually mining TBX and FBX at the moment, I'm not sure the diff would really increase much." Sure, the difficulty for TBX and FBX would increase a lot, but what matters is LTC.

I'm sorry, did I mention anything about control? Control isn't the only possible objective of manipulation by fear. Giving some value to some 7M premined stash, on the other hand, could be one.

Me neither, but it keeps the CPU busy and the house warm, so... 

You can't really compare cryptocurrencies to real atoms with a real intrinsic value. But anyway, if one can extract both copper and gold from the same mining effort, it does increase the rate at which both are mined, because one can suppose that people mining only gold will likely switch to mining both at once, and same for people mining only copper.

If everyone switches to merged mining, all those merged-mined chains will end up with the same diff. So it is really like mining 50 LTC + x TBX + y FBX from the same effort as it would have been to mine just 50 LTC or just x TBX or just y FBX. So it is really like mining one of them and getting the other 2 free.

That's splitting the value of a chain into 3. Of course, maybe you think a LTC being worth 0.002 BTC is too much and that making that 0.0006 would be better?
That's also reviving dead currencies for the profit of the few people who stashed a lot of them (not to mention the 7M TBX premine) and will thus find them gaining some value again.
Finally, that's opening the road to merged-mining of, say, 50 meaningless cryptocurrencies (if we merged-mine 3 currencies, why not add some other new ones to the party?). BTC is already not getting much of a merchant adoption, I'm not sure any merchant would bother dealing with 50 different and weak cryptocurrencies, they already most often don't bother with several real currencies...

From the number of people actually mining TBX and FBX at the moment, I'm not sure the diff would really increase much. Also, did it ever occur to you that this virtual attacker could do merged attacking, too?
I like how we see so often the same technique as politicians use: spray fear to justify anything. "Booo, beware of the naughty attackers, we should add trusted nodes to check-in every other block." Reminds you of something?

Thanks, the LTC address in my sig it is, then

Barely 5% improvement on my AMD laptop but indeed about 40% better on my Intel server. Pretty neat

Oh, then yes that's because of inherent problem with the calculation. Or also maybe just because noone ported the project to GPU It's also funny to see how some BOINC projects will work only on NVIDIA or only on AMD GPUs.

If you're referring to LTC, they don't view this as a handicap. It's on purpose because they think that it helps more casual people (who usually don't have big GPUs) become miners too. And probably also because a new GPU cryptocurrency would have to face the competition of BTC.

The point of Folding@Home and alikes is to compute as fast as possible, so I wouldn't say they can be "GPU-hostile" as in "sabotaged not to work in GPUs", but if you look in BOINC you'll find some projects like QMC@Home that work only on CPUs.

The good way to do that is to adjust the difficulty. Not to say "be there at this precise time for your chance to get much more coins than you should". The difficulty was maladjusted on purpose.

Fixed typo

Con: first-48h adopters were able to stash in the 10k LTCs already. Check out how long it would take now with a standard quad core CPU.