The encoding is very simple but recovery in part depends upon prior knowledge of what might be encoded in a message. When you examine the reasoning of people who look at this with no assumptions about prior knowledge though (and in the case of Callimahos, someone who regularly dealt with material without tremendous prior knowledge) I don't know that that holds up particularly well. How can you tell whether the recovered graphical symbols are well formed or not without prior knowledge of what they're supposed to look like?
In some ways, I think morse may actually be easier to recover with zero prior knowledge because it's quite clear that the time domain is dominant and at that point it's just a matter of identifying word and symbol boundaries and then analyzing those (though again, some thought about how an alien intelligence might boot-strap meaning out of a series of symbols would presumably be helpful).
This is not about AM or FM modulation, that's quite easy to detect and is based upon the fundamentals of electro-magnetic radiation: there are only so many ways to manipulate a wave (manipulate amplitude, manipulate frequency, or manipulate phase). CW has all but fallen out of use here, but looking at the RF environment, it stands out
The point about power, of course, is likely more significant, though if one posits an extra-terrestrial intelligence far in advance of ours, technological questions about what kind of signal detection is possible become complicated and entirely speculative.