True. This leaves open the very real possibility that any "designer" tests I define in terms of known processes may be doomed to fail.
OTOH, the means by which a particular physical process can be harnessed are probably fairly constrained. It certainly seems reasonable to base scientific tests on the assumption that there will be commonality. If that assumption turns out to be incorrect, then gosh, I guess my tests didn't support the hypothesis: which is exactly how it's supposed to go.
What would be bad science would be to simply assume we couldn't recognize non-human design, and not try to test it at all.
That being said, where's your baseline for non-human designers? Where's your collection of known products of non-human designers, from which you can draw inferences about unknown artifacts? Hate to say it, but the answer is: you don't have one, and hence it's of a wholly different nature than spotting signs of human designers.
Actually, we do have a collection of known products of non-human designers. For example, it's known that crows and other animals invent and use tools. True, they're not terribly sophisticated tools. If we take sticks as an example, we find that animals use them in the exactly same way you or I use sticks as tools. The point being, the evidence points (weakly) to there being a likelihood of commonality in human and non-human design.
What you are lacking are data points requiring an explanation other than natural selection. That would seem to be the minimum requirement for a paradigm shift.
I don't know that, and neither, I suspect, do you. We know that human processes are constrained, of course - some of the proposed candidates for Grand Universal Designer are not, shall we say, posited to be quite so limited. In which case, there's no reason to expect commonality, nor does the lack of it count as evidence against a presumably ineffable designer ;)
What would be bad science would be to simply assume we couldn't recognize non-human design, and not try to test it at all.
I'm not saying you can't test for it at all - I'm saying you can't test for it the way you test for human design, because you don't have the tools (facts and inferences) that you use to determine human design. That's why the two cases are not at all analogous - to do it that way, the way you examine things to decide if they're human artifacts or not, you need knowledge about non-human artifacts that you don't have. Worse still, you need knowledge about non-human artifacts that, thus far, you can't get.
Or you have to find some other way to test. Dembski thinks he has another way - I think it's complete bunk, but there you go.
For example, it's known that crows and other animals invent and use tools. True, they're not terribly sophisticated tools.
I believe you've mentioned the crows before. The problem is, the crows very neatly refutes your argument, insofar as nobody recognized the artifacts as tools in and of themselves - the only way anyone knew that the crows were using tools was by observing them in the act of using them. Do you expect to be able to observe the Designer in the act of designing, so that you may recognize his handiwork later? ;)