Ahh, so now it sounds like building it up piece by piece – unison, m2, M2, m3, M3, 4, etc.
That is actually the GNU Solfege is physically laid out. It also returns us to the main idea. When I looked into the science of methods that worked best in learning this, there is almost zero study on the subject. Its like we have various teacher doing different ways here and there without any actual tracking of students to see if any one method is better than any other, with the sole exception that everyone agrees that singing intervals makes the learning faster. Because of this blind spot I simply looked at what was easiest for me as I tried to learn it, and that was having octave as easiest (although I never though about unison). After that the next easiest for me to get 100% on was the 5th and then the 4th. This pattern reminded me of the overtone series so I simply based the learning patter off the overtone series inclusive of its inversions. Thus instead of this learning pattern
unison, m2, M2, m3, M3, 4, tritone, 5, m6, M6, m7, M7
I got this via the overtone series (mindful of inversions)
octive, 5th, 4th, M3, m6, m3, M6, M2, m7, m2, M7, tritone.
To which, thanks to you, I would now add unison at the very start. In short the method builds a gradual hierarchy of dissonant's, so your learning from most consonant to most dissonant. I can not say its any better than learning it the u, m2, M2, etc way, and as far as I can tell, no one has ever studied it to see – all I can say is that it fit, at least to start with, with what I found by far the easiest things to learn. For whatever its worth I did have a music teacher from Australia tell me he thought it was an excellent idea because in his experience of year of teaching aural training, kids very show errors when tones are close in the overtone series, say a 4th and a M3. I know it seems odd but apparently the guy has tracked the error types and discovered overtone proximity to by high on the error type list. Thus, his thinking was, that by teaching based of overtone series proximity the student becomes and stays critically aware of the very factor causing so much problem in many people right from the start (I just learned this yesterday). I would have never in a million years have guessed that, but there it is.
I have noticed that it self reinforces melodic direction study. It does this by grouping inversions. In this way ascending and descending study's are reinforcing the inversion relationships which then make it all that much easier to notice – maybe (again, no proof). Anyway, so there we come full circle.