Doing a good job of constant prop means doing a thorough job of implementing the best algorithm in the best paper. And to do that, you need to build a form of SSA. And to do that you need to compute dominators. And to do that, you need to build an appropriate form of the CFG. Don't need a PhD to do all those things (not inventing anything new here, just using other peoples' beautiful results), but it's still plenty of work to build a fast & thorough implementation, with lots of testing required.
Can't possibly be economical; but of course, the experience will teach you a lot.
Has anyone used this approach to memorize music?
E.g., for playing an instrument.
I used to be able to remember plenty when I was young and practicing a lot. And certainly professionals can remember significant amounts (and not just the music, but performance details, fingerings, etc).
Seems like a different mechanism, relying less on visual and spatial associations, more on hearing. But perhaps it's similar in that you learn a structure and attach details to it.
Or, having developed a good memory for music, can we use it to help memorize random facts? Perhaps associating them with places in a song.
The circle of fifths can be studied (and often is) with the loci method. Visualising the circle of fifths, with the relative minors, as a real space that can be moved around in is very helpful.
Music theory is full of little mnemonics, and is itself a sort of system of mnemonics:
Instead of an infinite gradation of frequencies, they are reduced to 12 repeating notes within each doubling, so that a 440 Hz vibration is an A, as is a 220 and an 880 Hz vibration.
If instead of doubling a frequency, we add a half to it again for a 2:3 ratio, we get a perfect fifth. By repeating this process (with some fudges) we end up with twelve notes, the circle of fifths, and the vast majority of Western music.
Rather than simplistically giving these 12 notes 12 letters, they get 7, in a clever system that allows for all 24/30 major and minor scales to have one of each letter A to G.
The 7 notes in any particular major or minor scale can be combined with each other to form chords. Major and minor chords each share a root and a fifth; the third is different by a single semitone. The combinations of three notes that sound nice together are thus greatly simplified.
Once the circle of fifths is internalised, all 24 (or 30 - ask Wooten) major and minor chords can be constructed, transposed, and played with.
So, when an experienced musician hears a song they can put all of that background to work. Instead of remembering each indivudal note played, the notes are chunked into chords, hich are chunked into keys and progressions. Once contextualised with similar pieces the memorisation becomes easier and easier. By the time all of that contextual data is assimilated, with music that is _felt_, there isn't much need (I'd imagine) for good musicians to use memory palaces for fingerings or performance details. I'd say a lot of that knowledge is in the hands, the breath, and the whole body.
I use a mnemonic technique to remember song forms when I'm playing drums. I draw a map of the song, each measure is a square. Squares are arranged in lines in such a way that it makes sense musically (usually 4 squares per line). The parts of the song (verse, chorus, etc.) are separated. The squares are mostly empty, except when they contain something important to remember.
The drawing takes up more space than the typical informal drummer notation (for example: intro 4 bars; verse 8 bars; chorus 8 bars, etc.), but it is much easier to remember (at least for me).
If I'm learning a song that isn't recorded yet, I sketch this map during rehearsals. Before each rehearsal I try to visualize the maps of all the songs.
If I'm learning a challenging song from a recording I use a simple python program that draws the map from a given text file and allows me to play a loop over selected measures. This way my brain starts constructing a map of the song without conscious effort.
This sort of maps are some kind of constructed memory palaces in 2 dimensions. I personally wouldn't use an existing memory palace to remember a song, because the songs already have enough structure.
I need this technique only to remember the big picture. When I need to learn the changes or the melody, I try to rely more on music theory and "audio memory". If you are interested in learning jazz standards, you might want to check the book "Insights in Jazz" written by John Elliott.
As someone who, back when I used to play (semi-professionally) I preferred to memorise all my music, I would say that memorising pieces for me was "effortless", and thus needed no technique, provided that when I would first study a piece, the intent was specifically to memorise it, rather than, say, to sight read, or analyse it.
I always found it extremely odd that, if the memorisation intent was not there, no amount of exposure would lead to effective memorisation. E.g. I could sight read an accompaniment 50 times and I would not remember most of it; once I decided to study specifically with the intent of "memorising" the piece, typically a handful of passes would be enough.
Scott Adams recently said something similar in one of his coffee podcasts. He made an experiment with himself: he has a short "skit" he always starts his show with, and he made a bet with himself how long it would take, if ever, to memorise that skit, simply by reading it, with no intentional effort to memorise it.
Yes, memorizing everything (all the music) helped hugely in practicing. I used to believe it was because I didn't have to read the music, and could focus on fingers, etc. But maybe the memorization effort (and it wasn't hard, as you mentioned) was enough to establish a framework, and the bulk of my practicing was hanging new details onto the framework.
In the renascence there was a standardized system of memorized hand symbols to recall music. For guitar you can memorize fingerings by creating a person action object system in a way that tells you the fingering, the string and the fret number in one image. You can then use this to memorize classical music. With music you are going from conscious to unconscious procedural memory. But you can "double encode" using memory palaces and systems for a "backup". (It can actually help you learn it faster because you can get rid of sheet music)
Compilers have parsers, not languages.
Many languages have more than one compiler (e.g., C), and the various compilers may use different techniques for parsing.
Indeed. But consider the 2nd sentence, the conclusion, and the twitter post.
On the other hand, I certain agree with your idea about teach/using handwritten recursive-descent parsers.
Here's an old book that presents it pretty clearly, along with a nice approach for error handling. Maybe you can find it in a library. https://www.springer.com/gp/book/9783540082408
Happens in other fields too.
When I worked at Tera, I (and plenty of others) would talk to Burton Smith in the halls.
Whatever the topic, he always made us feel included and smart.
Then, after the conversation, as we moved apart, our IQ would drop and our understanding would fail.
