antoinebalaine

That's not perfect pitch. 90% accuracy in pitch recognition based on memory retrieval is not perfect pitch. As for the bunch of comments here claiming they acquired AP at adult age for "only" one or two notes: that's not perfect pitch either. AP can't be acquired at adult age - your brain only learns to recognize pitches during your neural plasticity phase as a baby (provided it's exposed to enough variety), just like it acquires speech and color-discrimination. Also, AP is not a party trick. Not having it is akin to being color blind - even if you could "remember" what colors are supposed to look like, you would _still_ be color blind.

That the article's abstract and researchers are willing to claim the opposite despite their own evidence : this smells like butt-hurt denial.

Isn't this the case in most universities?

I'd argue that if you've never longed for automations, it's because you've never written a hard score. Any ambitious piece written at the churn pace that a composer needs to make a living - any such piece uses algorithmic tricks under the hood. There's human decision and judgment in using those, but it doesn't diminish their value - much less having them as automations.

Using vim to edit large dictionary files. I forked a steno dictionary for Plover, and edited its 300k entries in vim. I've ben using regex daily since then.

The real reason is because we tune instruments differently. Anything about "why would you want to be a copy cat of Bach?" is just non-sense - musicians have been copy-catting each other endlessly since music existed, and still do today.

The writing technique of the baroque era revolved around line conductions that dodged harsh-sounding intervals in the temperament of the time.

Nowadays, we tune using equal temperament. Equal temperament doesn't have harsh-sounding intervals (except flat 9th on a major chord), which makes classical writing technique obsolete.

The shocking thing about Spotify's UI was that it was never modeled to be a record collection or a record store. It was always modeled to be a database front end.

The ui design and the business model always reflected that prerogative: poor discoverability, poor user-favorites, poor classification, absolutely no blobs, credits or liner notes. An artist uploading is just another entry in the database - the service sells access to the DB, it doesn't sell music. By extension, the user is also more and more a "yet another entry" in the whole process of "consuming" the data.

It's just dehumanizing, yet consumers would rather this than having to deal with the friction of purchasing records (which are outrageously expensive, which also points to the record industry's responsibility in this...) or pirating under they eye of copyright trolls. Long story short: they'll fail, sooner or later.

If you're OK with switching from gapps, a de-Googled android works just fine. I've had to keep a spare googled-android around for the rare occasions I need to use lyft and Uber. That's it.

Looking forward to discovering this standard. After 2 years working on parsing ABC, I realize how difficult it is to represent notation. Kudos on this effort!

Checking in with USG before publishing leaks is actually fairly standard procedure in order to avoid operatives on foreign soil being endangered. Wikileaks did so at everyone of the major releases they did in conjunction with NYT, WSJ, spiegel and Le monde - it's also the reason they stopped collaborating with some of those journals when they found out some of their journalists hadn't followed procedures. And it's also the reason why claims from USG that WL was endangering people's lives abroad were always bogus.

About NYT withholding documents: well, that's just another nail in the coffin of the paper's credibility, along with yet another betrayal of the trust they got from WL...

I use the French bepo layout for everything, but all my languages are Roman languages.

Here's the context: the fr gov's been acting against national interest for more than 20 years in a row. The people know it. The gov knows the people know. The gov wants to pretend it's not doing anything wrong. It could keep pretending credibly, if only people would keep their trap shut on the web. So now gov wants to control what people can see and say on the web.

I saw your design and it's fascinating. The whole concept of writing without distractions is really awesome. Good luck!

Kagi's made the internet pleasant again for me - the very low amount of noise in the search results, and removing the pervasive sensation that I was constantly getting tricked into garbage results, that's been a very positive shift. For 10$, it's been totally worth it - search remains an invaluable daily door into the internet, and good quality pay-for-search promises positive compounding effects: low noise-to-signal ratio, a potential shift in business incentives away from advertising, better access to information, a reduction of bias, etc.

A month into Kagi and not one fallback-search I've made on Google yielded better results. I'm glad to finally have the option to use an essential web service without being turned into a product for advertisers. Would definitely pay more if I had to.