The general problem with Alexa and skills (and really with voice interfaces generally) is that the voice recognition has gotten pretty good but conversational speech understanding really has not. As a result, commands are sort of like a wizard's incantation (or a CLI). You have to pretty much nail the exact syntax you need and there is very limited discoverability. That's fine for a few things but remembering syntax for a bunch of different seldom-used skills quickly breaks down.

Useless for anyone who listens to music through Spotify, though, which is still the majority of the streaming market.

Exactly! It's super annoying that Apple started discriminating 3rd party services in Siri support. I tried Apple Music, it's ok, but I support Spotify for a reason - I do not want to live in a world of conglomerates in power of every aspect of my life. And I have the right for that decision, so Apple could respect it and allow me to listen to music on my preferred service, since I already spent a lot of money on their devices and going to spend more, if they stop annoying me.

Hard to know what goes on behind the scenes in any particular case but I think this "discrimination" goes both ways.

For content like that to be a first-class citizen on iOS/tvOS, the providers (Spotify, etc) would need to expose the relevant metadata, APIs, etc to Apple.

I know that on Apple TV, there were a lot of providers who balked at providing that information, preventing Apple from fully implementing relevantly simple features like system-wide content search.

I certainly don't blame the providers, because they wanted to maintain more branding and autonomy rather than being a faceless Yet Another Source of Content on AppleTV, because at that point they'd be essentially faceless interchangeable.

Spotify on iOS exposes metadata and basic commands. When music is playing, the track info and album art shows up on the lock screen, and you can pause or skip using the lock screen buttons. Apple makes a standard API available for that stuff, and apparently Spotify is happy to participate.

Seems to me that Apple could pretty easily let any app opt in to Siri integration too. All they'd really have to do is pass commands saying "play song/album/artist named X" to the app, and let the app handle it from there. Add a setting somewhere to let the user choose which music app they want to use for Siri, or make them specify it as part of the command, and off you go.

If Apple offered this and apps like Spotify refused to participate, then maybe we could talk about this going both ways.

Other than setting a default app, isn't the SiriKit SDK exactly what you describe?

Edit: Nevermind, looks like Music apps are explicitly not allowed.

Exactly! Apple has explicitly opened up Siri to certain kinds of third-party apps... but not music.

And in many cases the model Apple enforces is not particularly generalisable. For example, the Apple TV content search model only allows for TV shows to consist of x seasons of y episodes. The BBC hasn't been able to provide comprehensive metadata because many shows (such as soaps and current affairs, or un-numbered specials etc.) don't fit into this model.

What about Amazon Prime Video? That still isn't on the Apple TV but is available on every other device.

Not on the Chromecast.

At least Prime video has been announced to be coming to the ATV, although I'm not sure what the hold up is.

I wouldn't mind they kept me within Apple Music if Apple Music, iTunes were Great, or good, or at least half decent. But no they are a piece of crap. Every single part of the UX in Apple Music Sucks, iTunes has become an after thought. And to add insult to injury, the collection of Music is only good in UK or US, in Japan and SEA region it absolutely horrible.

So to me Apple Homepod is just another way to try and sell customers into the Apple Music Services Revenue, bound and pulled by the iPhone halo effect.

Just to be absolutely clear - Apple doesn't "discriminate" against 3rd party services, they just don't support whole classes of applications "yet" (in quotations marks because I honestly would have expected them to expand SiriKit to music and podcasts before... sending money)

I presume you have sources backing the claim that Spotify is the majority of the streaming market. Be that as it may Apple Music is available in far more places that Spotify.

This is as of early last year : https://www.musicbusinessworldwide.com/apple-music-is-now-in...

And as far as I can google, Spotify still isn't available anywhere in Africa officially as of September 2017.

Spotify being a majority of the streaming market numerically doesn't matter too much if its market reach is limited.

Based on numbers [0] released by both companies over the summer, Spotify has twice as many subscribers as Apple Music does.

Surely their having achieved this with limited market reach is positive for them, and not for Apple?

[0]: https://www.theverge.com/2017/7/31/16070982/spotify-60-milli...

I think it shows they're both very strong competitors. Apple Music, despite being much younger and being 'less cool', has managed to obtain half the market share.

Spotify, despite being in less countries, has managed to go deep into the countries it does support.

> I presume you have sources backing the claim that Spotify is the majority of the streaming market. Be that as it may Apple Music is available in far more places that Spotify.

Yes, a very simple Google search for "music streaming market share".

This is something that I've never understood. This happens with Google Home too.

Something like Harmony uses native API's on Alexa, but third party API's on GHome. So on Alexa I can say "Alexa, turn the TV on" while GHome I have to say "Ok Google, tell Harmony to turn on the TV. Google Home will then reply "Ok, here's Harmony", pass the query to the Harmony "app", and Harmony will spit back "Ok, turning on the TV".

It's great for when I've lost the remote, but a terrible UX. What I don't understand is why can't I as a user determine what skills/intents get promoted. I understand the "Tell X to Y" for security measures, but if I could have a whitelist instead of a bunch of 1st/2nd party "power skills" it would make the entire experience more intuitive. I personally do not use skills just because of how clunky they are.

> I briefly tried some of the "skills" but they're awkward as hell. You have to say something along the lines of "Tell <app name> to <command>". Yikes.

That's a little like saying "some websites are rubbish" though. IME skills don't have to have cumbersome conversational UI - I use mine (in conjunction with a Logitech Harmony Hub) as a sometime replacement for a universal remote: say "Alexa, turn on Sky/Netflix/xbox" and on come three devices, all switched to the right inputs. And in the case of Netflix, it launches my TV's Netflix app too. The phrase feels natural to me.

That’s a special case as far as I’m aware. Alexa supports ‘Turn on <noun>’ inherently, the hub must be exposing endpoints in the right way for that to just work. 3rd party skills have to follow the more awkward syntax however.

Ah, I didn't know that. I thought the third party Yonomi skill had just somehow registered "turn on" as an invocation, but upon further reading it seems that's merely an implementation of the PowerController API†. Thanks for enlightening me.

† https://developer.amazon.com/docs/device-apis/alexa-powercon...

That sounds a lot like AppleScript. Which was not a Good Language.

The skills work well with Dish: "Alexa, tune to ESPN" or "Alexa, play The Big Bang Theory".

https://www.mydish.com/support/products/hopper/how-to/alexa-...