Obviously you won't get a 1:1 correlation with legacy desktop apps. ReBirth is an example of an app arguably not worth porting to ipad. Unsurprisingly, the ipad does better in areas where it innovates and sometimes exceeds desktop capabilities - for example apps like Konkreet Performer, Traktor, Reactable, Samplr, Lemur, Impaktor, Electrify, all the Korg and Moog apps. There are a ton of DAWs to choose from but they are among the less interesting things out there.
Well, touchscreens are trendy right now and they're moving a lot of hardware and software.. but for real hardcore musical creation?
They won't be a centerpiece like a desktop running Ableton Live until they get a hell of a lot more powerful.. and frankly, the first thing most people will do with their touchscreen musicbox is find a board covered in keys, knobs, faders, and buttons to attach to it.
I think people might underestimate just how tactile music is and how important it is with electronic/computer-powered music to see what you're doing at the same time as you're doing it or--conversely but no less importantly--to be able to do what you're doing without looking at it.
I mean, there's talk of tactile feedback touchscreens and everything else, but really--this is a solved problem, it's called a button.
You're obviously correct that tablets are CPU-bound compared to desktop which is why the more exciting stuff is native on the ipad rather than being a port. Traktor is a good example - the ipad version is very different and stripped down yet more powerful in a number of ways. When you talk about 'hardcore creation', remember that Brian Eno is a big fan of the ipad. Now you can get Reactable for ipad and even an emulation of the Tenori-on - you won't find anything remotely like those capabilities (or e.g. Samplr or Impaktor) on desktop.
And the controller market is much larger for desktop than tablet - the tablet is a very popular controller for the desktop (especially Ableton as it happens). Consider that for a long time Lemur has been a high end pro choice precisely due to it being a soft controller and they are now working with the ipad.
I'd like to see Eno do a show with naught but one ipad. I think Reactable is basically a toy and as for the Tenori-on.. I don't know if you've ever played one, but it was kind of terrible.. I think it probably would do better as an app than a discrete physical device though. It looked cool and novel, but the build quality was poor and the interface actually wasn't worthwhile, it was a toy. It's so much easier to address tone using a keyboard, and you don't need a crazy 16x16 grid to address time. A linear array of 16 buttons, any of which you can hold to address a specific point in time, and then manipulate that point in time by pressing keys on a keyboard or turning encoders to set parameters works very well.
Samplr looks cool, but I don't actually see the capability. Controllerists hack together things like this in PureData, Max/MSP, and Reaktor all the time. Just watching the video, the issue of the finger covering the display is huge. I see him changing parameters that are tightly clustered by pressing on them and then sliding his finger left and right. So.. the nearby parameters are inaccessible when doing this? What happens when my hands get sweaty?
I think touch screens have their uses but they simply don't replace tactile controls. Having an integrated computer and screen that can present a large amount of information is itself a huge advancement. The overly touch-centric UI is a regression, in this musician's opinion. I see a lot of novelty, but I've yet to see anything that I personally would even consider bringing onto the stage. Keys work, knobs work, faders work, decoupled controls and display work. In all the novelty I don't see any real tools for managing the complexity and demands of live performance, I see complexity added without practical affordances.
Tightly coupling control and display is just not such a great idea and I really don't think it takes much imagination once you remember all of the different sensors and hard input devices that exist to come up with another way of implementing any interface that works well on a touch screen.
I'm more interested in the possibility of apps w/ dedicated controllers, or multipurpose controllers, than I am in the standalone apps. Unfortunately, these devices don't really put configurability very high on their feature sets.
If you haven't done so recently, go play a modern hardware synth or sequencer. Then think about the explosion of possibility that 'desktop'(Which is really a misnomer with laptops and small headless computers on the table) solutions offer... then revisit a tablet app and you'll find yourself asking, as I found myself asking, why do I have to obscure the information with my finger and why is such a large part of my screen given over to taking input?
NO tablets are great for playback and controller use cases for composing/creation not so much.
Now Live on a Surface pro might be good for music creation.