that's unfortunate if true. imo, wcf data services is very flexible and easier to bolt onto legacy asp.net than webapi.
the biggest obstacle with ios development on windows is the requirement to build against apple binaries. if they remove that obstacle by either providing a "macbook-in-the-cloud" or a totally hands-off compilation step, i would be on that in a second... seems like a great way to get more people into azure hosted services. imagine tight VS integration with 3 or 4 different project target types with seemless cloud-build functionality for windows phone, android, ios, etc... PCL for business logic, platform-specific project templates for ui bindings... the number and quality of windows phone apps would increase as the marginal cost of developing a windows phone app in addition to ios/android would be small. this would be my dream scenario, but then again, this is all rumor at this point.
webforms is still very big in enterprise and i don't see that changing any time soon. there are too many old vb6 guys who migrated to webforms who would be lost doing anything else. webforms are also a little easier to secure "out-of-the-box" than a bunch of http endpoints if you don't know what you're doing. if you do know what you're doing, then yes, you can do anything on a webform, especially if you use the webform as simply an html/js delivery mechanism. if you remove all unnecessary httpmodules, turn off viewstate, ignore the Page event model, and enable static ids (or better yet, don't use any server controls at all), you can get excellent performance/throughput. this is a trend i've noticed as business users are asking for modern js/css ui features but the business unit cannot get away from webforms for whatever reason.
@dentaku: well, there are people who care about music and dsp on windows but they're all doing it native. it's funny, i see a lot of libraries in java for midi and audio, just not a lot out there for c#. last i read, java vm performance is close to c# vm performance so i don't know why that is... could it be the people who would be most interested in doing foss midi/audio tend to favor the platform less bound to a specific ecosystem? i know mono eliminates that argument, but maybe the perception is still out there.
Yes, but that goes back to my initial point. If Node.js can't do it, then Web API probably can't so use WCF.
If you're going for speed to develop a LOB application, nothing beats ASP.NET with something like Telerik controls. Also, I'd see the MVC crowd abandoning MVC for something new before the webforms people.
As for ORMs, who cares, you're passing back json anyway. Well I'm an ORM hater anyway. I like to actually write my queries in sql to see what's going on and make sure i don't get an surprises.
...well there is that whole decade + worth of mature .Net code you'd have access to ;)
i kinda get what you're saying about writing raw sql. i don't like too much magic either. check out some of the micro-orms like petapoco or dapper. with the use of t4 templates, they give you the rich statically typed experience of an orm with the ability to extend your entities (using partial classes). Zero magic.
I'd rather write something like DB.Save<MyNicelyTypedClass>(MyNicelyTpedObject) than the dozens of lines to figure out upsert logic from your primary keys.
wcf is much bigger than the web. you can do .net remoting, com interop, lot's of different types of endpoints with rich meta-descriptions of interfaces so clients can easily get what they need to know about a service. i think mvc was a reaction to the rapid rest-like development that was hot with ruby on rails at the time, so they bolted a framework onto existing asp.net/iis. i think webapi represents a ground-up approach on how ms SHOULD do rest-like development. whether you like it or not, node.js is hot right now, so the decision to go with it for azure is likely a marketing thing. if you're already bought into the ms stack, chances are you're going to use azure anyway if you needed cloud, so the marketing right now is focused on getting those other guys who wouldn't otherwise be considering azure. if i had the luxury to start a brand new project with my choice of tech, i would go mvc/iis (webapi is just too different for me to learn it all right now), but enterprise developers rarely have that luxury. i might even go with wcf data services. you get odata and other nice rest-like features, but it's still on the back on the familiar asp.net iis stack.
wow, where did this come from? there are some impressive looking demos people have created with it already out there. i know the visual programming paradigm is always an inch deep, mile wide, but i really like the ui they've created for the object "brain". if they open up some real scripting abilities, something like this could be really cool to prototype on.
you sir, are awesome. thanks a bunch. proper plugin delay compensation isn't even implemented in ableton yet! lol @ petproject. you're doing cooler things for fun than most of the devs i've worked with professionally. i'll be sure to mention you if anything comes out of this.
seeing the midi and audio driver wrappers would be nice, but i think naudio might do some of that + i'd probably want to release anything novel i come up with for free. the host and vst integration seem like such cool projects to keep under wraps. are you planning on a commercial release/licensing agreement? codeplex/github and the c# music community would greatly benefit from opening those up. i had a quick look at the pitch tracker. awesome "divide and conquer" optimizations on the autocorrelation algorithm.
the non-realtime spec was more of a realistic limiting factor on the time i can spend on it (and my own dsp math deficiencies). it doesn't need to respond to realtime midi input... i'm thinking more like a standalone redrum step-sequencer in propellerheads reason, not a full featured daw with sample accurate time precision.
if naudio or other library already has some basic effects baked into i think i could certainly transform a wav buffer before mixing it down or sending it to the sound device, but i'd be out of my league if i wanted to write my own. heck i'd even be okay with doing some p/invoke to use any el cheapo "system" reverbs/delays that may be exposed through win32. i basically have some compositional aids and interface ideas i'd like to play around with.
i do remember reading all your interesting posts on the subject and i'm glad you chimed in. are you sharing the source on your project and/or do you have a project page? the high level modelling is what i'm kind of interested in at this point.
edit * heh... sounds like your SortedList is a lot like how i do it with 2d points. i thought there would have to be a better way if basing it off midi, but i can see the benefit of having that layer of abstraction between on/off events in the file and the in-memory sequence if you edit the sequence after starting playback but before the "Playhead" gets to the event.