Cockos Incorporated Forums - View Single Post

Airal · 08-06-2017, 12:23 PM

karbomatic: Wow, we keep going in circles here.

But you seem not to realize that reaper is using 20+ year old technology. Winmm was written around win3.1 and that is what reaper uses and 99% of other daws(a few of the progresses daws out there are just now updating their code base for it).

I am not arguing at all that windows could write a bridge for it... what I am arguing is the reality of the situation... which is, that hey haven't... At least officially.

It is not a question of whether reaper should or should not have to do anything but a question if does it need to do it to support a large set of modern devices.

---

Have you actually ever written a device driver? Or worked through the ddk? Be honest! I have! Now it's been over 15 years since I've done it, and things have changed quite a bit since then... but it is not an easy task and it is NOT the same as writing a C# app or even a typical C/C++ app. It is more akin to assembly and it is not object oriented(all win32 stuff, but you can't just do stuff like you can in ordinary programming). There are pitfalls and order of operations that must be taken or you will crash the OS. So, it is entirely different in that regard. A kernel mode driver can crash the instantly while a user mode app will rarely crash the OS(well, today at least). Also, debugging was usually done by console methods rather than what we use today.(I'm not sure how far it has advanced).

Not only that, but the amount of documentation and examples is much smaller.

For doing certain things, yes, it is not too complicated. But for many things, it is much more complicated than your standard programs(specially today with most things being abstracted away).

Either your a pretty good programmer and fail to take in to account the collective of modern programming or you haven't done much in depth programming. Most programmers are script kiddies that couldn't write a proper app if their life depended on it. Ruby, python, etc are not real systems programming languages and, IMO, are not really much of a programming language in the first place. In fact, Object oriented assembly is where it's all at, but no one has bothered to write a compiler for it.

Essentially what you are arguing is that bare metal microprocessing programming is equivalent to high level javascript web script programming like greasemonkey. If you really think your average programmer can walk in in and write a well written device driver in the same time, approximately, that they can write a some web script or whatever, then you are pretty ignorant of the state of affairs. Again, they may be the same for you depending on your skill level, but you are not everyone else.

C# and .NET are easier languages not because they are higher level, in any real sense, but because they are more unified, better put together, and have more plumbing. The .NET is a very nicely designed framework. Basically the only one in the world that actually was put together well IMO. It's almost a pleasure to use.

But, I actually use D to program in now days and it's pretty sorry. It has one of the worst libraries I've ever had the displeasure to use... but the language itself is awesome. The meta programming capabilities blows anything I've ever used(20+ languages) out of the water. I can do things that I've only dreamed of doing in other languages. (it's compile time capabilities are astonishing). It's not perfect, and the language itself could be much better organized, but it is extremely expressive(relatively speaking). As far as drivers in C#, one could do it if the kernel allows IR to be ran... I've not looked in to it so I don't know. The issue isn't so much with instruction amount... you can write whatever core optimized functionality you need in native code and bring it in if you want.

Anyways, this is starting to get pointless. Neither of us can prove our opinions... I think I have more facts on my side, but I'm sure you think the same. One of us is almost surely more right than the other though and I'll just assume that that person is me

Cause I'm sure you'll do the same.

Cause, at the end of the day today I still don't know what your ultimate point is. Is it that cockos should simply not implement any code to try to use midi over bluetooth and wait for someone else to make it happen? If is, then that is just your opinion. My point is that reaper cannot use these devices(or many of them)... and that is fact. It's ultimately up to cockos to decide what they want to do.

My main issue with you is the mentality that you take about leaving it up to someone else to make something work(assuming that is your position, which is what it sounds like). If everyone had that mentality nothing would get done. Everyone would just point the finger at someone else... that is also fact. Any time anything gets done is when someone decides to do it. That might be trying get others to do the work, as I have tried to do with this post. But it's better than pointing fingers, which is as close to absolutely useless as one can get. I could, of course, write my own DAW and implement design my own processors and audio interfaces and all that. I'd love to do that one day but, of course, there is only so much a single person can do.

08-06-2017, 12:23 PM	#37
Airal Banned Join Date: Nov 2015 Posts: 406	karbomatic: Wow, we keep going in circles here. But you seem not to realize that reaper is using 20+ year old technology. Winmm was written around win3.1 and that is what reaper uses and 99% of other daws(a few of the progresses daws out there are just now updating their code base for it). I am not arguing at all that windows could write a bridge for it... what I am arguing is the reality of the situation... which is, that hey haven't... At least officially. It is not a question of whether reaper should or should not have to do anything but a question if does it need to do it to support a large set of modern devices. --- Have you actually ever written a device driver? Or worked through the ddk? Be honest! I have! Now it's been over 15 years since I've done it, and things have changed quite a bit since then... but it is not an easy task and it is NOT the same as writing a C# app or even a typical C/C++ app. It is more akin to assembly and it is not object oriented(all win32 stuff, but you can't just do stuff like you can in ordinary programming). There are pitfalls and order of operations that must be taken or you will crash the OS. So, it is entirely different in that regard. A kernel mode driver can crash the instantly while a user mode app will rarely crash the OS(well, today at least). Also, debugging was usually done by console methods rather than what we use today.(I'm not sure how far it has advanced). Not only that, but the amount of documentation and examples is much smaller. For doing certain things, yes, it is not too complicated. But for many things, it is much more complicated than your standard programs(specially today with most things being abstracted away). Either your a pretty good programmer and fail to take in to account the collective of modern programming or you haven't done much in depth programming. Most programmers are script kiddies that couldn't write a proper app if their life depended on it. Ruby, python, etc are not real systems programming languages and, IMO, are not really much of a programming language in the first place. In fact, Object oriented assembly is where it's all at, but no one has bothered to write a compiler for it. Essentially what you are arguing is that bare metal microprocessing programming is equivalent to high level javascript web script programming like greasemonkey. If you really think your average programmer can walk in in and write a well written device driver in the same time, approximately, that they can write a some web script or whatever, then you are pretty ignorant of the state of affairs. Again, they may be the same for you depending on your skill level, but you are not everyone else. C# and .NET are easier languages not because they are higher level, in any real sense, but because they are more unified, better put together, and have more plumbing. The .NET is a very nicely designed framework. Basically the only one in the world that actually was put together well IMO. It's almost a pleasure to use. But, I actually use D to program in now days and it's pretty sorry. It has one of the worst libraries I've ever had the displeasure to use... but the language itself is awesome. The meta programming capabilities blows anything I've ever used(20+ languages) out of the water. I can do things that I've only dreamed of doing in other languages. (it's compile time capabilities are astonishing). It's not perfect, and the language itself could be much better organized, but it is extremely expressive(relatively speaking). As far as drivers in C#, one could do it if the kernel allows IR to be ran... I've not looked in to it so I don't know. The issue isn't so much with instruction amount... you can write whatever core optimized functionality you need in native code and bring it in if you want. Anyways, this is starting to get pointless. Neither of us can prove our opinions... I think I have more facts on my side, but I'm sure you think the same. One of us is almost surely more right than the other though and I'll just assume that that person is me Cause I'm sure you'll do the same. Cause, at the end of the day today I still don't know what your ultimate point is. Is it that cockos should simply not implement any code to try to use midi over bluetooth and wait for someone else to make it happen? If is, then that is just your opinion. My point is that reaper cannot use these devices(or many of them)... and that is fact. It's ultimately up to cockos to decide what they want to do. My main issue with you is the mentality that you take about leaving it up to someone else to make something work(assuming that is your position, which is what it sounds like). If everyone had that mentality nothing would get done. Everyone would just point the finger at someone else... that is also fact. Any time anything gets done is when someone decides to do it. That might be trying get others to do the work, as I have tried to do with this post. But it's better than pointing fingers, which is as close to absolutely useless as one can get. I could, of course, write my own DAW and implement design my own processors and audio interfaces and all that. I'd love to do that one day but, of course, there is only so much a single person can do.