A VSTi essentially has two parts; the noise making bit which is primarily generic platform independent code and the user interface which (in the packaged form of the VSTi) is tied very much to the OS. In order for the Oasys to load a Windows VSTi something like WINE (a Windows API emulation) would have to be folded into the Oasys Linux, and that in turn relies on a Linux Windowing manager being in place as well. Since the Oasys is only using a tiny amount of Linux and (I beleive) using it's custom UI management, adding WINE and a Windowing Manager would be quite a large undertaking and reworking of the Oasys.
The Oasys is not just a Linux box running the Oasys as an app.
The Open in Oasys means that it is not a 'closed' system like the Triton or M3 which use hardwired chips for their noise making parts and have finite extensibility. Instead it uses software and that is open to expansion beyond what has been traditionally available. Software is mutable, hardware isn't so much. The M3 received a significant update recently but that was all to the UI which is software based, the noise making part is set in stone unlike the Oasys.
It's not Open in the computing sense, I can't develop and load my own EXf for example. There is no third party API, or any other means of hooking in third party code.
Ok so it's a little more complicated than I thought But are we now looking at massive development costs and time, or do you think this would be very doable ?.
I know you can only give a rough guess, but your guess would be better than mine since you program all day long everyday.
It really would be a lot of work IMO, not just in terms of writing new code and integrating something like WINE, but also re-architecting the system to be more Linux and less Oasys. I suppose Korg could take WINE and refactor it to work with their UI primitives or provide their own emulation layer that took Windows calls and mapped them to their own graphical system. Yuck ! Lots of work and lots of compatibility testing. Korg could more profitably spend their development time elsewhere, I suspect.
With that Muse Research box I believe they have to get the VSTi developers to tweak their code to work on the box, and that is probably a better way to go. However, how inclined would a developer be to doing that for such a small audience. Only a smaller developer would see it as an opportunity I suspect.
While what you say is largely valid, the fact is that creating Karo sample programs is vastly more straight forward than developing a quality Plugin. And developing even wonderful Combi's is similarly more straight forward - despite the obvious unique talents of EJ2 and other (that I for one do not have).
Sharp - I do think the development costs are high, and mark-ups low. Many music technology companies fold – is a tight business. I have a guru engineer friend who has done plenty of designs for Digidesign, Focusrite and so on over the years - he's sough after across the world - and from a chat with him recently, the economics of music technology companies is very poor and has been since the onslaught of DAW's over hardware. Many household names in traditional studio and music technology circles are just about hanging on in there right now.
So while OASYS may be Korgs showpiece - well - they've already done it. There's no golden rule that says they have to support it forever. As we've already acknowledged on this forum over the months - there are many hardware aspects of OASYS that are already looking a little dated - such as inability to connect via USB to a computer, no DAW plugin capability, its internal HD size, processor speed and so on. If they were showcases three years ago, they are not so right now.
Furthermore, just because this forum has acknowledged that OASYS doesn't have to be profitable does not mean that's the way Korg think on it. OASYS was / is a labour of love alright, but they still have to balance a profit and loss sheet every quarter, and I can't see any company putting money into low profit areas if they can use the same workforce developing newer and more profitable products.
I personally suspect that OASYS was a bit of an experiment for Korg - and I suspect that may have hoped to sell many more units than they did; and for it to be self sustaining economically; and although all of this is my opinion only, and I will continue to use OASYS for the foreseeable future (indeed I have made the decision to purchase a 2nd one over the coming months for the long term because it's simply that important in the long term for me) I do not believe it to be economically viable for Korg; and even though there may be some more updates; it has been simply too long already since the release of MOD-7 for me to believe that OASYS will enjoy many more major updates or support. We may see the M3 sequencer updates come our way - and I hope we do - but I'm not holding my breadth .
Re. 3rd party development - while OASYS uses the Linux Kernel, virtually none of OASYS's OS or real functionality is based in Linux. I'll be delighted to stand corrected on this – but isn’t Linux used for OS book keeping type stuff, with all of the real OASYS functionality being very low level code to maximize speed. If so then all of the usual API's, Layering, standardization and so on typical of software development within an OS environment is totally bypassed in OASYS. This would mean that any 3rd party would have to recode their plugins completely. They would also have to learn the entire OASYS / Korg in-house way of designing and coding. Considering a typical engineer salary as being for argument sake $100,000 pa then you could see it costing at least several hundred thousand dollars in training/up skilling, development, testing, publicity, release and support for any plugin of the quality of say an NI or Arturia Soft Synth. So this would not just be one person hacking code, it would be at least a small team, a strategic business/engineering commitment with Korg - and it would not be cheap.
In my view, either Korg would have had to sell a lot more OASYS's, or it would have to demonstrate reusability of any 3rd developed plugins across a suite of profitable Korg products before any 3rd party could afford to put their eggs in the Korg basket. The bottom line to me is that the economics do not work out.
Remember - Stephen Kay is working individually so overheads are low. Also - he's extraordinarily talented and well beyond your average excellent engineer, perhaps doing similar work to 3-4 engineers - yet - he has still dedicated about the past decade of his life to Karma - AND it's on the Karma, OASYS and M3, with software for all of those as well as for the entire Triton range and now the M50. Plus he's selling his software separately so it’s another revenue stream.
Overall, OASYS is no more open than any other workstation. Again, I believe the openness to be a vast in-house DSP library for Korg that enabled Prophecy, Z1, MOSS boards, OASYS-PCI, Karma and M3 and so on - and we are the winners. But just because it's OS is on a HD and bootable, and it runs on an Intel chip, makes it only marginally more accessible to 3rd party developers than, say, the Fantom, Motif or PC3.
But are we now looking at massive development costs and time, or do you think this would be very doable ?.