[LAU] LV2, DSSI and the future of plugins

Hi list, I suppose I could've just addressed this to drobilla and got most of my questions answered but it concerns us all really as even if A3 and qtractor were to achieve feature parity with Cubase and buds tomorrow (obviously I'm stretching things a bit there) we still wouldn't see vast droves switching to Linux DAWs for a few reasons such as hardware support, people liking what they know etc. but most importantly the dearth of quality native plugins available for Linux seems to be a primary showstopper for most. DSSI has a few things coded for it but not much and its still very early days for LV2 so the open plugin format of the future is still anyones game and it may not necessarily be either of those that succeeds and gets widely used of course. I think a couple of VSTs have been ported over to LV2 but I'm not aware of any that have been ported to DSSI and I think that the ease in doing so is quite an important factor in the success of any such format, if not the be-all and end-all. If anyone here has any experience with coding and/or porting VSTis - what is currently lacking from LV2 or DSSI that could potentially cause problems for someone wanting to port their big beefy synth or snazzy FX from VSTi to LV2 or DSSI? I already know about the incomplete persist LV2 extension but I'm pretty sure that won't be the only thing needing work. A very important factor for such a format would definitely be that the major hosts (commercial, foss or otherwise) for all major platforms would be able to easily implement support for it and that plugins would be easy to port between the different platforms. I'm not aware of any DAWs for Windows that support LV2 or DSSI yet but I could be wrong? There's nothing stopping a closed source, commecial app vendor adding support for either format is there? Another factor I see as increasingly important is that the plugin format should be able to take advantage of OpenCL to take advantage of the superior processing power of todays GPUs. Quite how we'd convince Steinberg and co. we need a replacement for VSTi and get them to support an open standard though is anyones guess :/