Just a note to peter - I've passed the R1.26 driver source to ATIK so that should be available to you shortly.
I'm quite happy with the drivers as they stand in the OSX form.. being a C++ driver with an Objective-C wrapper it should be easy to top'n'tail.
Here's the drivers working with the new ATIK One 6.0 at a recent SGL star party last week:
Where next with the drivers? Well portability I think is the next target.. although the base level transport is more efficient than usblib as it uses direct user buffer filling straight from the OSX USB I/O user driver interface rather than having a separate buffer.. I'm going to add a libusb transport and C++ based management. I'll also add GP guider support at the same time to maintain complete coverage..
Here's my crop of the daytime moon with the ATIK One - single exposure with red filter.. no darks or flats (and a dirty objective lens!):
We're trying maintain unix-wide source compatibility of our driver, we have still much more users on Linux than on MacOS X. We are just working on a version with separated USB driver and INDI wrapper (both C++) so the USB driver can be used with another software as well (e.g. lin_guider). If you will have similar philosophy maybe it is a time to merge our USB drivers to a common shared library and use it both for INDI and AOSX.
You are right, libusb is probably not as efficient as native API, but portability is a priority for us (and for couple of MB/hour it makes no difference).
I think it's important to recognise there's a slightly wider set of use scenarios than just pure astro (although the majority overlap):
* OSX portability - with minimal rework required for windows, OSX and linux.
* Driver transport that allows usblib and a direct usb transport - for maximum speed.
* There may be features that the drivers expose that may not be needed by INDI but these need to be available.
Accepting that means a considerable reduction in duplication all around, whilst providing both DSO imaging and realtime imaging.
Btw - I'm quite happy to work with INDI.. recent day job changes in responsibility means I have less time than I used to! Perhaps I should put up a hello introduction post..
Hi Peter,
I've been focusing on finalising some outstanding issues with the ATIKOSXDrivers - I found one really nasty issue caused by a bug on my side but masked by OSX returning a single error codes for a group of issues (including buffer under run). I only found this out by looking back at the old open source Apple OSX source code..
This caused issues with the 4000/11000 which have been resolved, and the upfront flushing now takes 9 seconds rather than the original 1m30s!
So - release R1.30 is current out in binary format and I'll be sending the sources to ATIK as per usual.