@ikir
Ah reinventing history is fun, isn't it ?
Yes Eyetech did tell costumers that the A1 was a product with edges, but they also went to any extreme to make clear that all these problem would only ly in SW or FW, that the HW would be identical to the consumer-release, and that all bugs were just fabricated by the competion.
Bout Ben Fudd, yes he has been prooven wrongs with what he stated back than (which changed on an allmost weekly basis, but lets forget about that).
The A1SE/XE couldn't do (reliable) DMA back than, and even now they can only do it with extra HW (SIL-controller) and special kernels.
It maybe that MAI finally found a way to get reliable DMA, but thats 2 years late for companies that wanted to sell Articia-based products in 2003. You know, time is an important factor in the real world.
Bout the MikroA1. I do remember being told that DMA would work with the old A1s cos there is this new revision Articia, the new firmware, a new kernel and whatsnot, only to find out days later that bug was still there (till it was declared a feature :-o ), and for now it seems pretty similar with the MikroA1, I still haven't found a clear answer wether it still needs special kernel or if it has some sort of cache-coherncy fix added somewhere.
And I'm certainly not gonna trust Ben or one of the fanboys on this one...
