Let's all remember what I2S stands for: IIS > Inter-IC Sound. Meant for interconnecting DIT/DIT to DAC chip and so on. Was never meant for outboard connections hence that lack of standardization. That's what the AES/SPDIF format was created for. I wouldn't be surprised if it actually performed worse when using outboard connections.

Well, it looks like at least two recent posters disagree and think that I²S is better.
There's some sound tech to support their opinion; I2S includes the clock signal; AES and SPDIF do not.

However, there is some truth to your concern; I2S is only designed for very, VERY short runs. Maybe for runs of over 30-60cm (guessing) the other connections are better. But if you can manage with a super short interconnect, I believe it's the tech of choice.

I think that people's frustration is that the best protocol just hasn't been developed; it's been ignored by most of the industry, over the years. But when the majority seem so obsessed with spending more and more money to get the best out of a usb port, who can really blame the industry?

To evaluate decent transports, is it enough to have a look on the trace of SPDIF signal on a scope? What do you think of this guy?

http://www.lampizator.eu/lampizator/TRANSPORT/CD_transport_DIY.html

http://intona.eu/en/products

The problem is not jitter but noise, and from what I have gathered this noise is very difficult to eliminate even with galvanic isolation after the usb receiver and reclocking/buffer after the isolator.  Or one could also try to limit the work the PHY must do (by improving signal integrity) and thereby minimise the noise at the source.