Suppose you migrate to "smart sensors". Things like PIC chips are little more than the cost of their sand these days. The pressure sensor could be built into the pitot assembly with no long tubes to collect crud. The sensor data could be massaged through a calibration table custom to at least the probe design if not the probe itself and then transmitted digitally to the newer ADIRU equivalents. Such a concept would allow for multiple sensors and multiple types of sensors for the same data in the hope that when one type of sensor for a given data item suffered accuracy problems due to ambient conditions the alternate type would not.
I think you'll find that integrated or close processing of ps data has been around for some time, though don't know if it's used in any current civil ac. Some of the benefits are the elimination of measurement delay caused by tube length, simplified unit replacement and, as you say, connection via standard ac data buses.
The probe issue still nags and nags - Most of the probes i've seen in the past have heaters but no temperature sensing and looking at the data sheets for the two AI fittings mentioned earlier, it looks like they are similar. Why is this of interest ? . Without sensing, it's not possible to have closed loop probe heat control. This means that the probes are getting full power all the time they are switched on. With 300 to 500 watts input typical, it wouldn't surprise me if the probes get almost cherry red hot on the ground, which can't be good for probe life. With a temperature sensor and a decent controller + enough heat capacity, it should be no problem to maintain temperature at optimum levels irespective of load and would also eliminate ground overheating. Otoh, perhaps they use a resistance element with a positive tempco, but it's still a bit primitive.
Iirc, simple 3 term pid controllers are a first year electronics undergrad subject and have been used in industry for decades...
Chris