Cable length

Uh oh! C'mon Sean, I'm not too sure about your explanation of this, and I'm tired. But the biggest issue re the problem with longer speaker cables is the fact that the amp sees them as part of the passive crossover network of your speakers! If there's even a small impedence change then the crossover will shift, resulting in a spectral coloration.
If you run XLRs you can run HUNDREDS of feet between a robust CDP and a great, transparent pre with NO issues. The recording pros do it all the time! I chose to run longer, asymmetric (yikes, eh?) XLRs between pre and monos, and shorter speaker cables. Glad I did. I'm forgetting some of the other reasons for this, but it also has to do with the source component's signal being voltage-driven, while the amp's issues in driving lines and crossover networks and drivers is a current thing, which is therefore more sensitive to cable length loading. Recording engineers ALWAYS run long mic cables to pres rather than long speaker cables. Can we get an engineer in here who hasn'y forgotten what he learned in a marijuana haze back in school in the 70's (me!)??

I would guess in Simon9's case there would not be much difference as the two situations are not much different. Sean makes good points:
Along the same lines, if the signal is distorted due the materials used to construct the cables themselves OR the cable is of a poor design geometry resulting in improper impedances OR outside influences such as RFI and EMI enter into the equation, those distortions will be amplified further down the chain. It is not uncommon for problems with RFI / EMI to manifest themselves through interconnects due to broken grounds, poor design or poor connections. These problems are FAR less likely to arise with speaker cables unless using highly reactive ( inductive or capacitive ) cables or drastically long lengths.

To top it all off, interconnects use small gauge wires.

But it seems that if an interconnect is WELL designed (shielded etc.)and of sufficiently low resistance then long runs are okay. Have any measurements been done to show actual loss of these low level signals?
Shorter speakers cables would better damp the speaker drivers and while less immune to RFI, wouldn't the higher current in speaker cable make transients, phase shift, etc. more effect of the cable than low level signals in I/C's?
Well, I know what I mean, just can't say it :-(

Check this out:

http://www.tnt-audio.com/clinica/wirecalculator.xls

and this

http://www.tnt-audio.com/clinica/cat5questions_e.html

for some info on the affeexts of cables length etc...

These are out of Revel Speaker manuals:

NOTE: “Loop resistance” is the
DC resistance measured at one
end of a loudspeaker cable, with
the other ends shorted together.
Use high-quality speaker cable with a maximum total loop resistance
(for each wire run) of 0.07 ohms or less. Use the chart in
Figure 14 below to determine the maximum wire gauge.

A high loop resistance will result in mis-termination of the
filter networks, and serious degradation of sound quality.
Use high-quality connectors on the loudspeaker cables.

WIRE GAUGE (AWG) LENGTH (FEET) LENGTH (METERS)
6 87 27
7 69 21
8 58 18
9 43 13
10 34 10
11 27 8
12 22 7
13 17 5
14 14 4
15 11 3
16 9 3
17 7 2
18 5 2

We are talking about nano seconds of difference in electrical signals and until someone can validate the human brains ability to differentiate these miniscule packets of information I would prefer to put a small pillow behind my bald head and change the sound of my system that way or cup my hands near my ears as my wife shakes her incredulous head.