Well what would be closer would probably be to use cpole~ and use the second article's values for frequency. Then you feed cpole~ some kind of pulse (not sure what kind, band limited, dirac, rectangle, maybe katja's pulse generator), and then convert polar coordinates to real and imaginary parts. For polar phase, you exponentially (using ead~ or something) ramp the phase of cpole~ corresponding to the frequency at that moment. the polar magnitude (constant) is the "decay time". Then take the real part mebe?
edit: see above for an example by katja (the resonator setup)
however this would only be an approximation because the bridged t thing seems to be more like a sharp band-pass filter (which cpole~ is the simplest version of). instead of using cpole~ really the bridged-t filter should be modeled and used. http://music.columbia.edu/pipermail/music-dsp/2005-October/064422.html
maybe vcf~ with high q would be more appropriate? anyway a look at the transfer function of the 808 filter and a bilinear transform seems like the kinda thing to do, i know nothing of analog filters and that stuff...
you can also use ead~ for everything, possibly (frequency and amplitude).
for modifying efficient-kick it would be closer to square or raise ^4 the amplitude of the vline~ before multiplying the cos~