Just to be clear, you, like R Mackey, Myriad, Newton's Bit, and Dave Rogers are writing about BV, BL and BLGB, published in 2007 and 2008, is that true?
I for one am not writing (present tense) about any of those things. I did so years ago; it's no longer relevant. The ship has sailed, reached port, sailed again, reached a different port, sailed, made a thousand more ocean crossings, been de-commissioned, and broken up for scrap.
Just for some historical perspective, though, I recall only sporadic discussion of
any specific mechanism of collapse progression, in terms of what members or connections failed first at each floor, prior to about 2007. There was no argument going on between "column-firsters" and "floor-firsters" or anything remotely resembling that. The argument was between conspiracy skeptics and the prevailing claim by truthers that progressive collapse of the towers would have arrested part way down without the involvement of some additional unreported sabotage of the structure (usually, demolition charges but other methods of weakening the structure were proposed) making the lack of any such arrest proof of said sabotage. The floor truss to column connections were occasionally mentioned as a likely point of failure. However, the emphasis was on proving that arrest was impossible, and doing so requires that even the best case for arrest fails. That best case involved the static and dynamic loads from the upper structure somehow directly transferring to intact columns below, which process, unlikely as it always seemed, was either explicitly claimed or tacitly assumed (e.g. by Szamboti's "missing jolt" claim) in a number of truther publications.
In 2007, NIST published the first "supplemental FAQ" to NCSTAR (which is archived
here). The first item in that document offered a concise argument against collapse arrest that was completely independent of Bazant's, disregarding the energy balance of column buckling and focusing instead on floor loading from collapsing or collapsed structure versus floor load capacity. Note that by doing so, NIST's answer also disregarded the possibility of falling loads bearing directly on intact column ends below without loading any floors. That brought floor failure more prominently into the discussion.
The NIST FAQ's argument against collapse arrest (impossible because no floor can bear the load of the rubble from above) does not explicitly describe a complete progressive collapse scenario of floors progressively shattering or shearing off leaving columns unbraced to fail a bit later. But it very clearly implies one. If the collapse can't arrest because the floors
would shatter or shear off under the load, the possibility that the floors
did shatter or shear off under the load follows practically automatically. That possibility had been discussed earlier, but as I said, the NIST FAQ brought it more into the discussion, often when truthers objected to because it disregarded column-end impacts. (By contrast, it was obvious from the start that Bazant's best-case collapse arrest mechanism of every column segment folding into thirds with 180-degree bends never actually happened; any of thousands of photographs of the rubble make that clear.)
Keep in mind that many truthers' arguments were still based on describing the collapse as an impact between two large sturdy structures, the "upper block" and the "lower block," somewhat like a vehicle collision. Bazant's model still represents a good argument that, even if that were the case, collapse would not arrest. The NIST FAQ argument, and the scenario it implies, is more realistic but less airtight as an argument against an expectation of collapse arrest.
My own thinking about actual collapse progression mechanisms, as of a year before the start of this thread, is indicated by my proposed
intermediate-scale physical model of progressive collapse. That model was explicitly designed to collapse by severing of floor-to-column connections from the load or impact of upper floors. It's those floor-to-column connections whose strength would have to be calibrated, as described and illustrated in the design, to create the demonstrated effect. The columns in that model design at that scale, or any realistic-looking columns at that scale, would not be expected to break at all, which is why continuous columns weren't used for the model (since the terms of the challenge required the entire structure to collapse to the ground). It is, in short, a design for a physical model of runaway floor-tray destruction by progressive severing of floor-to-column connections.
Somehow, I managed to come up with that design a year prior to the OP of this thread which supposedly enlightened us all about that possibility.
Given the above-linked NIST (2007) and Myriad (2009), which are only the easiest for me to find of numerous other examples, I don't think M_T's claims to priority for the OOS model (except for the name), or his insinuations that skeptics were blindly adhering to the wrong model prior to that, hold any water at all.
