The general consensus is that they should be compressed somewhere closer to a 50% state of charge. My approach is that I compress them before I do my top balance process*. They arrive at 50% state of charge, so that works out well for me. Of course, I have to remove them from compression to transition them from the top balance configuration to an actual 16s battery. So is that bad? I dunno, but it's worked out for me now on my third DIY LiFePO4 battery.
WRT the Kapton tape, many of the off-the-shelf batteries are using Kapton tape or something similar to "compress" or hold the cells together. If Kapton tape was the only option, I would compress the cells with a clamping system, put on the Kapton tape and then remove the clamping system. I used a set of pony pipe clamps and aluminum plates to compress the cells during top balancing and I would do something similar to get the Kapton tape on.
IMO, springs are overkill, trying to satisfy the compression requirement after we've already handled 99.5% of the requirement. But I applaud the folks that have done the math and figured out which springs to use and exactly how tight to make them.
* - top balancing seems to be going out of vogue in favor of slapping the cells together and letting the BMS balance the cells. This assumes that you have a BMS that has enough balancing amps to do the job.
