When the partition was larger (500GB) there was enough space to store the entire TP for 8 days. I'm wondering if the skip ahead logic also parses down the full transponder to individual streams per channel and discards other other channels besides the big 4 to conserve space.
Since we don't have access to the code, we can only guess.
Do we really know the partition was ever a full 500GB, or if it was that they were ever using it all?
I have never believed they are capturing the whole transponder as a single stream since getting the Hopper. It has never behaved that way and doesn't make much sense. They have to be able to demux in real time to watch the PTA channels live anyway.
What makes more sense - developing the code to reliably handle trick play of a separate "full transponder" file format that has to demuxed while playing, or going ahead and demuxing as it comes in, storing the individual streams, and using the proven trick play code.
Consider the scenario during PTAT with PTAT and the other two tuners all recording. All Joeys and Hoppers are watching separate PTAT programs from earlier in the night. Is it more efficient at playback to read 4(+) x 40mbps streams and demux for each display, or read already de-muxed 6mbps streams directly? Which would be easier to implement?
As you said, all we can do is guess until someone is curious enough to crack the case and look at the drive contents.