4 ms·

While I agree that the compression breakthrough doesn't make much sense realistically, I think it makes a great premise for the show. The plot needs Richard to

by eduren 9y ago

While I agree that the compression breakthrough doesn't make much sense realistically, I think it makes a great premise for the show. The plot needs Richard to actually invent something revolutionary (and thus fail to get it off the ground time and time again), that is also simple enough for the audience to understand. Any algorithmic breakthrough that is real-world feasible would either need too much vague handwaving from the writing ("A brand new way to compute data!"). Or be close enough to reality that neither the layman nor the expert is happy (imagine the show having to explain why the big-O of Richard's new magic sorting algorithm matters). Compression is the logical choice because it's: 1) easier to explain why fitting more data into less space is significant (laymen already have some rough visualization of compression), 2) lends itself to being "black boxed" in the plot ("Look at all the companies we can build with this one breakthrough"), 3) just barely feasible enough that if some SV nerd actually did make the discovery, then events would play out roughly the same as the show depicts (winning techcrunch, getting funding)

monocasa 9y ago
And I'd add that most compressors make a lot of assumptions about the general case that reduce that real world compression ratio when compared to theoretical maximums. That's why stuff like kkrunchy can't be beat in their domains (first and last rule of optimization, "know your data inside and out!"). If someone was to stumble upon a heuristic that gives you noticeable gains for general compression as well, it could be close to as game changing as Richard's stuff is. Obviously we have a good handle on "theoretical maximums" thanks to information theory and so that end of the plot is BS, but looking past that embellishment the show isn't outside the realm of possibility.
thomastjeffery 9y ago
It's perfect, because there is one impossible plot device in the real world. The attention to detail used to accomplish that is very appealing to me. It's my very favorite aspect of the show.

by eduren 9y ago

monocasa 9y ago

And I'd add that most compressors make a lot of assumptions about the general case that reduce that real world compression ratio when compared to theoretical maximums. That's why stuff like kkrunchy can't be beat in their domains (first and last rule of optimization, "know your data inside and out!"). If someone was to stumble upon a heuristic that gives you noticeable gains for general compression as well, it could be close to as game changing as Richard's stuff is. Obviously we have a good handle on "theoretical maximums" thanks to information theory and so that end of the plot is BS, but looking past that embellishment the show isn't outside the realm of possibility.

thomastjeffery 9y ago

It's perfect, because there is one impossible plot device in the real world. The attention to detail used to accomplish that is very appealing to me. It's my very favorite aspect of the show.