Splendid point about seed generation - it's what made the original 8 bit versions work the way they did. Storing that planet data in memory alone just wasn't possible. Truly an amazing feet given the memory those poor little beasties had. Anyone interested should check youtube and look for the John Snow video named "The Making of ELITE (Computer Videogame)". It's a really, really good watch.If you weren't handling data for more than one player, weren't processing the BGS at all, etc. the "offline" version of the game wouldn't be noticeably bigger on disk than the current online version.
Incorrect - it's much more abstract than that. It would take that much to store the entire galaxy in a conventional database, but they don't - one of the most impressive Frontier achievements is that generation as needed from a tiny amount of "seed" data is fast enough to look like it's all there already.
Set your bandwidth monitor on and try doing a 20kLY route plot. The bandwidth use will be in the routine bytes/sec of "just playing the game" (assuming there's no other players about), the route will plot in a matter of seconds having passed through millions of stars worth of space, and you can even do things like apply a star-class filter to it to make it have to generate more than just "a system exists at this position" without really slowing it down.
No data transferred from servers, no gigabytes cached and retrieved from disk, and all with fairly low CPU and memory usage, too. Really impressive.
Remember that Terabytes are huge. If you're just storing text-style data, mostly numbers, the occasional string of words, even getting into gigabytes is tricky.
A complete EDDB data dump for inhabited space - and that's provided in a fairly inefficient uncompressed format for ease of use - takes less than a single gigabyte. Even accounting for various hidden variables and the markets for all the new Odyssey settlements EDDB doesn't know about yet, "non-player" data isn't going to be more than a couple of gigabytes total.
Once you add in players it gets a bit bigger - >12 million accounts potentially generate quite a bit of data, but they'd still be able to have 100kb each before the data got to a terabyte - a very active account probably has something like that or even more from exploration data, but the majority won't. Of course, for an offline version, you only need your own data and a couple of megabytes won't matter much. They might be up to a terabyte or two by now, but I wouldn't be surprised if it was a lot less.
Completely 100% agree about what an offline mode would require - in that, from our side, it's little data. The only thing I doubt is de-coupling the current code base. Whether FD decided they didn't want offline mode because they wanted to 'guard' their code/server data, or that technically it didn't fit in with how they were developing - is irrelevant (for us). It's just that we're stuck with what we've got. I suppose what I am saying is it's less about the TB involved here and more about the wrapping of code that manages that data - it's been built for that in mind, if you see what I am saying.
Worse: let's assume that, on paper, a team could sort this out in a very short (== cost effective) way - it's really down to FD. Which is why it didn't happen from the start, and even less likely to happen now.
So, theoretical ease will never be trumped by the practicalities involved and at the top we have the locked gate that is the decisions of FD. And I want it known that this breaks my heart because, heck, I would love an offline mode!