Posts by BenWilson
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Hard News: Winning the RWC: it's complicated, in reply to Emma Hart,
"Don't be a dick about stuff you don't like." Because you never know what the personal, emotional associations are with the people who can hear you.
It's a good rule of thumb, but it's not universally applicable, just like all the "don't be a dick" rules. Sometimes, being a dick is the only way to move forward. Some issues are just a clash of dicks.
-
Seems like Quantum computing has matured since I last cared to puzzle with it. Interestingly from the wiki article:
There is a common misconception that quantum computers can solve NP-complete problems in polynomial time. That is not known to be true, and is generally suspected to be false
That suggests that it's actually not likely they will crack public key encryption after all. Interesting. But they still sound awesome.
-
Hard News: Thinking Digital, in reply to Rich of Observationz,
And both my published Spectrum games work on emulators that can be downloaded from a range of sites.
My old VIC20 games that I wrote are all sitting on an audiotape in a box, somewhere. I think they're better off there, really. If the tape has turned to spaghetti, it's only reflecting what the code on it was probably like.
-
Hard News: Thinking Digital, in reply to Russell Brown,
Virtual machines would be more straightforward, surely? Leo's current project is putting MAME -- and with it, every arcade machine ever -- on an old Xbox. I am quite excited about this.
Heh, yes, MAME is fun. I had it on my last PDA, an iPAQ, along with every arcade game. It was fun to play some of them again, although the main thing it brought home to me is actually just how far games have come. Also, the whole credits and lives thing is just gone, except in arcades. We don't like games that you actually have to win without endless go-back-10-seconds restarts the way they are now.
As for whether they are simpler, well that depends on the purpose. If you just want to read an old file, do you really want to boot up a 100 year old operating system and use a 100 year old word processor to do it? Then what are you going to do with it? Mail it to yourself using a 100 year old mail client, made for a 100 year old TCP/IP stack? Print it out using 100 year old drivers? I can just see people reveling in how shit Outlook was in 100 years. What the FUCK is that paper clip?
-
Hard News: Thinking Digital, in reply to Rich of Observationz,
. Most breaks into encryption systems result from incorrect "packaging" rather than frontal assault on the algorithm.
Indeed, although with a massive breakthrough in computing power, frontal assault becomes feasible. I don't know if the solution to that, if quantum computing becomes practical, is to just use longer keys. Seems likely to me that new algorithms would be needed. There are algorithms that are harder than even the hardest ones that we like to set as puzzles for cryptanalysts. And then there are harder still. I'm pretty sure that there's no hardest, just as there is no "biggest number", even in transfinite numbers.
So it could be an arms race. Or it could be that an algorithm is discovered that is not practically crackable at all, in any amount of time, on any kind of computer. Then cryptanalysis becomes a thing of the past. I hope that happens.
-
Hard News: Thinking Digital, in reply to Ian Dalziel,
when do we hit peak data storage overload?
Tricky question. At a technological level, storage has been growing at similar rates to computing power, ie exponentially. Which makes the storage of old data not particularly problematic in terms of capacity - every couple of years we make more storage than all the data in the world to that point.
However, we rapidly create data to fill that. So we always need more.
With current technology, there are theoretical physical limits to how much data can be stored, given space, just as there are theoretical limits to computing power, given energy inputs and space. Since we're increasing both exponentially (and it's a high base - doubles every 2 years) , we should hit that barrier. Hard to know exactly when. Could be 10 years, but you never know with technological breakthroughs - the theoretical limits are still incredibly high compared to where we are at now.
Also, the concept of quantum computing contains various promises that are mind boggling (to a computer scientist anyway). If they become practical, the impact is actually hard to fathom. A machine with 500 parts, that has processing speed and storage potential of 2^500 is something that it out of this universe. That's around 10^166 (extremely rough calc), which is around 10^66 times the number of atoms in the universe. Given how our computing works now, that kind of thing is simply ridiculous, bordering on totally unbelievable.
The ramifications of the existence of even one of these devices is huge - the entire Google cache could be stored virtually unnoticed in a corner of it. The combined computations of every humming server in the world today could be cranking away in it. Cracking current encryption would be an afterthought, although it seems likely to me that you would just make new encryption that used the entire capacity of the device, plus sufficient multipliers that make it impractical to crack. Not my field, but it certainly is possible to create mathematical problems that would tax even a device like that. Encryption will always be an arms race, no matter what technological breakthroughs are made. It's a great pity that such a thing occupies so many of the world's CPU cycles, to be honest - the arms races even eats our processing power.
Such futurism aside, the practical question of data overload is probably more dominated by administrative issues than technical ones. Most of the data in the world is unstructured, and this makes it effectively useless. Formats change gradually but constantly, invalidating old data. Essentially, there is a steadily rising base maintenance cost for keeping data in any kind of shape. This suggests that any data keeping organization will eventually become "data bound" expending nearly all of their resources just on maintaining data. This is ridiculous, which is why it doesn't happen. Instead, data is lost. Data will always be lost, even if it is stored, because data without appropriate conversion is meaningless. We have to work to keep our data meaningful, and we have to keep doing it constantly. There's only so much data that can possibly remain meaningful.
It's one of those things we have to accept - every moment of our lives, we forget 99.9% of our sensory input stream. Does it really matter? So long as we hold onto the main threads, the important bytes, the rest can sequester down to random bits, and most certainly will. That's life, superimposed into a digital universe.
-
Field Theory: All Blacks v Wallabies Tensionfest, in reply to Lucy Stewart,
Tell that to Jerome Kaino.
OK, I meant a big guy who is meant to score tries on account of their blistering speed and power. You train people differently depending on the work they are expected to do. That is why American footballers couldn't run a mile to save their lives, but are some of the fastest sprinters in the world.
-
Hard News: Thinking Digital, in reply to Paul Campbell,
if historians want to read word95 docs 100 years from now they will need to have working machines that run win95 and word95.
Well, they need something that can convert the files, anyway. The data storage medium is reasonably irrelevant if the data is being copied/backed up. If it isn't, it won't last 100 years.
But I agree about being able to read the doco now, wholeheartedly. Not so sure about PDFs, the format has been open and unencumbered since 2008.
The idea of all software used by the government being open source is quite a big ask - I'm sure they use an awful lot. But they could make it a mission to only develop that way from now on, and to push for it whenever it is a reasonable alternative to something they are using or going to use.
-
Hard News: Winning the RWC: it's complicated, in reply to Simon Grigg,
Talking to others, I know this sort if thing is not uncommon.
Yes, that is quite shit. I have a whole bunch of vague statements to fob off persistent inquiries, and changing the subject is the best tactic. It's annoying to have to do, but that kind of thing is by no means limited to rugby. Being asked how much I dig some kind of music that I don't know about is in the same bag. Or books. Or religion. Sometimes the safest is to just support them all. Or pick an obscure one, and claim it's what you're into.
Unless, of course, you like having arguments about the subject. Can't rule that out. I've been known to like arguing about some things.
ETA: An especially good fob is to claim that you get so into it, that you can't handle it any more. "Can't watch the All Blacks any more, get too tense, have to take my pills!"
-
Hard News: Winning the RWC: it's complicated, in reply to Russell Brown,
There's actually a massive e-waste problem with old CRT sets developing -- they're valueless and people just throw them out.
I'd throw mine out, if I could lift it.
Last ←Newer Page 1 … 500 501 502 503 504 … 1066 Older→ First