Comment Re:ESA link is borked for me. (Score 1) 97
Glad (kinda, in some perverse manner) that I'm not the only one.
Glad (kinda, in some perverse manner) that I'm not the only one.
the flight takes 14 hours,
10 hours, nonstop.
Turn trans-Pacific/Atlantic into a weekend trip instead of the current 3 days of travel time and there's a market for it.
The Concorde turned a 6 hour flight across the Atlantic into a 3 hour flight. Why, then, was the Concorde economically unfeasible? Cost
Those sub-orbital flights will cost a lot more than the Concorde flights. People will say, "$1200 for a 10 hour flight, or $5000 for a 4 hour flight?" Sure, a handful will pick the $5000 ticket ($20,000 when you add in spouse and a couple of children) but most will say, "4 hours is not worth $15,200."
Trying to balance a big pencil on a postage stamp that's moving unpredictably and simultaneously in 4 axises (pitch, roll, yaw, altitude) doesn't seem to have very high odds of success. And the worse the sea is running, the lower the odds.
If it works, though, count me really impressed by what would surely be a Crowning Moment of Awesome.
Increasingly, companies want to make a "me too" product or do things based on what focus groups tell them is good.
We won't mention that vertically-launched multi-stage rockets powered by RP-1/LOX are pretty "me too".
If you want a repairable computer with a separate chip for every application, I have a coal plant to sell you
Or an original IBM PC/XT. All the chips were socketed, and every IO device (except maybe the KB) required it's own card on the ISA bus.
Yearly major releases and more frequent bug fixes was a solved problem long before billg co-founded the Evil Empire...
if a fire starts in your home, you want the system to respond appropriately in milliseconds,
Given the time it takes for people to become alert from a deep sleep, the time it takes a telephone to be automatically "dialed" and connected, and the fired trucks to arrive, ISTM that a 5ms response to a fire alarm signal is no more needed than a 500ms response. OTOH if houses were nuclear powered or flew at 1000 MPH then 5ms response times might actually be considered slow...
holding back new research tools because amateurs and politically motivated groups could misuse them is very scary indeed.
An analogy: we hold back guns from four year olds -- even when we show it to them and say, "Very dangerous! Never touch!", but not from legally competent adults; when said four year old gets his hands on a gun, bad things can happen.
Likewise, we should not hold back *copies* of data from the world. However, so as to protect the "chain of provenance", edit privileges should be limited in some way, so as to prevent abuse by sock puppets and the anonymous. Maybe something as simple as requiring editors to log in using a cryptographic certificate signed by a trusted third party which requires some form of official ID and manual verification.
(1) Wikidata would either have to keep (many) multiple copies of possibly quite large data sets, or keep diffs. How much of a strain does it put on a busy server to generate a dataset from a huge original and lots of large diffs.
(2) Not too many people pay attention to Wikipedia changelogs. If only the current form of the data is easily visible, that's what most people -- especially amateurs and those with political motivations -- will use.
We've identified many deep problems with scientific research on this very forum
Most revolving around laziness and academic corruption. Allowing data (for example: historical weather gauge readings, or IQ scores, or any other data having to do with hot-button topics) to be edited is an invitation to socio-political fraud on an unheard-of scale.
I can't be the only one who thinks that is a terribly bad idea... It would rip the guts right out of repeatability, and confidence that "this" is what $RESEARCHER found.
"How fast does each individual core have to be to run a single threaded Word Processor at an acceptable speed?"
WordPerfect 6.0 ran great on a 286 w/ 640KB, and WordStar was zippy on a 4MHz Z80 with 64KB (it was the floppy disk IO that hurt performance).
So... the answer to your question is: not very!
And of course, every HTML document is a tree, and any tree can be parallelised.
No one has answered the question about disk and network slowness.
What's the human-perceived benefit of rewriting Firefox to get a 1/2 second speedup in page rendering when I'm still waiting 3-4 seconds for some ad server to send me the rest of it's crud (ABP needing to be blocked so that videos on ESPN will play)?
First thought: Why the hell aren't those two (total of) 10,000 page "books" split into their constituent 50 "actual" books?
That's the kind of parallelization and work optimization that needs to take place before algorithm changes.
indexing and searching get pretty easy boosts from parallelization.
How much indexing and searching does Joe User do? And what percent is already done on a high-core-count server where parallel algorithms have already been implemented in the programs running on that kit?
Web browsers especially love tiled parallel rendering
Presuming that just a single tab on a single page is open, how CPU bound are web browsers running on modern 3GHz kit? Or are they really IO (disk and network) bound?
Those who can, do; those who can't, write. Those who can't write work for the Bell Labs Record.
