They probably check what's passing through the upstream filters, and is handled by a spamfilter.

My spam filter catches some 45-50 spams a day. It misses those with attachments (irritating - get 5-10 of those daily) and a few others that are hard to classify as spam (rather legit business related but not my business), but overall doing a decent job. I'm getting a similar number of legit mails a day, a large number of those being stuff like meetup and facebook status messages. So that'd be indeed about half/half for spam vs ham.

However, that 45-50 spam a day is AFTER greylisting, which basically requires every first-time sender (FROM, TO, source IP) to retry 5-30 minutes later. Many spambots just don't. I have this active for years, and when I installed it the amount of caught junk went down from some 300 a day to 30 a day. If this 10-fold difference is still the case, I'd be getting some 450-500 spam a day, or a 90% spam to 10% ham ratio. I'm not about to switch off greylisting to try this out, though :-)

How to measure (and slow down) sending rates of some bulk mailer, who sends e-mail to hosts all over the world?

The source doesn't have to follow the protocol - they just send out as fast as they can. The destinations (plural - probably big big numbers) have no idea what other destinations get from that source. Most of them will see just a few e-mails come in, that's their share. All the spammer has to do is not send everything for Yahoo at the same time, but intersperse it with mails to other destinations and Yahoo doesn't throttle him - or maybe Yahoo does throttle, but that won't stop the spammer from targeting other destinations in the meantime. They probably do that already, simply by having hundreds of outbound smtp sessions running at the same time.

You see how well that works for traditional paper junk mail, where the cost of sending out mailings, even delivered door to door, is easily an order of magnitude higher than the number you suggest.

It's totally non-existent thanks to this cost, right?

You mean - just like human drivers are tested in various conditions, like dry sunny weather, during heavy rain, with ice and snow and at night, just to make sure they perform well under those conditions?

Rear-ending is something that always comes up in /. discussions about driving, especially by US drivers (the site is rather US centric). Everyone and their dog seem to have been in at least one such accident. I have never been in such an accident, nor have I heard of any European friends that had such an accident.

Add to that, statistics show that US drivers have far more accidents, injuries and deaths per distance (per km or per mile, whatever you like to use) than European drivers, especially those from western European countries. This while US streets are wider and straighter; quite some Americans are scared stiff by our narrow, winding roads - we're routinely doing things like driving 80 km/hr (the legal limit) on country roads, and not slowing down for oncoming traffic while the road is so narrow there's not even a line in the middle... because the road simply is plenty wide enough for two cars.

Much stricter driving training does help a lot.

Interestingly you don't mention how much harder bad weather conditions make driving for human drivers, as well. There is a reason that many more than usual accidents happen when the weather is bad, when it's snowing, late at night (sleepy drivers - never heard about a robot getting sleepy), or when the roads are bad and human drivers think they know it all and can continue at top speeds.

Of course they start in good weather - that's also how you got your driving lessons. First make sure you can handle the good weather situations, then add bad weather into the mix.

It's old, it's pretty much done for, and preceded by many better protocols (some of which have also been seriously damaged since, like RC5). It starts to sound a bit like kicking a dead horse.

It's not about the technology, it's about the content of the sheets. That usually also needs updates now and then. New products, new (hopefully improved) sales procedures, etc.

It depends what the people in the company have done before.

When Henry Ford started building cars, he started off at what was then pretty much the state of the art. His main achievement is the mass production part. Car production was expensive, so cars were not for the masses, and Ford changed that. He figured out how to build a regular car (state of the art for his day) and turn it into something that's economically viable and cheap enough for the masses.

This company may well be populated with smart brains that have developed and built aircraft before. It can build on existing technology, supersonic flight is well understood. I don't think there are many technological hurdles to build a supersonic airliner (after all, it's been done before: Concorde). The trick is to make supersonic flight fuel-efficient so operating costs can be cut, as that's currently the big hurdle, and one of the main factors that eventually killed Concorde. Can they do that? Well, time will tell. But to say off the bat "they can't build supersonic because they didn't build anything before" is a poor argument, imho.

Economics have changed: the world population has risen a lot since the advent of Concorde and has become a lot richer as well.

More importantly though: technology has changed. Engines have improved - higher efficiency. Aerodynamics has improved - less drag. Materials have improved - less weight.

I wouldn't be surprised if you can build an aircraft now that can operate at similar speeds as Concorde, but at half or less of the cost. So if you can still sell the tickets at the price of Concorde, it may very well be economically viable.

That's where the FlashBlock extension comes in handy. No flash, unless you explicitly allow it to run.

So your corporate training programme itself hasn't been updated as well? It's just weird to think that 10 year old slides are still useful for training. As if time has stood still in other parts of the company as well. Doesn't sound good to me.

You have to manually search for where the toll roads are and think on how avoid them. Drag the marker around sure but will that give you the best alternative? Can't have it avoid them in the first place and have Google come up with a good alternative right away. Same for avoiding - or preferring - ferries. Also no way to set a preference for secondary roads vs motorways.

So OK those options kinda exist, but it's gone from automatic to manual. How's that better?

Another option that Google doesn't have is to search for fastest route vs. shortest route (often got really interesting results, albeit often totally impractical) vs. most economical route (that option is keeping in mind that cars run most efficient at 80-90 km/hr, so roads that let you travel at that speed use less fuel than motorways where you're doing something like 120-130 km/hr - could even set fuel economy for your car at different speeds); by selecting a slower but shorter route on a regular commute I could easily save 30-40% of fuel, and the 1 hr trip took maybe a few minutes longer.

In Netherlands they do the same for bicycles. All directions same time green! Great fun. Suddenly the intersection is full of cyclists (going in all possible directions), seconds later the intersection is clear, and somehow they manage not to hit one another. Also fun to navigate as cyclist. Added bonus: extra green light cycles for the bikes, cutting waiting time, and far fewer cyclists run red lights!

That's exactly how I learned traffic lights work in The Netherlands. Yet pedestrians you don't normally have to worry about, as they do not get a green light even when you're turning left at a green disk shaped light. Basic safety arrangement. Pedestrians get their own time slot.