I always kinda liked this one of a motorcycle engine with one end cut off. You can see how the valves work, and how they are designed to rotate under load, to even out the valve seat wear.
"The government" is virtually the only entity that has ever taken any interest in space flight. The cumulative government investment in space research, development, and operations is orders of magnitude higher than the the cumulative private investment, and even today in the supposed golden age of private spaceflight more than 80% of the money is coming from taxpayers.
The analysis conflicts with personal recollection. It seems to show leaving SF at 4pm and being able to get to roughly central Mountain View in an hour. Google Maps says that if I leave today at 4pm, that trip takes "typically 1h - 1h50m" which in my experience is about right. One hour is a good outcome, not a normal one. In a car, at rush hour, sometimes it could take an hour to go four blocks down Battery Street in SF.
Also, of course, the article has neglected other modes than cars. If you leave on a rush hour Acela train from Boston you'll be in Providence in 33 minutes, well outside even their 10pm driving radius for Boston.
I commute by car from Sunnyvale, CA to SF (Harrison/8th) almost every day, and I'm flexible with when I can drive there, so I started to track google maps time to find the better time.
1st wave starts at 5am and at the top of it at 8am then it's getting better.
there's a spot between 10:30 and noon when you can drive within 40 to 50 minutes, then it's getting worse (2nd wave).
if you would decide to drive at 5pm it would take almost 1.5 hours to drive to Sunnyvale
the best time to drive from SF to the Valley is after 8pm, there is a small increase in between 10pm and midnight - roadworks while traffic is still present.
it's interesting that the pattern is accurate up to the minute, it's 10:30, not 10:15 or 10:45 when there's a spot to drive, almost every day
Acela isn't exactly great comparatively as it's not an economically viable daily commute option. I don't know what it is for Boston to Providence, but DC to Baltimore is $40 one way in 25-30mins. Compared to 15-20 mins more for the commuter train at $7 one way, it hardly seems worth it.
Acela is really priced for expense accounts. Given that the entire Boston-NYC route was electrified for Acela anyway, it rarely makes sense for me as an individual to pay for Acela vs. the Regional. It saves maybe an hour and gives you a slightly more comfortable car but it's 2x the price.
I rarely take it but MBTA commuter rail has really gotten pricey. There are passes but I'm almost $30/day to commute into Boston by train between parking, the ticket, and the subway on the other end.
I have a feeling that Chicago will switch to a distance based tap on/off system at some point. It just makes more sense to understand where people are actually using the system, and also prevents train dwellers riding nonstop all day.
I don't know if anyone will care, but I saw Grand Rapids, MI mentioned so I thought I'd throw my 2 cents in. I like taking Amtrak to Chicago from Battle Creek, MI. Usually about 3 hours, and that's my preferred way to go to Chicago for the weekend or longer.
Day trips are more interesting because of the South Shore Line.
It's a electric rail that runs from South Bend, IN to Millennium Station in Chicago. Take something like a Cubs game. I drive 2 hours to the Dune Park station
From there I park for free and ride the train to Chicago and then hop the Red Line to Wrigley. Even if it's a night game I can still make it back to Millennium Station for the last train out at 12:45AM.
The South Shore Line is also great for conventions on the weekend because it stops right at McCormick Place.
Yeah, also it is not averaging well given geographic constraints, so the 4, 7, and 10 distances are the same(ish) for all directions except north and south and directly along i90, which makes their distribution seem oddly narrow
Portland's is also misleading. Getting off the freeway still involves massive long lines to get anywhere.
This is a city that just replaced the Sellwood Bridge, an aging bridge that was a huge bottle-neck because it was only a single lane both directions. They spent $324M to build a brand new Sellwood bridge that is still only a single lane both directions.
Tacoma street East of the bridge wasn't gonna get any wider no matter how many lanes wide they made the bridge (which is mostly one lane each way). It would make no sense to widen the bridge beyond the width of Tacoma.
The road to the east of the bridge narrows to one lane each direction for 100 yards, then returns to two lanes wide. And there is clearly room to make that hundred yards two lanes wide.
And worse, eventually PDX is going to wake up and realize they need to widen key surface streets to address their traffic hell. That will be expensive enough on it's own. But after you already blew $300M on a bridge to ensure it will always be the bottle-neck, it will be far more expensive.
And then returns to two lanes, rinse, repeat. Not sure how you think widening Tacoma would happen. You'd need to tear down most houses and businesses that line the street. I live in Sellwood, trust me, the bridge is not the bottleneck.
I've driven over it many times, it's clearly a bottle neck. And read my post again, the worst part is they spent hundreds of million on a bridge that will become THE bottleneck if they ever try to improve traffic flow in Sellwood itself.
From what I heard, the old bridge was pretty much falling apart (it had temporary girder braces and such), so it seemed like it needed to be replaced. Oregon in general has lots of aged bridges in need of repair or replacement.
Regarding Sellwood, I have heard that the new bike lanes on the bridge are an improvement. That is just second hand information though -- I don't ride on it.
The bike and pedestrian improvements are huge. The old bridge had no shoulders (death trap to ride) and one ~3' sidewalk on the North side that was shared between pedestrians and bikes heading West and East.
Edit to add photo...
You can see the old span alongside the new bridge here:
The bike/ped improvements haven't done anything for the massive traffic jam the bridge causes. Even at non peak hours, such as 2pm in the afternoon there can be a quarter mile of cars lined up waiting to get on it from the west side.
You're seeing both the old span (moved North from its original location) and the new bridge under construction in that photo. Once the new bridge was finished, the old span was torn down.
Agreed. It shows a far more optimistic view of how far one can get in 1 hour at 4pm in Atlanta, however my wife would consistently spend nearly 2 hours in traffic to travel 21.4 miles. Without traffic that is normally a 30 minute drive.
"Using billions of anonymous measurements from cell phones and vehicle sensors, Here Technologies, a location platform company, calculates how traffic conditions change throughout the day"
Emphasis on the word "billions". I find my own can be quite variable too. Large sample sizes should make that number more predictable.
More predictable does not imply more consistent. It just means you can nail down numbers like 10th and 90th percentile more easily. They'll still be far apart in many areas.
The 4pm travel time also changes based on day, afair. I used to make the commute from SF to Menlo Park every day, and Tue, Wed, Thurs were much worse than Mon or Fri.
Lost me at "And lots of people learned that monorepos are really painful, because past a certain size they just stop scaling." Plenty of counterexamples of monorepo projects much larger than Linux kernel.
Actually, my experience is that monorepos scale better than non-monorepos. Probably the largest project that isn't a monorepo is OpenJDK--and even then, that's only barely true, as the entire JDK (effectively, rt.jar) is itself a single repository.
If you're arguing it's not a (implied-to-be-evil) monorepo because, well, it's really managed as independent fiefs who do all their work in separate integration stages... well, that's basically how every other major monorepo project does it. So if Linux isn't a monorepo, then there's no such thing as a monorepo in practice.
