This is great. Although looking at the UK map makes me sad, all those abandoned lines.
It makes you wonder what the country would be like if we'd leaned into the rail system instead of tearing up a chunck of it following the Beaching Report. The most interesting ones for me are the additional lines that crossed the Pennines, linking the north together much better than it is today.
I know there's a lot of (very justified) vilification of Beeching, and Marples who was transport secretary at the time.
But actually Britain had leaned in hard to the railways, by undertaking a large modernisation scheme [1]. The problem is, they did it a bit too early, in the 1950s, when electricity seemed like a risky bet so they stuck with steam which was the UK's core strength.
Of course, that turned out to be a mistake in retrospect. In the 1960s, investing in the railways in any form at all would've seemed like making the same sort of mistake as just a few years ago.
One interesting analysis I saw about the Beeching cuts said that the worse sin is that they did not preserve rights of way i.e. when closing a line they should have built a road and run a bus service (as
Beeching's report recommended) or at least preserved it for potential rail reopening, but in many cases they sold the land off to developers so the chance was lost.
Ian Hislop, the editor of Private Eye magazine and regular star of the satirical TV show Have I Got News for You, made a documentary 12 or so years ago - Ian Hislop Goes Off the Rails - which looked at the Beeching cuts and, more broadly, the history of the UK rail network. It's nicely done, and never descends into an anguished, hand-wringing rant. Sadly, it's no longer available on BBC iPlayer, but it's on YouTube (split into six episodes):
The northeast US is also dense with abandoned lines that can never be rebuilt. I assume they are not on the map because they've been built on top of or otherwise completely ripped up.
Rails to trails is actually a (somewhat subversive) way for rail right-of-way to be legally preserved.
From a practical/political standpoint, it'll be interesting to see the backlash when a railroad attempts to start exercising its right to reclaim some of these properties, which are largely and falsely believed to have become community property.
> Although looking at the UK map makes me sad, all those abandoned lines.
Germany is just as bad. Between 1994 (the begin of the privatization) and now we lost 6.200km of rail infrastructure [1] - something like 15-ish percent.
My s/o and I went on a two week railroad trip through the whole of Germany. Seeing rotting or half-ass dismantled rails, former shift yards and smashed-in former other infrastructure just hurted to watch. Everything has gone downhill, there was barely any investment in upkeep of tracks and buildings, only in new construction of expensive billions-euros high speed rail... the reason is simple: DB Netz, the privatized infrastructure operator, has to pay for upkeep on their own (from usage fees) while the federal and state governments (and in some cases like the Munich S-Bahn also the counties and cities) pay for all the high-profile projects.
We had so, so many industrial areas served by rail as well: in 1994, over 11.000 companies and industrial zones had their own railway attachment - today, it's barely 2.300, a reduction of 80% [2]. No wonder our highways and side roads are overcrowded with trucks. The reason for that is that unlike the US, Europe still has chain-link and buffer couplers, which means shunting yards are extremely staff-intensive to run, and shunting from and to the industrial areas is expensive as well... which means that, thanks to privatization, DB Cargo was more or less forced to shut down the industrial zone supply because the absurd losses this branch brought in could no longer be cross-subsidized by passenger rail.
Not much different. Service frequencies and speeds on those lines we’re not suitable for modern day, without large capacity increase in cities they couldn’t really be increased.
Most of the closed lines were seriously seriously unviable and had really poor service frequencies.
There was definitely mistakes (Pennines) but overall most of the lines would have had to be closed at some point. They often connected tiny towns/villages by circuitous routes, often with stations ages from the actual place they were meant to be serving.
Not everything! While it has rail lines, and abandoned rail lines, and surface light rail (streetcar/tram/trolley) lines, it is missing abandoned streetcar/trolley lines. (They existed in a lot of North American cities until the 40s-50s ish)
Despite that minor absence, the map you linked is clearly superior
While I appreciate the sentiment, the OSM wiki specifically says[1] "In locations where the railway has been replaced by new buildings and roads, the mapping of such features becomes out of scope for OpenStreetMap. Historical mapping can occur on Open Historical Map."
(Just checked - OHM doesn't seem to have any data whatsoever - present or historical. Maybe I don't know how to use the interface, but all I see is land/water borders)
Right. In the two cities I have personal experience with, the tracks and/or sleepers are still there, just paved over. They are evident (and briefly verifiable) when potholes develop or roadwork excavations go into those areas.
Regardless of whether it's more suited to OSM or OHM, I just think it would be a neat feature to have on Open Railway Map.
That's really interesting. In the UK you can cross reference with the National Library of Scotland maps[1] which allows you to overlay quite a few different maps of an area over different time periods.
[1] London Kings Cross area with large rail yard to the north. By changing the transparency you can see things like ST Martins College and the new Coal Drops Yard development. The new British Library is also built over Somers Town Goods Depot.
Me too. Clickbait Case's Fault Really. (But that's present in the submission, it's not just HN's auto-clickbaitiser. (But I'll take the opportunity to say I wish it wouldn't do that anyway.))
The title reminds me of an oft-cited anecdote of how someone running an online shop of this or that sort can tell that an order is from the US — it will be the only one where the provided address doesn't include the country.
On a similar note, it's also a common cliché in the context of international group introductions:
"Hi, I'm Olaf and I'm from Lund, Sweden"
"I'm Francisca and I come from Córdoba, Argentina"
"My name is Jayalalithaa and I'm from Erode, India"
"Hey everybody, my name's Peter but my friends call me Pete. I'm from Columbus, Georgia, and I'm super excited to be here"
The population of other national subdivisions is greater than that of Sweden and of most US states. That doesn't make sensible to introduce oneself with "I'm from Minas Gerais", "I come from North Rhine-Westphalia" or "I'm from Kaduna State" in an international context.
It might make sense if Francisca had just said "from Córdoba" and Jayalalithaa "from Erode, Tamil Nadu". Since everyone was talking about countries rather than states, one would reasonably expect Peter to do the same, and thus assume he is from Georgia rather than the USA.
On the other hand, for example every country's stamps say which country issued the stamps... except the UK. Our stamps just have a silhouette of Liz on them (eventually they'll have a picture of Charlie instead) because that's just how it's always been.
Interesting, I looked at the collection that's from (at the Smithsonian) and it didn't have any more in the pages I scanned. I wonder what the story is.
My American sister in law does this when she's posting stuff from America to her non-American relatives. Her parcels tend to bounce around America until some postal worker figures out she meant the original Worcester, or wherever.
For fun I once send a postcard to my parents from Spain to Germany without a ZIP code, without the country, misspelling the town and street, and writing "the last house on the right" (in German) as a house number.
I worked a summer as a postal carrier in the US, and I was surprised at how far off from the official address standard you can get and still get delivered.
For reference, our address standard is ## Street Name,<cr> City, State, ZIP.
There was one neighborhood in particular that had a name like "Edgeview Cottages", and instead of getting letters addressed to 20 Apple St, you'd see Edgeview #20. It's crazy how people expect the post office to be able to figure this out.
Embarrassingly, I didn't believe it when I was told USPS OCR rejects less than 5% of hand-written address. I assumed that even if you could read the writing, the format would be too irregular to parse a lot of the time.
Part of the reason I was wrong is that so much of the address is redundant. I think they only use four letters of the street name, for example, before they essentially always can narrow it down to one possibility in their database.
A quick check shows plenty of rail lines that haven't existed in several decades: the tracks lifted and the lands sold off long ago. Not very useful if you're looking for information.
Does anyone know of any current or past projects (AI or not) that measures how many miles of train carriages pass through a particular intersection or particular line?
I know there's an API for AMTrak, but anything similar for commercial trains/lines?
I'm often stopped by trains, (what feel like) over 10 miles long, doing 30-50mph. Some carriages carry uncovered coal, and are left parked near my neighbourhood for days.
I'd love to track how many miles of carriages, and what type of carriages, pass through my neck of the woods, either with something like a Vizycam, and train it to count the carriages, or even by sound or vibration, and get a basic count, at what times, what dB, etc.
The railfan forums often have very knowledgeable people, and many trains are run on schedules (our local track has a train one direction in morning and the other in evening). If you find someone who railfans in your area they probably know most of what you want.
It shouldn't be hard to count cars if you wanted to, either by sound or laser or something similar. People have even done it by earth vibration.
Not what you asked for but I've seen recent articles and videos stating that trains are getting longer, which means more time consumed by those waiting at crossings.
Just a footnote: this includes a lot of lines that don't exist, have been abandoned, or repurposed.
As an example in Kansas, the steel rail lines from Osawatomie to Osage City have long been ripped up and scrapped. They are now rails-to-trails projects. Several other of the railroads marked are the same way (I know because well I visit these places quite often). Great use of the land btw, and makes for some beautiful passage across the state.
Connecting multiple carriages together to form a train was an amazing innovation - those carriages are separated by inches, whereas trains are separated by miles.
If everything could be separated by inches, we could have a network with 1000x the throughput. 1000x the throughput means we can use all the existing (expensive) rails, and be able to transport everything the interstate road network does today.
And, with enough innovation, that's possible. Imagine an infinitely long train. If more cargo needed to be added to the train, while in motion at say 60 mph, one link in the train could be broken, the gap widened to fit an extra carriage, and then an extra carriage comes in from a side rail, just like merging on a highway, and the whole lot is reattached again. All while still moving.
Infinitely long trains already exist - for example an escalator is a continuous 'train' of steps in a loop.
Disconnecting trains while in motion already exists - it was done regularly on the british railway 50 years ago.
And the rest is just a coordination and timing problem - one that computers are really good at.
Eh, not really that simple. The coupling forces with very long trains can cause the trains to break apart. You need to distribute locomotives throughout the train to provide power not just insert new cars. The freight railroads are already running very long trains based on computer modeling and they’re getting pushback from the govt. https://www.trains.com/trn/news-reviews/news-wire/railroads-...
Rail doesn't need technological pipe dreams. A well run rail network is already really good at moving freight and passengers.
The issue in the US is decades of cost cutting by the freight operators. This includes a lack of infrastructure investment, optimizing cost over speed, and overworking employees to an extreme amount.
The technologies for better rail already exist. We just need to actually use them in America.
> If everything could be separated by inches, we could have a network with 1000x the throughput. 1000x the throughput means we can use all the existing (expensive) rails, and be able to transport everything the interstate road network does today.
The problem: most rail runs above ground, with level road crossings. Running trains at the same spatial density as cars would massively impact wildlife, and you'd need to rebuild all the road crossings as well. Additionally, you'd need to reconstruct a lot of bridges and other infrastructure as it was never built to handle that amount of load and traffic.
Another limiting factor is that switches need time to operate, you don't want to end up with the tip of the train going straight and the rear part of the train going off. That requires a further minimum distance between trains.
Edit: It's not just the US (!) Maybe a 'North American rail network map'?