Luciano Special Technical Collaboration Challenge - August, 2017

Posted by Luciano on 30 July 2017 in Korean (한국어).

I’m sure many of you have noticed, I no longer post my monthly challenges. I’m sorry about that, to those who miss them. In fact, I had hoped the community would take up the torch and run with it, but it hasn’t really happened. The closest has been what niels20020 did with the Weekly Word concept - it has been a nice effort and some great results.

This challenge is different. First of all, it’s an invitation for collaboration in my country, Ardisphere, rather than a challenge for you own work. It’s also quite detailed and difficult. It’s for ADVANCED mappers only, and especially for those obsessed with railroads.

Some of you are quite good and knowledgeable about railroads. I’m not. I enjoy thinking about and mapping railroads, but I am not any kind of technical expert.

As you know, I have been placing contours on parts of the Ardisphere. Once contours are in place, I can start to think about very realistic railroads and highways (i.e. “engineered” to match slope, terrain, etc.).

I want to have a freight / passenger line between Las Vegas, DP, and Narao, DP.

I built one a few months ago, but it was “over-engineered” and was quite baroque and implausible, and it required some excessive modifications to the contours to make it “fit.” I have deleted this draft (you might still see it on some lower-number zooms on the TOPO layer, which seems to update rather inconsistently.

Your challenge, should you wish to try, is to design a contour-realistic railroad for this stretch (see map).

Actually, I’m not very well-educated about what realistic standards would be for a 1910’s era railroad through this kind of terrain. So that’s part of the challenge, too.

In fact, this whole challenge is Thilo’s idea, more than mine, which he suggested after providing much-needed criticism of the original railroad draft I built.

Challenge summary:

1) Research what realistic standards would be for a 1910’s era railroad (freight and/or passenger, if those are different). There needs to be thought about curve radii, maximum grades (inclines), maximum tunnel length, maximum cutting depths, maximum bridge lengths and heights, etc.

2) The route MUST include Ciudad Sokolov (about halfway, near the western border). Other towns to try to include would be Cuaquensenutes, Paso del Miércoles de Ceniza, and La Apoteosis.

2) Use the TOPO layer imagery in JOSM to design the railroad (if you don’t know how to use the imagery layer, I suppose part of the challenge would be to figure it out - as I said, this is an ADVANCED challenge).

3) As you draw the railroad, record elevation, incline, etc on the ways and nodes. You are free to download some of the other railroads in the Comala region that I’ve built to see how I do this. I recommend the Comala-Yahuas stretch, which I’m pretty happy with.

4) DO NOT UPLOAD your work - this could create problems both because there may be more than one person working on it, and because I may be working in the region on other things, and conflicts are difficult and unpleasant to deal with. SAVE your work, and upload the OSM file somewhere online so that Thilo and I can look at it in JOSM.

5) Results will be announced in September. Prizes: a) satisfaction knowing you’re the best railroad designer in OGF-land; b) a certificate of recognition from the Ferrocarriles Federales (S.A.-Ardesfera) Oficina de Ingenieria, dated 1912, and signed by Presidente Selenio Quiroga.

Happy mapping.

Comment from Leowezy on 30 July 2017 at 07:52

Great. Just the distraction I need from my uni exams xD I’ll see what I can come up with. I think you already gave quite a lot of background info, I’d just like to make sure that the railway was intended as a main line back in the 1910’s and built to the “highest standard of that time”? Or was it a minor side route even back then?

Comment from zhenkang on 30 July 2017 at 08:15

That is quite tough, I wonder if you can draw contour lines on JOSM first though. I thought of doing the same for my country, with Kasai rainforest at the western border sharing with abandoned Troipoline, but after developing my cities and reworking the road systems in Singkangia.

Does railways needs 2 parallel rails or one?

Comment from histor on 30 July 2017 at 08:36

Most existing railroad lines are build before 1910. Or in other words - all main-lines are build first, so that the busiest railroads are this, which are build as the first lines - logical!

In 1847 the main line Hamburg-Hannover (Germany) was build with a minimum radius of curves with 701 m and a bed of 28 englisch feet (for the second track). In the 1850ies and 1860ies was layed the second track all the line long. So 1910 is a very late point of design. O.k. - between Hamburg and Hannover most is flat land, but some hills the line could not avoid.

Comment from Luciano on 30 July 2017 at 08:55

@Leowezy: This is intended to be a “main line” standard. However, it’s a bit of a “national vanity” project - I’m not sure the traffic was there - it was more about asserting sovereignty in the region.

So, yes, “highest standard of the time.” But not parallel tracks - it’s meant mostly as a freight line, and given the topography parallel tracks in that era would have been overkill.

@zhenkang: I’m not sure what your question is. I draw all my contours in JOSM (see That’s the only way I know how to do them. I think Paxtar has a different way of working (something raster based), and other users might be doing other things. Thilo’s conversion program converts JOSM to raster, so a raster system might be smarter, but I have no idea what tool I would use for that.

@histor: 1910 is only late from the European perspective. The vast majority of major Latin American railway projects were built in the first half of the 20th century, e.g. Argentina: in 1890, the system had less than 10000km; by 1920, the system had almost 50000km. Since Ardisphere is vaguely Latin American in character, I was imagining this sort of timeline.

Actually, if anyone doing this challenge would prefer to imagine an earlier timeline, I have no problem with it - I picked 1910 more or less at random, based on my knowledge of Mexican and Argentine/Chilean history.

Comment from Ūdilugbulgidħū on 30 July 2017 at 18:18

Great challenge! I don’t have great technical railway knowledge either, but fascinating looking at railways that might cover similar terrain as Comala. I’m thinking the Salta-Antofagasta line connecting Argentina & Chile (map). That might ultimately go over the same sort of ground and date from around the same time you envisage? I guess there would also have to be a good reason to build a line here in the Ardisphere. The Cuaresma gorge between Vegas and Cuaquesenutes looks particularly challenging. How are you planning on facilitate collaboration - it wouldn’t make sense for different people to draw their own line in its entirety? Will you set up a collaboration page in the wiki?

Comment from joschi81 on 30 July 2017 at 19:23

Hi Luciano, quite interesting, your challenge! Like for @histor, the “detailed” part of my railway knowledge is based upon the European standards and timeline. It seems also that European railways (down to detailed articles about single lines) are much more common for European railways at Wikipedia at the moment. I have difficulties finding what I’m looking for in detail.

By the way, here’s a Wikipedia list of adhesion railways with steepest gradients. Indeed, the Lisbon tramway ride up to Chiado district on Calçada de São Francisco is an amazing experience!

Ok, but this does not help much here. The basic question is the one from @Leowezy: What standards do you want your railway line from the 1910s to have? In my opinion it must be a single-track main line, with some double-tracked sections and many passing loops at stations or between stations.

And certainly without zig-zags. It would be realistic, maybe, to have some “special features” like zig-zags or sections with steeper gradients where additional bank engines are needed. But maybe there’s a way to prevent this, even within the most complicated section between Las Vegas and Cuaquensenutes.

I share here what my findings are, as I don’t know yet if I will finally participate in the competition - and maybe this information helps others.

This Wikipedia article says (in German) that for main railway lines, the maximum gradient without need of additional “helping” (security) measures is 1:40 (25 ‰) on main railway lines. A good example of a main-railway line with gradients completely below 20 ‰ is the Schwarzwaldbahn, but you see the interesting alignment in the middle of the railway line, which was necessary to prevent steep gradients. Something like that is necessary in any case between Las Vegas and Cuaquensenutes.

At least if it needs to be a main line with very “smooth” gradients and if you want to continue with approximately the same gradients that you used up to Las Vegas - about 22 ‰, if I calculate right. But this would mean that we need about 24 km of railroad length to manage the 600m of elevation distance between Las Vegas and Cuaquensenutes (5km as the crow flies between both places).

If we go up to 40 ‰, that is the maximum gradient for branch lines in Germany, only 15km is needed.

If we look further up to Cueganabaquias, it is at an elevation of 1450m, 360 more than Cuaquensenutes. The distance between both places is 11km. We would have a gradient of 32 ‰ if we follow the river bed (approximately, we should not have a curve radius smaller than 200m, I propose from what I read about railway lines). I think that for a railway line of the type that is needed here, 32 ‰ still is ok, but of course trying to stick to a maximum of 25 ‰ would make this challenge even more interesting, because we can’t directly follow the river bed then. :)

Two more interesting examples from main lines in Germany:

Zig-zag at Elm (Schlüchtern) and Schlüchterner Tunnel - From 1868 until 1914, trains had to change direction (zig-zag) at Elm station, before the “Schlüchterner Tunnel” opened in 1914. In 2011, a second tunnel was opened, as this is part of the Fulda - Frankfurt main line, one of the busiest north-south routes in Germany.

Spessart Ramp (see also at Wikipedia) - The new alignment including several tunnels only opened this year. It is part of a busy east-west connection between Nuremberg and Frankfurt. Before the opening of the new line this rear, bank engines were needed for longer freight trains to pass the ramp.

Ok, so far for now. I have I conflict here. I want to participate in this challenge, but I also have my own challenge to continue with my city of Siuvna and to have the general layout (with all landuses) ready before October… Like always: Too many interesting things to do in OGF. :)

Happy mapping!


Comment from Leowezy on 30 July 2017 at 19:54

I’ve made a quick draft and shared that with Luciano; I went with a max inclination of 50‰, and took 80‰ as my absolute maximum (mostly to account for mapping inaccuracies - xD ). I don’t know too much about railway operating, but I thought that for such a relatively short route, trains might as well run with bank engines all the way instead of just at a few difficult spots; especially considering that the amount of trains taking that route would most likely be very limited anyways, as the population centers it connects are pretty limited in size even today, and I guess even more so back then. I would argue that the resources needed for the operating the bank engines for such a limited number of (nevertheless strategically relevant) trains would by far be compensated by the much easier and shorter alignment.

I have to say that I went the exact opposite way than joschi proposed, and build dozens of zig zags. My arguing was that the main goal is, like Joschi stated, to get to the top of that mountain chain (I think we all agree that a base tunnel is out of question), and that very long detours would be needed for that. I tried to cut back on bridges and tunnels as much as possible, as I think building them in such isolated locations would have been a major obstacle back then.

Another major point I realised was the curve radii; the terrain seems to not only be rather steep, but also very “loopy”, which makes it very hard to align the tracks on the mountain side. I went with a radius of ~300m, a bit less in specific conditions, and still had to build quite a lot of bridges and tunnels. That was another reason I found it hard to get by without zig-zags.

I must say, Joschi’s points really unsettled me about my concept. But I guess that’s the point of the challenge after all, to see what different types of concept the community can come up with. And Luciano will be spoilt for choice in the end ;)

Comment from Luciano on 30 July 2017 at 20:30

Just a few short comments for the moment, from a moment of insomnia, middle-of-the-night Seoul Time.

@Leowezy - wow! Saw your work so far. Very well thought out. To be honest, I simply hadn’t even thought about doing switch-backs. It’s a pretty elegant solution, and your observation with respect to the Huancayo line (Peru) is exactly right - that’s the type of railway this is, I think, both in terms of timeline and style.

@Udi - actually, I hadn’t thought to set up a wiki collab page on it. If there is sufficient interest, I might, but in fact I’m not sure it’s entirely necessary - I feel like I’ll get more interesting results WITHOUT people seeing too much of other people’s work: Leowezy putting in switch-backs, and joschi81 rejecting them, is a good example of that.

@joschi81 - I appreciate all your numerical observations. They help me a great deal not just for this project but for thinking about other railroads, too. The version I had built (and deleted) was all 25 ‰ with zero switch-backs. It was quite horrible, because I had these implausible spirals, which abused the terrain (i.e., having access to the contour map made me behave very badly, as I rearranged canyons, etc.)

General thought: I’m not sure trying to tightly follow the Río Cuaresma gorge is the best approach - it’s too steep, as joschi81 points out. The general configuration of the land is a kind of “wall” and perhaps the best solution would be a kind of zig-zag, 5-10 km long, up the wall, and bridging the canyons? Then we get the 24km of distance needed from Las Vegas to Cuaquensenutes per joschi81’s analysis. The train doesn’t need to go right into Cuaquensenutes - just somewhere near - I could even imagine a kind of “old town down in the canyon, new town up near the railroad” built in response to whatever is the final result…

Comment from Luciano on 30 July 2017 at 20:44

Another small point, in case I have not been entirely clear about it - population centers are more sparse on the map than they “really are”, because the mapping hasn’t been advanced that far along. Thus, although Las Vegas is a small town, Cosmópolis (about 10km south of Las Vegas) is a good sized town (maybe 50000 hab.), Narao is similarly sized, and of course Comala (another 50 km southeast) is a major city (200000 hab in the metro area), so far completely unbuilt - but ¡check out the 1000 AD historical snapshot pre-colonial city of Quelepa, in development there!

This railroad would have been about connecting the city of Comala with its northwestern hinterland and getting a direct connection to the international frontier at Narao (and hence connections onward to whatever Dandria may some day become and beyond that, countries like Zylanda, etc.).

Comment from Ūdilugbulgidħū on 30 July 2017 at 22:47

I guess its a competition, so it means people will come up with their ideas and you’ll chose the best, or the best sections. I’m not going to commit myself to that amount of work, so I’ll just put some thoughts here in the hope they might be useful. I was wondering if any rail line at all up ‘the Cuaresma wall’ was actually realistic. Is there an alternative of a line going east from Las Vegas up the Arroyo S Miguel and looping around there? That could eventually top-out on the highlands a couple of km north of Vegas. There’s enough distance to get the line up to the right altitude there, its more a question of spacing for the loops at a couple of points in the line.

Comment from Luciano on 30 July 2017 at 23:34

@Udi - That is an excellent idea. In fact, before making this competition, I had been looking at the map and thinking something similar - i.e., just avoiding Barranca del Cuaresma altogether. If we select that option, the line through the actual town of Las Vegas might become a spur line, maybe originally to mine nearby or something.

Thilo had a proposal to go over the pass to the East, via Quebrada Mogonte / Arroyo de Bronce. Realistically, this might be an easier option, from an engineering standpoint, but I decided somewhat arbitrarily that Ciudad Sokolov had to be included in the line, for some obscure geopolitical reason or point-of-pride, or maybe there was a corrupt official from that town? Anyway, it’s just a constraint I decided I wanted to make.

Comment from wangi on 31 July 2017 at 01:36

Good to hear somebody else thought about Arroyo de Bronce :)

Enjoying the discussion.

Comment from histor on 31 July 2017 at 09:36

Nartao as railway-knotpoint has it importance for the railway-line to Zylanda. For my opinion the better way is along the river to Oscuridad and Flecha Caida, up the valley at 121,570° latitude (Arroyo de Bronce), a tunnel of ca. 4 km at the summit near Gugresejo and then south of the river at the “small rooster” and Rio Cuaresma.

Ciudad Sokolov can be junctioned with a line from the north.

What gradient is to use too is a question, what locomotives run over your rails. So in Würtemberg / Germany they made the trace more steeper, because they us stronger locomotives. And sure in the mountains you can have curves smaller than 700 m - but then you lost speed. There is always the question, what you want - but “you can’t always get what you want”. A radius of 200 m is too few.

Comment from histor on 31 July 2017 at 09:39

South of the summit-tunnel you can build a branch-line to Ciudad Esoterica (with a second tunnel).

Comment from Sarepava on 31 July 2017 at 17:11

What sort of budget was available to build this line? Most railways of this period were speculative enterprises, where a company would sell futures in order to raise the construction costs, and then make back what they’d borrowed when the line was in operation. Depending on the available funds, more or fewer tunnels and bridges can be built, more navvies hired to build it etc. This all affects the geography of the line.

Comment from Luciano on 1 August 2017 at 00:25

@Sarepava - Budget is a great issue to raise. This could lead to all kinds of constraints. It’s notable that one of the real-world routes mentioned above (Salta-Antofogasta) took something like 40 years to complete, in typical Latin American fashion. It’s possible, for example, that this line was originally proposed in the 1870s but wasn’t completed until 1912 due to budget constraints, with perhaps several “false starts” leading to a suboptimal route, too. The presidency of Yuna Namgung (1900-1910) introduced a lot of reforms, including battles against corruption and new public-works spending. This may have been what finally resulted in the line being completed.

@histor - A 4km tunnel sounds very modern, and unlikely in a middle-income country (at that time) for 100 year old project.

One thing to note. This is a HISTORICAL project and railway. There is no reason to suppose that perhaps a newer, better-engineered railway might be built on a better route, in more modern times - we’ll get to that eventually.

I know I almost never mention it, but Ardisphere will eventually have high-speed rail, with at least a few major lines (i.e. VC-CyC, VC-Jeongto), and perhaps international connections. But I don’t think it’s appropriate to build it or even plan it, yet. I remain committed to generally filling in the map historically, first. With many exceptions and anachronisms, mostly I focus on my “1930s” snapshot - what did the country look like at that point in time, before the push for modernity led to motorways and high speed rail?

Comment from Luciano on 1 August 2017 at 23:20

Here is an update which I am duplicating here from a private conversation with @Udi, just to keep the conversation all in one place for future reference.

@Udi sent me an OSM file with a partial route:

Spent a great few hours in the late 19th century imagining I was surveying the canyons and pueblos. Did you know there’s a rare species of ibex there as well? ;)

Happy mapping


I answered:


Ah, I love the amazing, implausible diversity of animals in the Ardisphere. I needed a new species, definitely. This shall be el íbice paraménico (Capra udis?).

I have taken a look at it this morning… main thought: it is excellently realistic. With that long detour eastward, however, I actually doubt the southern terminus would be at Las Vegas (except maybe as a kind of spur.

Much more rational would be to come up the Cuaresma directly from Sequiyá / Tierra Amarilla. It also occurs to me that another plausible route could be to exploit the already high elevation of the Comala-Yahuas line at the pass just north of Monte Mirón, and branch over from there. This feels very realistic and also allows Comala to be more of a rail hub rather than just an “end of the line” type place.

I drew a 3-minute sketch.

The more I think about it, the more I like the idea of dropping Las Vegas as a southern terminus, and branching over from Comala. It solves the elevation problem, because it allows a much longer, gradual climb “diagonal to the grain” rather than “against the grain” of the landscape.

Comment from Ūdilugbulgidħū on 2 August 2017 at 09:20

I actually had the thought that a line from Comala itself to Gugresejo might be a more cost-effective and easier solution - avoiding the Rio Cuaresma to Gugresejo section altogether, connecting Vegas/Cuaresma on a branch line from the south. And if that’s the case it makes the route beyond Cuaresma a bit obsolete (though I quite like my solution there!). That line would follow roughly the route you sketched to Comala (I thought a slightly lower route might get it out of the Rio de la Virgen valley more easily). Worth coming back to - but beyond the scope of your challenge!

Comment from Luciano on 6 August 2017 at 04:55

I have received a quite complete and detailed solution from joschi81. It is similar to the route that I had originally drawn, and includes a similar “spiral” approach at the southern end, along the Río Cuaresma.

Except for those southern spirals, I’d be content to use it “as is”. I would like to see what else people might do, however, so I will continue to leave this challenge completely open.

Login to leave a comment