Simutrans for Android: is it possible?

[Tontarelli]

Well, I am a fan of simutrans, just love it. So I have device that runs on android, I'd love to play it there. Is it possible? I've seen Czech android discussion but I don't understand a thing.

[prissi]

Since android uses SDL, download sourcecode and compile should work straight forward.

[ӔO]

I have attempted to compile an app for android with the drunken 10 step guide, but I'm totally stuck at step 4, trying to compile SDL 2.0. 

I will say, that this is not for the faint of heart and the guides for compiling SDL for android are outdated by quite a bit.

[Vladki]

I have tried to play openTTD on android tablet, but it's almost unplayable without mouse. So it's not only about compiling, but also about some adjustments of user interface for touchscreen.

[prissi]

OpenTTD has double size interfaces. The UI of standard also moves (slowly) to a scalable interface. That would also allow for larger buttons. Apart from dragging (which is not really needed), I am not sure where mouse sopport is indepensable in simutrans.

[Fabio]

Touch devices can use long press.
So, press can be left click,
long press ctrl + click,
normal drag move the map or the dialog,
long press + drag could be drag tool...

[sdog]

building ways is almost impossible without draging ways. The method clicking at start and end tiles produces not very predictable results (rail tracks with sudden 90 degree curves). Building double track that way is not possible. Ctrl-click is also most of the time necessary to build double track, especially at diagonal track sections.

The interesting thing is, often adjustments to UI for tablet use also improve desktop use (recent example is gimp, but also gnome shell or ubuntu unity)

A way to solve the building and get rid of the ctrl click would be: Painting the way.
- The player paints a curve on the screen, where she wants the way to be built.
- A second click or tap on the line causes the way to be built.
- The chosen way is marked by a blue overlay (similar to the overlay indicating routes)
- way builder builds way connected between adjoining path tiles


Exceptions:
- diagonal too close to an existing diagonal same way type.
- ambiguous sections (eg way more than a tile wide)
- illegal way building

In those cases the problematic sections change are highlighted in red or yellow.

Yellow requires the player to touch/click again to build (that is joining parallel tracks for example) Red requires clarification by re-painting, or a click on a blue, unproblematic section that will in turn be built. If the red, illegal path, is not adjoining a blue proposed way it will just dissapear. (E. g. pull line crossing a river, with a way without fords. on either side of the river the path is highlighted in blue, the river tile red. touching either blue branch causes it to be built. when both sides are built, there are tracks up to the river the red path dissapears. [An alternative, make rivers yellow, clicking would build the default bridge for the waytype/speed] Same example, instead of touching blue and building, touching red would remove the whole path or cause any other resolution to be found)

[falconne]

A way to solve the building and get rid of the ctrl click would be: Painting the way.

I tried creating a design mode a while back. I ended up deciding it causes too much complication, and it would be better to have the bulldozer produce a full refund for a short period after building.


I think there's more design changes needed before the current UI can be effective on a touchscreen device, especially phones. The UI elements would need to be bigger, which means the dialogs would have to be less busy.

[Ters]

I can't see how Simutrans can be played in full on a handheld device. It would have to be some kind of Simutrans Light, more suited for simpler gameplay and smaller maps than requested by players on computers with larger displays. Simutrans Light could share the underlying code, but would have a different set of paks. Except that one might use Simutrans on a handheld device to check in on a running networked game, without intention of doing any serious building or administration.

[transporter]

I think Simutrans is only possible with something larger than the Galaxy Note. Since it has a stylus it would be more accurate. But using your fingers on a screen that small isn't practical. I downloaded TTD and it was really awkward to use.

[ӔO]

I was trying to compile it for my asus transformer w/ dock. One neat thing for this model is that you can use keyboard and mouse if you have the dock.

[prissi]

Just a remark: Until a year or so, way building with just clicking was the only way to do ...

[Vladki]

Just a remark: Until a year or so, way building with just clicking was the only way to do ...

Yep, and in those times, the only way to build double diagonal track was (for me) to build it tile by tile.
And that is maybe one of the reasons why I was not playing simutrans for some time in favor of openttd.
But now I prefer simutrans.

[jamespetts]

As Ters says, the full game will probably be a bit much for Android devices. Experimental has in some areas (particularly private car route finding, although that code is probably inefficient) been hitting the limits of what desktop computers can manage without unacceptable performance. Standard is lighter than Experimental, but one would still need to be restricted to a fairly small game (in modern Simutrans terms, at least) for it to be workable, and with a small size pakset (pak128, for example, would probably be too large).

I should note, however, that Standard (but not Experimental) has in the past sucessfully been ported to iOS.

[Ashley]

"ported" to iOS - kind of. The existing one lacks any kind of usable UI.

The main issues are adding hooks to many sections of the game to allow it to interface with the native UI elements provided by iOS or other mobile platforms. E.g. when you tap on a text entry field to begin editing it there needs to be a hook to allow iOS/Android to show the software keyboard and focus the entry box, and then to hide it again when done.

That's just one example though. The first step is to extend the current events system to add touch events, then provide platform-dependant hook methods for native UI functionality (which could benefit non-touch platforms too). It's a lot of work which I had planned to undertake once I got the OSX port working nicely and added to the app store (this would give a full working implementation of the game within a Cocoa shell along with package management for in-game download of paksets which is another needed feature on mobile platforms).

Anyway, as with all things I can't work on any of this until after my exams are finished. So if someone else wants to make progress in the meantime be my guest. Just don't under-estimate the work - it isn't just a case of compiling the source...

(In terms of performance, if you stick to smaller maps then it should run just fine. Recent phones are very capable computing devices)

[Ters]

[...]and with a small size pakset (pak128, for example, would probably be too large).

I actually thought large pak-sets, but small maps would be the thing for handheld devices. The tiles must be roughly fingre sized to so that players can reliably hit just one tile. As these small devices are getting quite high dpi, a tile size of 64 might be too small. The drawback of making it easy to click is that you can only see an area of a few tiles². To get an overview, one would have to zoom out. One could of course also zoom in to edit, but one would then have uglier graphics when editing. Or the paks could contain graphics for two zoom levels, so that both look as good as it gets.

What really makes the light version lighter is a reduction in tools and gameplay complexity. Otherwise the tool bar would fill almost the entire screen, or would have to be remade into some system of submenus which would make finding the tool more difficult.

I imagine the most complex dialogs will fill most of these small screens, especially since the size of clickable things would most likely have to be increased. They must either be simplified or removed, losing some functionality, or made into full screen dialogs. When entering a full screen dialog, the game should perhaps pause automatically, except the minimap. Dialogs that allow you to click on something to locate it in the map would need to put themselves into the background, from where they can quicly be restored as they were.

[ӔO]

I made a screen shot to see what sort of UI problems you would have on android tablets. Most common size is 1280x800, but the bottom bar eats up around 60 to 80 pixels, so you might be looking at closer to 1280x720.

The current icon size, I think it's 48 pixels, is not large enough. I think around 2x the size, 96 pixels, would be finger friendly. The android buttons are roughly 96pixels wide, but, clearly, they are more rectangular compared to the simutrans buttons.

Since I know space will be at a premium, 80 pixel square would be the most minimum for most fingers.

[prissi]

There are scrollable menu bars; z9999 did them a while ago. Those are shown, when a maximum menubar width and height are exceeded (specified in simuconf.tab, toolbar_max_width and toolbar_max_height).

[Bricktop]

Sorry for necro bumping, but I would like to take a look at the port that ӔO made.

[ӔO]

I never made the port, or at least I never got anything useful out of it.
The only thing I made was a 1280x800 screen shot that I used to see how it would look on an android device.

Either way, I've abandoned it, because now you can buy x86 tablets in a reasonable size.

I've tested simutrans out on an Intel atom Z2760 and i5-3317U and from what I can tell, a small game of simutrans standard will run on the atom, while the i5 will happily run simutrans experimental with 13000 convoys.

Seeing as how current ARM SoC vastly underperforms a Z2760 in computation, I don't see anything less than a Tegra4 or qualcomm S4 1.5Ghz will be able to run simutrans in any decent manner.

[Yona-TYT]

Expand to mobile platforms can give much fame to Simutrans, in my humble opinion






News of Simutrans on facebook

[jamespetts]

Expand to mobile platforms can give much fame to Simutrans, in my humble opinion

The trouble is that no mobile device has a processor capable of running all but very small games in Standard, and one that did would use all its battery in a few minutes...

[Fabio]

Honestly, don't mobile devices have at least the power of older desktops (e.g. Pentium 1 and 2) for which we still keep compatibility?
Sure it would be advisable a dedicated pakset, like 32 comic.
It might be a showcase project in google play (formerly Android market) which could introduce many new players to Simutrans and lure them into downloading the desktop version.

[Ters]

Maybe in some areas, maybe not in others. Different arcitectures might be hard to compare. Even different families of PC (x86-based) CPUs are hard to compare. The days of just looking at the clock frequency are long gone.

[prissi]

A Pentium 1 was with <120 MHz; those have indeed be surpassed by current smartphone. But thos had trouble to run a 128x128 map with 5 fps.

A normal smartphone is probably just enough to run a 256x256 map with 15fps. (Depends also much on the screensize.) But the interface is not up to touchscreens, and the battery will be empty in like in the batterery drain benchmarks (few hours at most).

[jamespetts]

I do have a pipe-dream about a "Simutrans Investor" app.: the main game (I was imagining Experimental, but in principle it could work with Standard) could be modified so that players have no "free" money with which to start the game, but must rather raise money, and then continue to raise money thereafter if they wish to expand beyond their current bank accounts. Players of "Simutrans Investor" (either on desktop or smartphone) would start with a pot of money and be able to invest it in different players' companies in the game (by way of bonds, for example), and their goal would be to make successful investments (monies would be lost if the players went bankrupt).

Alternatively, one might have a monitoring app for online games which will give players finance and line statistics, and the latest news/chat from the ticker.

[kierongreen]

The problem is current smartphones, while they might have a raw processing power similar to that of an early Pentium III are hugely different in architecture - they are likely to have 2 or more cores, which simutrans makes relatively little use of. Nevertheless, a reasonable sized map (up to 512x512 I would have said) should be playable on the latest iPhone, iPad, and top end Android devices. Tablets would be easier to use with the screensize, but touch interface is still a problem.

As for a monitoring app - the best way for this would be for the server to output statistics to an XML file (or similar) that was then read by the app. Transferring the entire saved game over a 3g connection (or even wireless) would not be fun, especially for other people facing a pause while this went on, and the phone would likely struggle to run the game even without most of the graphics.

I'm not sure how you would prevent abuse with a "Simutrans Investor" app - what is to stop people logging in on their phones to invest in the same company they are running on the computer, potentially many times?

[jamespetts]

As for a monitoring app - the best way for this would be for the server to output statistics to an XML file (or similar) that was then read by the app. Transferring the entire saved game over a 3g connection (or even wireless) would not be fun, especially for other people facing a pause while this went on, and the phone would likely struggle to run the game even without most of the graphics.

Yes, it'd definitely have to be done this way.

I'm not sure how you would prevent abuse with a "Simutrans Investor" app - what is to stop people logging in on their phones to invest in the same company they are running on the computer, potentially many times?

The only way to prevent this sort of abuse, I think, would be (1) to limit the number of investors per game; and (2) require some sort of identity verification to ensure that no investor was a player, or vice versa. The latter might be more trouble than it is worth. As I wrote - a pipe dream...

[Ashley]

If I were going to implement this I'd design a REST interface and build it into the game server, such that data could be queried by the app (or anything else which wanted to get at it).

As for mobile platforms, IMO (as I've said before) the biggest problems are the UI (interfacing with "soft" keyboards and touch events) and pausing/resuming the game when you open/quit the app. Processing power is, IMO, really not any more of an issue than on the desktop version - battery life is much more of a concern!

[hreintke]

LS,

More or less off topic but I agree with Timothy on the REST interface. Would open a lot of opportunities.

Herman

[Bricktop]

I actually think You're all underestimating the power of androids , with the flux of new cheap 7 inch nexus devices, and cheap 5,5 inch "phablet"s its only a matter of time people will be playing more "serious" games on their mobiles. Only thing we need now is a android sdl port or an actual tutorial . To all who still doubt how simutrans would be played on mobiles, please check the android port of Theme Hospital. Its called CorsixTh and it performs pretty decent on my tegra 3 device.

[AP]

Alternatively, one might have a monitoring app for online games which will give players finance and line statistics, and the latest news/chat from the ticker.

This would be VERY useful. Especially if it could provide alerts to key items, e.g. player-vehicles which had just become "stuck".

[Fabio]

I would think of a program & app to interface with a net game without downloading the save.
It could query the server and get
1) the mini map updates every minute or so without vehicles
2) a minimap vehicles overlay updated every second
3) some financial data
It could send back to the server some basic commands (e.g. Add a vehicle to a line cloning existing ones) which would executed if possible or ignored.

[jamespetts]

I think that the minimap thing would not be easy, as most servers are compiled without graphics, and the minimap requires the graphics modules that are omitted in the server builds.

[prissi]

The minimap is compiled without any graphic routines. However it costs very much time to compute it. Thus only the visible region is show, and it is only updated every month. A server continiously updating the minimap may be too much. Furthermiore I think a minimap without options is not very useful, some people want to see ownerships, vehicles or other stuff. The only way to compute it, is to actually work with the map data, i.e. join a game.

[AP]

It could send back to the server some basic commands (e.g. Add a vehicle to a line cloning existing ones) which would executed if possible or ignored.

I suppose there are two sides to an app for server games - one as a remote viewer (which I think is a valid proposition by itself, even without a minimap) and the other allowing remote interaction with the server game (possibly being a more complex undertaking?).

[jamespetts]

I suppose there are two sides to an app for server games - one as a remote viewer (which I think is a valid proposition by itself, even without a minimap) and the other allowing remote interaction with the server game (possibly being a more complex undertaking?).

Yes, allowing remote interaction would be a very useful thing, in fact, for players to undertake management tasks without having to be beside a full-blown desktop client. This might well be a fairly complicated undertaking, but it ought not be beyond the pale in principle. The issue is whether there are enough skilled Simutrans programmers ready to implement this feature, I think, rather than whether it is feasible in principle, and that will depend in part, but not entirely, on whether enough people think that it is worthwhile.

As for the map, I wonder whether it would be possible for the server to generate some sort of simplified version of the game map, showing just land and water areas, town hall locations and town names, and for the statistics to be able to identify stations by co-ordinate such that they might then be shown on a simplified version of the map based on that just described to give people something of a graphical representation of the game world without all the computationally intensive detail required for the full game. It would be useful to get some idea, for example, of where a rival has built, or where graphically on the map that problems on one's own network have occurred.

[prissi]

The computer intensive operation is not the display of the map. Most expensive are the connections and moving stuff around. Ironically, also the minimap is the most compute intense display, as it shows most of the map

[Fabio]

But calculating a mini map, maybe even with different displays, every month and caching it to be retrieved by query and not calculated on demand would impose a small overhead to the server processing.

[sdog]

There's a new version of an android up in the google play store. Developer is TMN Trend Media Network.

Edit:
just noticed this is the wrong android thread.

[Bricktop]

 did You make it?
it looks kinda suspicious.

[IgorEliezer]

it looks kinda suspicious.

And why does it look suspicious?

EDIT:

Topic renamed from "Android" to "Simutrans for Android: is it possible?".

[Bricktop]

I tried that tmn group version and:
Its kinda unplayable, looks ugly, and actually doesnt fit the tablet screen. (even with the tablet option selected), the on screen controls doesnt do nothing other than obscuring the view. Also there is a giant ad bar.

Only positive at the moment : the porting is possible.

[sdog]

i tested it too today, the only thing i was able to click were the ad banners. Nothing else worked.

[ӔO]

There was something odd about the cursor accuracy, but once it loaded the starting map, I could easily tell my TF101 was not up to the task.

[Lmallet]

I managed to get past the language selection screen on my Galaxy S3, but the cursor control is so bad I quit without managing to start a game.

[Yona-TYT]

I managed to get past the language selection screen on my Galaxy S3, but the cursor control is so bad I quit without managing to start a game.

Woe that recognize that "simutrans" was originally designed to run on a desktop PC.
I am surprised that can open in the  Galaxy S3

[Lmallet]

Woe that recognize that "simutrans" was originally designed to run on a desktop PC.
I am surprised that can open in the  Galaxy S3

I recall Prissi saying a few times that Simutrans should run on either a 486 or an early Pentium (that was a while back).  With a dual-core 1.5GHz CPU and 1.5GB of RAM, I think the S3 is powerful enough.

The real problem is the screen size and the button layout.  On one hand the developer of the Android version seems to have used some sort of emulator/virtual machine to make it work, and the built-in controls are not the greatest.  On the other hand Simutrans has a lot of small buttons, which is ok on a PC but not on a smartphone.  I tried openTTD on my S3 as well and it is suprisingly quite playable, mostly due to a better implementation on the Android side, but also the button layout makes it easier to hit them on the small screen.

[prissi]

Simutrans with not too large maps should run on a 600 MHz arm qithout troubles.

Simutrans can easily scale up the buttons. The problems are the dialoge, which are not completely dynamical (only a few of them are). On the other hand openTTD had awful scrolling (on Android) and required a lot of modifier keys (while simutrans could be used with only left mouse button and some scrolling support).

[Alasdair Green]

Hi,

I've just got an Acer D270, and thinking of changing it to Android, would Simutrans run on this? It's a conventional netbook, so no touchscreen play issues.

Thanks,

[Isaac Eiland-Hall]

No current release on Android, I'm afraid.

[Yona-TYT]

I hope someday we can play from an Android device


It would be very nice ....... Regards

[Ters]

Some of you Android users just have to figure out how.

[prissi]

There is a simutrans (obsolete version) for Android, which is absolutely unsuable with a touchscreen. It may work with the mouse tough. Unfourtunetaly most ports did not distributed the source code and hence violated the Artistic Licence of Simutrans. Thus I have to note google whenever I find them.

[Yona-TYT]

seems those from Open-TTD you they won us in this





It has an Android version
Google Play https://play.google.com/store/apps/details?id=org.openttd.sdl&hl=es_419

[prissi]

There is also a simutrans version, with is as unplayable as the OpenTTD version.

[Yona-TYT]

GUI patch Theme can improve that, is not it?

[kierongreen]

Potentially it could. It would also need somone with a tablet with a bit of time to port Simutrans.

[Carl]

Topical: Chris Sawyer is creating a new tablet version of Transport Tycoon.
http://www.polygon.com/2013/7/15/4524242/transport-tycoon-sim-heads-to-mobile-this-year

[Ters]

Considering how imprecise touch devices are, I have a hard time imaging that this could really be playable unless the touch screen is 20 inches or more. It's hard to hit the right place even when using a touch pad, where my finger doesn't obscure where I'm pointing and the cursor has a more well-defined hot spot than my finger.

[kierongreen]

With a larger button theme and a reasonable size pakset I don't think placing the cursor would be a huge issue. Of course using pak32 is probably asking for trouble...

[Markohs]

The UI has to be changed a lot, it's a first step, but there's a need to change the tools mechanics, all the dialogs, maybe even using the android dialogs to replace some of the simutrans one. It's doable ofc, but needs a team of people expert and with experience in mobile app design. Also some game mechanics might need a change.

Whould be nice if we had the inner simutrans code decoupled and modularized even more, in a plugin fashion, so we could just use the minimum simutrans code and replace some parts with touch-designed components.

[Ters]

Bigger GUI means less room fot the world. Increasing the tile size makes it easier to hit a tile, but harder to see where that tile is in relation to everything else. One would have to zoom or pan a lot, and build in tiny steps. That would certainly ruin the appeal for me. But there are as many ways to enjoy the game as there as Simutrans players it seems.

[kierongreen]

Well all the gui stuff is in that subdirectory, and all the interaction is dealt with in simwerkz already isn't it?


Edit:
And yes, I prefer a smaller gui which is why I'm keen that any theme code allows this as well as a more touch friendly large button gui.

[Chukwudi]

This is old topic, the S6 is easy powerful enough. But few apps handle(truly show off)it's 3d capability (I've found one, Tapet, and that just scream.) It's octacore, and very fast..   and real time professional audio. 

[Ters]

This is old topic, the S6 is easy powerful enough. But few apps handle(truly show off)it's 3d capability (I've found one, Tapet, and that just scream.) It's octacore, and very fast..   and real time professional audio. 

Simutrans isn't 3D, so its 3D capability doesn't matter to us. When it comes to porting Simutrans to another platform, it requires a display with a medium resolution at medium DPI (exact values depend on pak set size and GUI theme), a CPU and system bus capable of updating that display at 12 FPS (GPU isn't used), a mouse with two at least one button, and a keyboard with Shift and Ctrl buttons. And for music, it needs MIDI support. Pretty much everything but PCs fail hard at meeting these requirements. In fact, even some PCs are getting troubles with the first two requirements.

[Vladki]

I think the touchscreen combined with high DPI is the showstopper. Did anybody try to play simutrans using touchscreen? (e.g. on windows 8/10)

[jamespetts]

I think that the (Standard) developers have been looking into this issue (with the touchscreen) for a while; I know that there have been some interface changes, and further changes to the graphics are planned to deal with high-dpi displays.

[Ters]

Only the GUI themes have materialized, so we are still just scratching the surface.

[prissi]

You can scale up all menus independently. The only thing not finished is the proper support of larger font sizes.

[c933103]

I found a simutrans port on android here: https://sourceforge.net/projects/libsdl-android/files/Simutrans/ , however the apk is 3-years old and seems to have problem in starting new game on my android despite it can load the start menu screen. I contacted the repo's owner and he said "Simutrans is not supported, I only support OpenTTD. Please ask on Simutrans forums, I think they also have newer version for Android available.". So is his comment about the situation correct? And is there such thing on this forum?

[Ters]

I am not aware that any of us is working on Android stuff. It seems to me to be pretty much just Windows desktop and regular Linux desktop, with some occasional MacOS developers dropping by.

[prissi]

I have looked once at the Andriod port, but the SDL then did not compile at all. In priciple it shoudl be quite easy for someone who had written at least one Andriod App to compile SImutrans. (It would still need an ARM and x86 version though). I am lacking the time to look into this.

[DrPepper]

There was an app. And it did work. It was years ago though. Unfortunately it finally became to buggy to use and I ended up getting rid of it. I would love to have it on Android again. No longer shows up in the app store. But I have found openTTD.
Not the same but it probably the closest thing you can get to simutrans.

[prissi]

In principle it is very simple to make SDL2 compatible android apps, but the build process is totally incompatible with the normal build system on any other platform. Also the storage operations needs some tweaking to use the right pathc. So one needs a simsys_s2_android.cc, and android.mk and someone with enough knowledge to make this work. I have never programmed for mobile, and the divers manifest files and independencies really dulls me. I played around for the last three hours and I got a crptic building error, which may pages say, "Yeah, this happens sometimes. Go install the Android studio on a different comuter usually solves this." Really?

(Ok, error message was: No toolchains found in the NDK toolchains folder for ABI with prefix: mips64el-linux-android) But I was no building that anyway.

[Vladki]

Some time ago I tried playing openttd on tablet, but it was very weird. The interface is now based on mouse and changing that to touchscreen is imho the most challenging part.

[Ters]

The most useful target for an Android version of Simutrans would probably be Chromebooks and things like that.

[prissi]

Simutrans can be played by touchscreen, the functions have been added (i.e. dragging map by left mouse, switch to IME for input). I was playable (the old version in the playstore) The modifier keys are not really needed, most thinks can be achieved without them.

[Ters]

That the game is technically playable doesn't automatically mean that it is fun to play. I could probably play Simutrans using only my nose to give input, but it would not be the same. I won't even play Simutrans on my laptop without an external mouse, and I seem more comfortable using the touchpad than most, if not all, people I know. A touchscreen might be better than a touchpad. I only have experience from my phone, and find that precision is a bit lacking there as well.