Hi James, Ves and others.
recently I was working on signals for pak128.CS (standard and experimental as well), and I have searched and read a lot about the history of signalling in Czechoslovakia and Austria-Hungary. There's a lot of differences to British practices. I can't speak for other continental countries though, but you may compare for yourself.
The main difference was perhaps that all railways were first built as single track. Second track was laid only when the traffic was high. Even the first steam railway in Austria-Hungary - KFNB (which was quite important) was double tracked some 20-30 years after being built. Though on my search on early signalling I never found info about using tokens or similar things. Perhaps only some small private branch railways with low traffic did use that. What I read was that the early railways used time-interval signalling. Probably in 1870's-1880's absolute block signalling was introduced. (Sources are not really clear, sometimes contradicting each other). However, that means that both time-interval and absolute-block were widely used on single tracked railways. But IIRC, simutrans-ex implementation expects these two methods to be used on double track lines. The signalmen at stations, where trains could pass and overtake, had to agree on the direction of travel. (Which is sort of equivalent with the blue directional reservation in circuit block). They could use telegraph, and later telephone or special electromechanic devices for that. To notify the signalmen (watchmen) on the track a special optical telegraph was used - see http://www.laenderbahn-forum.de/journal/heldenzuege_1859/heldenzuege_1859.html
If you scroll down you can see on the picture something that resembles T-semaphore, but I think these were used to pass info from one signalbox to another, not to the engine driver. This was later replaced with bell signals - series of rings informed which way the next train is going, or if any exceptional case happened (like runaway wagons, help engine needed, ...) From all what I collected, there is no big difference between absolute and circuit block here - at least from the engine driver point of view. The difference is in the (speed of) communication between signalboxes, means of detecting that the track is clear, and human factor involved. You can see that even now there are lots of single track lines on this map: http://provoz.szdc.cz/portal/Show.aspx?path=/Data/Mapy/TZZ.pdf
(red lines - cab signalling, blue - circuit block, others - absolute block, thin black - D3 direct control via radio (verbal token block), dashed black - D4 something like RETB). The simplified D3 system is in use since 1960's, and D4 is brand new. So, not much token block in history anyway.
Time-interval is yet something different. Sources are not very specific, but flags and lanterns were used for that. At some time, the color was not important, but the movement was used to distinguish different signals. Some sources state that the earliest color flags were black=danger, red=caution, but no idea what was used for clear. (And in night
). Yellow/blue/white colors were used for similar purposes as optical telegraph - send help engine up or down the track.
There are contradicting sources on early signals (for time-interval), and whether classic (upper quadrant) semaphores were used for time-interval or not. It seems that lower quadrant was not used here at all. However I found a lower quadrant signal in rule-book from 1914, but it was a fixed signal meaning "proceed slowly" - i.e. speed limit sign.
I found that a T signals were used in czechoslovakia at the beginning od 20th century: http://www.k-report.net/ukazobrazek.php?soubor=805778.jpg&httpref=/
However the text says that it is just an alternative to two normal signals. The signalbox you see is on single track somewhere in the middle between two stations. So instead of having one signal on each side of track, they built one T-signal - the train had to stop in front of that signal.
At that page, there is another interesting rule described - if the signalman does not get confirmation that the last train has arrived to the next station, and upon query does not get information that the train is still on its way, but according to schedule the train should have arrived few minutes ago, then next train can be sent on the track (with command to drive slowly). Perhaps some sort of fallback in case the telephone/telegraph is broken.
Another interesting thing is that Austrian rulebooks have "distanzsignal" and "vorsignal". Vorsignal means presignal, and shows caution/clear, just like UK distant signal. Distanzsignal means literally distant signal, but shows danger/clear. Its name is from the fact that it was built at long distance from station (500 m from outermost points), protecting the entry. It had no presignal on its own. It was supposed to be built on such place that it could be seen from big distance, so the driver could stop the train. Later when entry signals were moved closer to the station (50-100 m), these were probable repainted to vorsignals. You can see them here: http://www.modellode.cz/Doplnky-zeleznice.html
Picture 1,2 - Distanzsignal, 3 - normal and T-signal, 4,10 - optical telegraph (balloon signal), 5 - flag signal, 9 - vorsignal, 11 - shunt signals
I think the distanzsignal is technically similar to the swedish "plate signal". And it seems that it predates the use of absolute block, but the sources are very vague about that.
So to sum it up. For Czech signalling, the implementation of absolute block is not very useful. I'll use circuit block with normal_danger=1.
It would be nice if time-interval signalling is usable with single track - i.e. having the directional reservation as circuit block has. Otherwise I'll just fall back to 3-state or permissive circuit block.
For token block - to avoid most of the token block usage in simutrans, it would be sufficient if train does not revert to drive by sight when departing from station (unless it turns back).