When is a clean bridge not a clean bridge?

[Klei]

Quote: from mendel on 12:11 am on Nov. 7, 2001
My monster post deserved a monster reply. Argh. Getting out the scissors...
...
...
...
Here's a textbook example of what this rule forbids:
http://www.klei.dns2go.com/downloads/pontifex/cables/compcabl.pxb" target="_blank">http://pontifex.mendelsohn.de/forum/compcabl.gif" border="0">
...
...
...

Your bridge doesn't work (it twists and falls).
I've built one that works (click the image).

[Gray]

I'll try to indeed adopt your list to decide which trains are clean or not... anyway, my opinions:

You got two No. 3's there http://www.pontifex2.com/iB_html/non-cgi/emoticons/smile.gif" border="0" valign="absmiddle" alt=':)'>

I think that 6 is OK for a clean bridge, just think of something invisible holding up the anchors, it's a matter of presentation.

Don't need 12 either, why does Pontifex have Easy and Medium difficulties anyway? http://www.pontifex2.com/iB_html/non-cgi/emoticons/smile.gif" border="0" valign="absmiddle" alt=':)'> (I think 'lighter load' should only refer to difficulty settings, after all, the number of cars is set by the level author and changing it changes the way the level has to be solved)

13 is ... well ... it might also be a matter of presentation. For real bridges, not all cables carry max load all the time, most of the time the load is different. Remember, Pontifex exaggerates bridge elasticity to better show the effect of the load that the train places on the structure. So compressed, 'buckled' cables only result from the exaggerated visuals of the game, those cables are not compressed IMO.

Regarding 14+15: Define 'steep'.... everything but a horizontal deck? These two shouldn't be on the list, after all, why is the player allowed to build decks at any angle in complex mode? It could stay restricted to one horizontal plane... Only for the dirty bridges? Very few 'clean' (loosely defined at the moment...) record briges have a level deck now. Such a restriction would severely limit the scope of possible designs the player has at his proposal for every given level. We should at least allow 'smooth' (1 vertical HD unit 'angle difference') transitions between two adjacent deck pieces. But my last word is 'allow any incline, as long as the other rules are not violated'.

IMO, No. 2, 4, (and 14+15) should be allowed for certain levels where violations can't be avoided (in order to still draw a line between 'clean' and 'dirty').

My list would be:
Clean:
Never allowed: 1,3,3,5,7,8,9,10,11
Sometimes allowed: 4, 2(jumping)
Dirty: everything is allowed, though I don't like allowing 5...

[mendel]

I think so far Gray and I are in agreement now (finally!), except for one thing:

Gray thinks falling bridges are further away from reality than inclined deck ones.

I say, they're just as far away as anchors in the sky. Who said it was just a matter of presentation to imagine something that holds them up? I contend there's an invisible crane holding that up, and a clumsy crane operator drops it into place too forcefully....
Especially steel truss bridges are often pre-constructed, I think - I've seen this on a railway bridge in my home town. Ok , that was shifted sideways onto the final site, not dropped from above, but still...

Klei, thanks for showing that compressed cable can be used for "clean" constructions (except that it's touching the ground in both cases); I knew my example soon twists and falls, but it showed the compressive stress very well. On your go, the stress is more distributed, and the cables don't get as red(dish) as they do on mine.
Yours cushions the fall even better than mine does; if one could just come up with a version that doesn't need the ground anchors, that would make dropping things much easier...

[VRBones]

Just to keep the thread going :

In my mind, slanted deck is a bug. Not because it is unrealistic, but it appears incorrectly in the designer section, therefore unintended by CL.

A brief note on point 2. Unless you build perfectly flat bridge deck, your train will ALWAYS leave the track at some point. You just knocked out 90% of the record bridges and the main advantage of complex mode.

Personally all these rules just make me want to build dirty bridges. Anything goes apart from obvious bugs.

 

[orealis]

with an arching deck.. 40 billion miles of cable (feels like it) i made a "rubberbridge" when the train passes it it compresses it so its almost flat and when the last wagon passes it bonces back and throws the train off the bridge and onto solid ground.. i have been trying to make a more bouncybridge that will set word records in traintossing.. but IF U REMOVE SLANTED TRACKS I CANNOT DO THIS ANYMORE.. SO.. NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
OOOOOOOOOO..dont remove!!

[Gray]

Would you mind sending me your creation so I can upload it to my page for everyone to enjoy?

Simple answer - when it breaks links.

However, timing seems to play a major part in whether some bridges break or not. I've recently had to withdraw a couple of my records as I stumbled across certain timings where they popped links.

It's obviously impossible to test any bridge on every single timing. It's even impossible to test for the same timings as these vary according to what system the testing is done on. Gray and I agree that the only "standard" timings we have are zero delay and after settling - however, I don't really think it's fair to evaluate a bridge on just those two.

We could adopt "contest rules" and just test bridges after settling - this would certainly simplify matters, but the purist in me isn't too keen on this idea.

A second approach would be that bridges had to work with any delay, and were "innocent until proven guilty" - so if anyone ever managed to get one to pop links, it'd be disqualified.

Another possibility would be to include a "set delay" feature in the game, and have a standardised set of delays to test each bridge.

Does anyone else have ideas or opinions on this? ChronicLogic, I know you're out there lurking (lurk lurk!), have you guys decided on a policy yet?

[Andy24]

It would be cool if there was an automatic delay. mabey 90 frames or something.

that is about three seconds

(Edited by Andy24 at 10:52 am on Nov. 3, 2001)

[mendel]

500 frames = 30 seconds at normal on a very fast system

"Clean" means for me, show the bridge to an architect, say "build me this" and not get laughed at - I think that was the rationale behind BB clean, as well.

This means the train has to stay on the tracks at all times,  must not get "wet", and of course the bridge must not change in any way through use, sustaining a potentially infinite number of runs. All this on "hard".

For those of us who find that "reality" is no fun, that's what dirty records are for... anything goes as long as the level completes at least sometimes :-)

[Gray]

I agree with mendel, I'd use my BB cleanness policy pretty much completely for Pontifex - i.e. as realistic as possible is the most important aspect.

The 'innocent until proven guilty' concept is basically what I use for BB records. If Pontifex gets a 'measure delay/set delay' function in a future version (start train manually, then display delay in frames/enter specific delay in frames), people will even be able to post the exact critical delay times when they find them.

[Andy24]

my definition of clean is anything that doesn't break links on hard mode. if it hasn't broken links yet it is clean.

(Edited by Andy24 at 10:51 am on Nov. 3, 2001)

[falkon2]

Great! I can get all Ũ records that way! Nothing to break =)

[Gray]

Andy24, that was indeed an insightful reply - I think we all agree that difficulty has to be set to hard, and that broken links are not allowed. The question was, how to tell exactly whether a bridge gets broken links or not? Did you read Chillum's post thoroughly? The first line contains your reply already.

[Andy24]

I think that bridges in which the train jumps or gets wet are ok

but I agree with the infinet run thing though.

[mendel]

"innocent until proven guilty" works only if the bridges are published; this means it's not a good idea for a contest (unless you want to demand your awards back - and remember, this could be the result of a simple oversight), but could work well for a records table.
Obviously dirty bridges are much easier to handle :-)

Placing a horizontal bar at the top end of the grid on a default level takes ten seconds to fall on my machine. Call this machine- and bridge-specific time 1 (one) BF. So one could announce that the bridge would be tested at 0 BF (=TT), 1 BF, 2 BF etc.
The standard bar to use is a 10m horizontal light steel bar, and the cost of 趚 is subtracted from the budget.