Since coming back, I've had a love-hate relationship with the new LANDMARKS system.
On the one hand, I think eliminating the tedium of walking to common destinations is great and I think it's also incredible
for new players.
On the other hand, it's very
odd that adventurers can suddenly walk through even the most obscure shortcuts. It's strange that the default method of getting between many places is to treat the rivers as highways. It's weird that using landmarks is actually better
than knowing the way and moving manually for most people in most situations.
I would like to propose a solution that is, I think, tractable - and then I'll talk about the benefits I see:
- Player does WALK TO TARGOSSAS
- Calculate path from current room to nearest major-road room (the nearest room that is part of the main loop).
- Calculate path from destination to nearest major-road room.
- Calculate path along road.
- Execute the autowalking as normal: current room to road, along road, off road to destination.
Now some crucial points and some benefits:
- These things can be precalculated. I don't know how the current system is implemented, but it has to deal with a combinatorial explosion of starting locations and destinations. Under this system, for every room/NPC location, the path to the major-road room can be precalculated and stored - it's just one path per room. Similarly for every destination. After that, calculating the shortest path along only the highway loop is completely trivial. Computationally, this is an easier problem (in that it can be turned into a simple lookup - the initial calculation is harder, but whatever) than the current pathing problem, so it can't be infeasible from a resources standpoint.
- This avoids problems like the issue with rivers interrupting autowalking - there aren't many rooms where newbies without waterwalking can still be swept away without any real recourse, but there are still a few. And things like river currents could be more liberal without having to worry about newbies getting stuck on them while autowalking.
- It eliminates the weird issue of people autowalking through areas they don't even know exist, which is weird for immersion. Similarly, not everyone in the game should be some incredible trailblazer that knows every shortcut in the world at 18.
- It makes learning your way around as a newbie easier. Instead of watching the map go crazy while they walk through a river, three forests, and a swamp to get between two places and then a desert, a lake, and a bog the next time, newbies see the same roads each time and can actually start to learn where things are. I think this is actually a more substantial problem than people realise - right now it's very easy for a newbie to get around for quite a while without actually developing any understanding of where things are in relation to one another. This hurts immersion, makes the world feel less cohesive than it is, and it also makes things more harrowing when, for whatever reason, landmarks aren't an option.
- It makes roads feel more like roads. Roads that actually function like roads.
- It makes LANDMARKS a convenience rather than an optimising strategy. These routes are slightly less efficient - they'd work fine if you're just traveling and don't want to bother typing in directions, but they retain an advantage to knowing your way around when you really need to get somewhere quickly.
However, even with your changes, it would make each move so confidently that (almost?) any human would quickly fall behind. It would have be wildly inefficient to be slower than I would be at just walking through, say, the highway between Ashtan and Delos (I know your example would have the walker use the the highway, but how many rooms it beats me by is how inefficient it can be for there to be no difference in arrival time). I would only beat it if I memorized the path and spammed the directions perfectly, and this isn't even taking into consideration the fact that server side queueing has no latency.
This is only really an argument to your last point. It would still be optimal unless they made it absolutely terrible.
As for the issue of knowing the relative placement of things, the maps should be placed more prominently on the site and updated.
For example, http://www.achaea.com/map/full-world-map-of-achaea <- still has Shallam
I didn't even know this existed until your thread: http://www.achaea.com/map/illustrated-map-of-achaea
That thing is pretty sweet.
Edit: the water walking thing is annoying though. I can see novices getting confused by that. I wasn't even sure how to avoid it because currents were new to me when I got back. I just raged really hard and started flying...
I imagine that would have a pretty huge impact on performance though, so your suggestion would be good as a sort of compromise. Although I'd add that, in cases where the destination is closer than the closest major road, it should just go straight to the destination.
Alternatively, have autowalking work through rooms you haven't explored, but move a lot more slowly through them, maybe 1 room per 2-4 seconds instead of 2 rooms per second. That would probably (if the speed difference is high enough) require the pathfinding algorithm to consider the cost per room since the shortest path could take a long time compared to the fastest path, so that might still be too big of a performance hit. Would also need to do something to prevent being swept away by rivers if it's slowed down to the point that it's a problem again.
The point is that there are a lot of places you need to walk to frequently and it's boring to have to input the same fifty movement commands all the time.
The other point is that asking newbies to memorise huge maps to get to basic locations has, historically, been a problem.
It isn't to be a faster way of getting places than walking manually. If it were, why even bother having it walk? Why not just have a message indicating you walked somewhere and just have you show up there with a short delay? Why bother calculating a shortest-distance path?
It's also not at all difficult to move as fast as the autowalker. With or without latency (server-side queueing doesn't do anything here - movement doesn't take a balance). And without really spamming (you get used to the speed at which you can walk pretty quickly since you do so much of it). I'm not entirely sure who you are and how well you know the room layout of the game, but most experienced players don't have trouble moving at the maximum possible speed - which is why things like the mhun bonus and the artefact that does the same thing exist (by your logic, they would have no utility).
And while Ashtan to Delos would be almost identical in terms of path and time, something like, say, Targossas to the Great Rock would be substantially slower going by the road.
@Sena my suspicion is that doing the pathfinding taking into account whether or not a room was explored by a given player is probably not computationally feasible. If not, that's a great solution too, but the reason I proposed what I did was largely because it actually reduces the computations necessary when a player starts walking. Well, that and my obvious obsession with roads.
Mostly, I just wish there were a reason to not use the autowalker sometimes. I wish that actually learning shortcuts (rather than simply having passed through the room one time at some point) were as useful as it used to be.
But it sounds like we just have opposing positions, so good luck. Not really sure what you plan to do about the Mudlet automapper though.
This might also go along with the highway 'extra move' mentioned in another thread, where using roads allows faster movement speed. (maybe with an 'out-of-combat' stipulation)