I'm going to reply to all the responses in one post. John Cater wrote: > > Hear hear! > > I'd agree with you on every point, and I hereby offer myself to > contribute to this effort - though it would be nice if we could use the > GTK editor... :-) Great. This is certainly a task that can easily be split up among several people - different people can take different areas of the world and fix it up. I'll take care of expanding and the initial tiling (simply because that will be done by scripts). Of course, people should communicate what map areas they will be working with. > > At the same time I'd like to shift a few of the existing maps around. > > e.g. Fold Port Joseph into Scorn port area, and move beginners 2 into > Scorn central. Certainly some shuffling can be done. Certain areas are not very large. I've always disliked the number of 'small' places that you get to by boat - they feel disconnected from the rest of the world. > > I suggest a new CVS directory, where the new maps go as we create them. > Maybe start-off with a blank canvas for the world, and go from there. > Personally, 32 x 32 appeals to me as a map size or maybe 24x24? Yes - a new cvs directory is probably appropriate. Something like 'maps-new' or the like - dunno - I would like to keep maps in the name, but generally dislike the idea of using relative terms like 'new'. As then a couple years from now, someone will say 'what is new about these'. Maybe maps-bigworld or the like? Each tile size is a matter of debate. There is no reason the tile size has to be a power of two (or nicely factored). Making the size 25 then means that the math is easy if you want to figure out how many spaces away something on another map is. I would prefer to keep the maps square also for simplicity. The main considerations for a map size are: 1) Number of objects. Making each map 100x100 would mean a lot of objects - thus longer load/save times, and more memory consumption. Given current hardware, both of these may not be that big a deal. 2) Ability to edit and have the entire map on the screen. The editors do have scrollbars - perhaps the point here is that it would be annoying to size something so that it is just a few spaces larger than will appear on your screen. I have no idea the typical resoluation people run out, but if you presume 1280x1024, that means that vertically, you are talking 32 spaces presuming no adornment for the editor at the top or bottom. 3) If some things like random encounters/objects showing up based on overall map attributes (and perhaps other random things based on total map attributes), you want the map such a size that the attributes describe an area of reasonable size. Now one thing I just thought of - if the world is expanded by 30, and the size of each map is also 30, that basically means each space on the current map would become an entirely new map. I'm not sure how useful the auto expansion would be in that case (other than to use as landmarks). > > At the same time, I'd like to see the directory structure for the maps > rationalized, so that maps in similar places were grouped together. Agree. I would think that maps should be sorted based on where they are located, not on author. > I've got a funky idea for the maps. What if monsters could apply exits? > > I can imagine some fun, when the monster you have been teasing follows > you outside - imagine being chased by a Balrog around the city! Note that monsters will follow players across maps that use map tiling. So for example, if you met a giant in the hills and decided to run back to town, it would follow you (presuming the speeds are roughly the same) > At least monsters should follow you up stairs etc. I've noticed tons of > people dropping down stairs, letting off all their mana and then > applying the stairs again to regenerate. Maybe only single-square > monsters that apply things (like vampires) could apply the stairs too > and follow you. I think all this will do is change the strategy players use. The other problem is that the logic here can be difficult. If the vampire is no longer on the same map as the player, it is difficult to know how the vampire should get to the player. If its standing on an exit, then yes, it could see if it leads to that same map. But if it is 10 spaces away, you would need to add code to have it search for exits. And different people may have different opinions on what monsters would really know the layout of the dungeon (eg, something like a black pudding probably would not, but a vampire should). But if the player now puts an extra map between him (eg, goes up two levels), this of course gets very difficult. The other potential problem is that if monsters follow players too far, they may end up where other players do no expect them. Think for example of monsters following some players out of houses in some of the cities. New players showing up may not be very appreciative of there being some nasty things about (and of course the player who lured them there, knowing he is out classed, may have decided to save hoping for a map reset or someone else to take care of them) Tim Rightnour wrote: > So.. I'd be more inclined to do it differently than you suggested. Perhaps a > skill such as mining, or herbalist. A herbalist could go out into the forest, > use his skill, and poof plants appear. Perhaps this could just be another > effect of wood-lore. Miners could go out (a pick axe or shovel might grant the > skill) and mine rock, and find chunks of gold or gems. That way the items > aren't just lying out there for all to find.. you have to go into the forest > and look for them. That could work - using the skills lets you find them if they are there. And of course you would get some small amount of exp for doing this. Of course, if nothing has been generated in that area (or someone else has come through and grabbed what stuff was there), your out of luck. > > > 4) Maybe add some partial line of sight blockage, eg, maybe a total of 4 is > > Neat. Perhaps it could work like light dark. As you look into the trees, it > gets hashed until it's black. Yep - that could get done. Scott MacFiggen wrote: > (my comments about making the game less blocky ommitted) > > I was actually considering taking this problem on for my next project. There > are some good articles on gamedev.net that go over ways to implement this. > The one I was going to try was adding a fringe layer to the maps. > Although the biggest problem with this is it breaks the new map command since > there are not enough bits left to add another map layer. Anyway the basic > idea is you have terrain precedences and a fringe bitmap for each > terrain border. So to take an example from the article, if you had > a forest and a grassland border, the forest would have precedence > over the grassland and stick out a bit into the grassland tile. > > It is a short article, URL is > http://www.gamedev.net/reference/articles/article943.asp That article does not seem to describe anything about terrain, but more overall game design. Note that the protocol/map currently supports three layers. So if you currently have something like FP spaces (F forest P plain), the layers as sent to the client for each would be something like forest, blank, blank then plain, blank, blank. You could put in a 'fringe', and then the stacking would be forest, plain fringe, blank, etc. There are already some maps that use something like this - rounded water spaces - look at wolfsburge/piratetown. The water of course is pretty easy - you just chop the image where needed. Doing forest would be trickier, as you actually need it to tile correctly with the piece adjoining it. But I would think that the same type of thing could be used without much trouble.