From: Erik V. <eri...@hc...> - 2005-12-26 12:17:24
|
I have started checking the tile orientations in Tile Designer vs. the files as provided with the games (1830 only so far). (the standard is that the position of the tile number defines either the S or the SW edge of the unrotated tile). It turns out that almost all 1830 tiles have a different orientation in TileDesigner! However statistically unlikely it may be, the only correct one (except the obvious tile #63) was tile #67! I have made a new version of the TileDesigner database, which I will send to Brett for conversion to .gif files. I have also doubled the sizes of the tile numbers and revenue indicators, hoping that these would get a bit closer to being readable. I will change the XML accordingly. As said, only 1830 has been done so far. Tile #45 exists in two versions in this game, I gave chosen one of these as the standard one, in line with the choice made in the PS18XX 1830 tile set. In other games, tiles have been printed in different orientations (i.e. with the tile number in a different position w.r.t the track). So we'll have to make the unrotated tile position game-dependent (it's a pity we can't move the tile number to different positions). The base rotation per game will be defined in TileSet.xml, I think by a <Tile> attribute like baseRotation="5" (the number 5 would apply to e.g. tile 7 in 1856). The purpose if this all is to keep the orientation indications of laid tiles, as printed in the OR and Log screens, in line with those used by PBEM players. Erik. |
From: John A. T. <ja...@ja...> - 2005-12-26 14:26:45
|
On Mon, 26 Dec 2005, Erik Vos wrote: > I have started checking the tile orientations in Tile Designer > vs. the files as provided with the games (1830 only so far). > (the standard is that the position of the tile number defines > either the S or the SW edge of the unrotated tile). > > It turns out that almost all 1830 tiles have a different orientation > in TileDesigner! However statistically unlikely it may be, the only > correct one (except the obvious tile #63) was tile #67! You will find that different games use different orientations of the same tile, and sometimes the same game uses multiple orientations of the one tile (I believe there is one in 1830 that has multiple orientations). My solution was to define a canonical orientation and store it in the database that way, and then for each game it could define a rotation relative to the canonical orientation (even multiple instances of the same tile in different orientations). -- John A. Tamplin ja...@ja... 770/436-5387 HOME 4116 Manson Ave Smyrna, GA 30082-3723 |
From: Erik V. <eri...@hc...> - 2005-12-26 15:05:27
|
> > I have started checking the tile orientations in Tile Designer > > vs. the files as provided with the games (1830 only so far). > > (the standard is that the position of the tile number defines > > either the S or the SW edge of the unrotated tile). > > > > It turns out that almost all 1830 tiles have a different orientation > > in TileDesigner! However statistically unlikely it may be, the only > > correct one (except the obvious tile #63) was tile #67! > > You will find that different games use different orientations > of the same > tile, and sometimes the same game uses multiple orientations > of the one > tile (I believe there is one in 1830 that has multiple orientations). Yes, tile #45. I have chosen the one that appears in the PS18XX tile set for 1830, as I think we should align with the PBEM community. > My solution was to define a canonical orientation and store it in the > database that way, and then for each game it could define a rotation > relative to the canonical orientation (even multiple > instances of the same > tile in different orientations). Yes, that is my approach too. The "database" now is Tiles.xml. I have just uploaded new versions of the tile XML files. However, after a quick check of your database and some of the games you have published I now see, that you use Marco's tile orientations rather than the ones that I found in the 1830 game box and in the PS18XX set for 1830. We clearly must make a choice which orientation set we will use. Whether or not a choice for whichever "canonical" orientation is important depends on the question whether or not we will display the tile numbers at all on the map. If we do, I think we should aim at displaying the tile number for each game in the correct place as often as we can. Now that you currently are the most prolific game publisher, the balance will shift your way inevitably... But IMO displaying tile (and revenue) numbers at all will only be worth while if we either succeed in implementing SVC graphics, or go for a separate gif file for each tile orientation (otherwise the number reading directions will rotate with the tile). And if we succeed in scaling the numbers such that these are readable and still do not overwhelm the map. I have my doubts on this aspect. At least we need a spreadsheet listing the orientations of all common tiles in all popular games against your database. Maybe I'll start one. Erik. |
From: John A. T. <ja...@ja...> - 2005-12-26 15:50:20
|
On Mon, 26 Dec 2005, Erik Vos wrote: > Yes, tile #45. I have chosen the one that appears in the PS18XX tile > set for 1830, as I think we should align with the PBEM community. I would think mirroring what is in the actual game is more appropriate. For example, in some cases the PS18XX tiles are totally unrelated to the orienation used in the original game (in some cases even the tile numbers, such as 18MEX, since they were developed before the game was published). > Yes, that is my approach too. The "database" now is Tiles.xml. > I have just uploaded new versions of the tile XML files. > > However, after a quick check of your database and some of the games you > have published I now see, that you use Marco's tile orientations rather than > the ones > that I found in the 1830 game box and in the PS18XX set for 1830. I assume you are referring to the web tile images? Those have not been regenerated yet. I am about 80% done with merging all the different tile sources into my database (kept in Informix using the schema I posted here earlier). Too many things going on, but it will get finished one day and then I will regenerate all the tile images. > We clearly must make a choice which orientation set we will use. I think the first question is what is the goal -- if it is to exactly mirror what came in the box, there are a lot of complicating issues such as using multiple orientations in the same game and (especially in 182x) different print runs having different orientations. If the goal is to support PBEM using only electronic maps, then choosing one orientation is best, even if you have to make clear that players using physical boards will have to carefully watch orientations -- otherwise, you have to find a way to disambiguate different tiles which have the same number but different orientations. > But IMO displaying tile (and revenue) numbers at all > will only be worth while if we either succeed in implementing SVC graphics, > or go for a separate gif file for each tile orientation (otherwise > the number reading directions will rotate with the tile). > And if we succeed in scaling the numbers such that these are readable > and still do not overwhelm the map. I have my doubts on this aspect. The way I intended to handle it was to automatically leave out components of the tile image depending on the size being drawn. That way, if you have a sufficiently large screen area or are zoomed in to the map, you will see all the detail available in the tile. If you zoom out, you see only the detail that is useful at that resolution. I haven't downloaded and built Rails since the tile code has been working, so I don't know exactly how you have implemented this. -- John A. Tamplin ja...@ja... 770/436-5387 HOME 4116 Manson Ave Smyrna, GA 30082-3723 |
From: Erik V. <eri...@hc...> - 2005-12-26 16:25:08
|
> > Yes, tile #45. I have chosen the one that appears in the PS18XX tile > > set for 1830, as I think we should align with the PBEM community. > > I would think mirroring what is in the actual game is more > appropriate. > For example, in some cases the PS18XX tiles are totally > unrelated to the > orienation used in the original game (in some cases even the > tile numbers, > such as 18MEX, since they were developed before the game was > published). I agree with that, but I would not go as far as preconfiguring different versions of the same tile in one game. The base orientation will be defined in TileSet.xml, and if people want to change the orientations defined there to match their own game box, that's fine, but not our concern. If a certain game has a consistent tile set of tile numbers, we'll use that. If not, we'll follow some common usage, e.g. Steve Thomas's PBEM tile sets. > > Yes, that is my approach too. The "database" now is Tiles.xml. > > I have just uploaded new versions of the tile XML files. > > > > However, after a quick check of your database and some of > the games you > > have published I now see, that you use Marco's tile > orientations rather than > > the ones > > that I found in the 1830 game box and in the PS18XX set for 1830. > > I assume you are referring to the web tile images? Those > have not been > regenerated yet. I am about 80% done with merging all the > different tile > sources into my database (kept in Informix using the schema I posted > here earlier). Too many things going on, but it will get > finished one day > and then I will regenerate all the tile images. OK. It would be great if we could leverage your database to provide tile sets tailored to each separate game, but we'll have to see how that works out. > > We clearly must make a choice which orientation set we will use. > > I think the first question is what is the goal -- if it is to exactly > mirror what came in the box, there are a lot of complicating > issues such > as using multiple orientations in the same game and > (especially in 182x) > different print runs having different orientations. If the > goal is to > support PBEM using only electronic maps, then choosing one > orientation is > best, even if you have to make clear that players using > physical boards > will have to carefully watch orientations -- otherwise, you > have to find a > way to disambiguate different tiles which have the same number but > different orientations. Yes, see above. The simple solution is one standard tile set for all games. Better, but perhaps less easy to create, would be a separate set for each individual game. > > But IMO displaying tile (and revenue) numbers at all > > will only be worth while if we either succeed in > implementing SVC graphics, > > or go for a separate gif file for each tile orientation (otherwise > > the number reading directions will rotate with the tile). > > And if we succeed in scaling the numbers such that these > are readable > > and still do not overwhelm the map. I have my doubts on this aspect. > > The way I intended to handle it was to automatically leave > out components > of the tile image depending on the size being drawn. That > way, if you > have a sufficiently large screen area or are zoomed in to the > map, you > will see all the detail available in the tile. If you zoom > out, you see > only the detail that is useful at that resolution. I haven't > downloaded > and built Rails since the tile code has been working, so I don't know > exactly how you have implemented this. We now have gif files with tiny numbers that are not readable at all. I have just created a new TileDesigner file with double size numbers, we'll see how that works after Brett has had time to create gifs out of that file. Erik. |
From: John A. T. <jat...@ja...> - 2005-12-26 16:52:02
|
On Mon, 26 Dec 2005, Erik Vos wrote: > We now have gif files with tiny numbers that are not readable at all. > I have just created a new TileDesigner file with double size numbers, > we'll see how that works after Brett has had time to create gifs out of that > file. I assume this is just a temporary solution and you still want to use the tile rendering solution I proposed, right? Pre-drawn bitmap images are not suitable for a scalable interface (not to mention different users having different resolution screens and preferences). -- John A. Tamplin ja...@ja... 770/436-5387 HOME 4116 Manson Ave Smyrna, GA 30082-3723 |
From: Erik V. <eri...@hc...> - 2005-12-26 20:29:39
|
> > We now have gif files with tiny numbers that are not > readable at all. > > I have just created a new TileDesigner file with double > size numbers, > > we'll see how that works after Brett has had time to create > gifs out of that > > file. > > I assume this is just a temporary solution and you still want > to use the > tile rendering solution I proposed, right? Pre-drawn bitmap > images are > not suitable for a scalable interface (not to mention different users > having different resolution screens and preferences). Absolutely! But even then we will need XML tile definitions like the ones we already have (Tiles.xml and TileSet.xml), because these will be used to enforce the rules. So these or similar XML files should somehow be derivable from your tile definitions. Erik. |
From: John A. T. <ja...@ja...> - 2005-12-26 20:50:22
|
On Mon, 26 Dec 2005, Erik Vos wrote: > Absolutely! > > But even then we will need XML tile definitions like the ones we already > have > (Tiles.xml and TileSet.xml), because these will be used to enforce the > rules. > So these or similar XML files should somehow be derivable from your tile > definitions. My personal preference would be to have a JDBC driver for access to game information, and one backend could be using XML or other text files. However, that may not have enough others who would use a database backend to justify the effort (I would have it directly access my database so I don't have to maintain multiple copies of the same data and worry about keeping them up to date). In any case, it is trivial to generate an XML file automatically from the database. -- John A. Tamplin ja...@ja... 770/436-5387 HOME 4116 Manson Ave Smyrna, GA 30082-3723 |