ID:1677913
 
It's necessary, but how can it be done?

One way is to supposedly use canvas.imageSmoothingEnabled = false, but I haven't been able to test this because the canvas uses the WebGL context instead of Canvas2D.

Another way could be to use canvas.style.imageRendering = "pixelated" in Chrome (beta 38), but that only seems to work with images in Canvas2D (I don't know how to use WebGL stuff...).

I've had success in Firefox setting canvas.style.imageRendering = "-moz-crisp-edges", but it doesn't work with map.zoom=0, it works when the canvas.width/height is the native view-size and the canvas.style.width/height is stretched.

Is there any discussion going on for how this will be done for us in the future? If it can't be done in Dream Seeker, that's fine, as long as the webclient will surpass Dream Seeker in performance (since it already surpasses it in UI features).
I was talking to Tom last night while we tested older boardgames and casual games for web-client bugs and he expressed that he thinks the web-client has the potential to deal with the various drawing issues and performance issues that are problem spots with the native DS client. Due purely to the fact that they're building this client from the ground up and not relying on old stuff that would be hell to revise, and the fact that the language/libraries the client are being made with are being actively developed and they've had luck getting issues resolved by the people working on the stuff.

It's just speculation of course, a 'it's possible' but personally, I can see the web-client outright replacing the native client at some point when the stuff gets hammered out and in good working order, because it has a ton more potential.
In response to Nadrew
Nadrew wrote:
It's just speculation of course, a 'it's possible' but personally, I can see the web-client outright replacing the native client at some point when the stuff gets hammered out and in good working order, because it has a ton more potential.
This is slightly off-topic, regarding the cool potential of the webclient: but I've got pointer-lock working and mouse movement detection (using MouseEvent.webkitMovementX/Y) for first-person games like Dawncaster (although that is scaled to poop due to map.zoom limitations) and whatever else. I copied the default web/map.dms, added some properties to the config, attached some event listeners, and you can get the mouse position, movement, pitch and yaw with winget(player, "map", "mouse-x; mouse-y; mouse-moved-x; mouse-moved-y; yaw; pitch").
I haven't done anything with the web-client myself yet, but that sounds really awesome.
In response to Kaiochao
Kaiochao wrote:
Nadrew wrote:
It's just speculation of course, a 'it's possible' but personally, I can see the web-client outright replacing the native client at some point when the stuff gets hammered out and in good working order, because it has a ton more potential.
This is slightly off-topic, regarding the cool potential of the webclient: but I've got pointer-lock working and mouse movement detection (using MouseEvent.webkitMovementX/Y) for first-person games like Dawncaster (although that is scaled to poop due to map.zoom limitations) and whatever else. I copied the default web/map.dms, added some properties to the config, attached some event listeners, and you can get the mouse position, movement, pitch and yaw with winget(player, "map", "mouse-x; mouse-y; mouse-moved-x; mouse-moved-y; yaw; pitch").

Keep in mind that you can always release your modified dms files as libraries that people can then include and use on in their projects. You'd just have to give it a new byondclass name.