Lostcast 67


  • Patron

    The notes I have for this episode are HUGE, so, with no further ado:

    • Iconsole looks pretty interesting! I wonder why they choose to use intel processors…

    • Guys, don’t talk about GDC anymore, ok? It just makes me more angry that I can’t attend to it!
      #Argh!

    • Geoff, how didn’t you know about that Playstation trivia? What a shame… XD

    • The roundtable was interesting, but I wasn’t interested in the topic of discussion.
      I mean, this is how to make a cross-platform HTML5 game:
      Step 1: Make a HTML5 game.
      Step 2: Done.

      HTML5 and cross-platform are almost synonyms, I think.

      (of course, I know it’s not always easy as that, but you get the point)

    • About formfactors. screensizes etc… I think in your case you could easily switch to vector graphics and never have those problems again, thanks to @richtaur’s art style.

    • I know cross-platforming is not always the best business choice, but the joy of seeing your game running in various devices is unvaluable.
      I remeber a Socket.io experiment I made were I could move a square on the screen using the Ipad and see the effects in all computers at home at the same time.
      So much fun!

    • As Matt stated, most “solo” game projects were actually made in teams.
      Except for Pixel. He’s crazy! And I love him.

    • That team-size conversation sounded reminded me of Cogs Postmortem, where Rob talks about what he calls “Maximizing awesomeness per capita”. According to him, awesomeness per capita maxes out at a about 2-5 person team, and that’s totally understandable.
      It reminds me of Amnesia Fortnight, actually. Watching it made it pretty clear how hard it to get everyone in the same page in a 10-15 people team. I mean, they spent the first week just figuring out what they were going to, how they were going to do it and who was going to do what.
      I personally think the “magical number” is 5 people.

    • @geoffb game jam games can be very meaningful. I mean, have you played Gods will be watching?
      Also, I think they’re pretty good for exercising game design, scoping and having concise ideas.
      Actually, I think they’re perfect for that.
      Game jams are very cheap in terms of time, but being so short, they teach you how to maximize your eficiency.
      I’ve participated of two Ludum dares as of speaking. Both were what I would consider as failures, but it’s a totally wonderful experience.
      Even if it’s not for the game creation experience, the community is just too great!

    • Bastion and Cave Story were for me those “Wow, everything came together at the end” games.
      Really guys, if you haven’t played Cave Story yet, do it NAO! I can’t belive a game this good is completely free!

    • You talking about game design reminds me of Mr.Blow’s lectures.
      Seriously, they just rocked my world about game design.
      My personal favorites are Conflicts in game design, Design Reboot and How to program independent games.

    • That “having an idea which has already been done” talk reminded me of a joke my mom told me:
      Two friends went to a video rent store. They were reading movie synopsis and stuff when one said to the other: “Hey dude, wouldn’t it be cool if there was a video rent store that rented books instead of tapes?”

    • I think the thing about being original is somehow not worrying about being original.
      I mean, just follow your heart and try to do what you feel like doing. Don’t try to do something that some game did just because it went viral, but also don’t give up on your ideas just because someone already did that.
      I know that you have to make games for a audience, but don’t make a game solely for the audience, ya know.

    • I’m pretty 'effing excited about HTML5 Unity. The number of games I wanted to play but couldn’t because there’s no Linux Unity plugin and because I was too lazy to reset my computer is countless.

    • Geoff said that maybe WebGL wasn’t fast enough before to HTML5 Unity to happen.
      I think that’s not exactly true. I mean, saying that WebGL wasn’t fast enough is saying that OpenGL wasn’t fast enough.
      The thing is: Javascript was, and still is too slow to run a 3D game as complex as some Unity games are.
      The problem is that ASM.js wasn’t still there one year ago.

    • I know Unity content is only a subset of online games, but they tend to have a much higher quality than other online games.
      Also, it’s almost mandatory for Ludum Dare.

    • I would totally be using Unity right NAO, if it wasn’t for the Unityscript mess.
      I don’t wanna learn another language ATM.

    • I think at some point Geoff and Matt will be so far away from each other that their only communication will be Asana tasks.
      Still, they’ll find some way to work together…

    • I tried to work on a game remotely with a friend of mine once.
      He has a HORRIBLE 1mb/ps connection (which is more like 500k in Brazil).
      Needless to say it didn’t work out well.

    • Guys, TotalBiscuit and the Cynical Brit are the same person.

    • That story of how Geoff got to that game is pretty interesting.
      It makes sense that Steam is getting their reccomendation system better and better, as it will be crucial when Steam goes open.

    • I think the last time you mentioned Spelunky was two episodes ago or so…
      Congrats Geoff, you just ruined it XD

    • I think the most important thing about challange is knowing what the challange is.
      I mean, SMB: you know what your goal is, you know you can reach it, you know you have all the knowledge and tools to accomplish it.
      You “only need to.” “I only need to jump a little higher”, "I only need to wait half a second before walking here"
      Maybe that has something to do with Tropical Freeze.

    • Maybe you shouldn’t make a sidescrolling AWL.
      I mean, Zelda went that way, but Zelda 2 wasn’t the most popular Zelda ever. =P

    • Raga in AWL2 = Kirby+Yoshi.


  • Patron

    ###P.S:

    I don’t remember in which episode exactly, but I’m pretty sure you talked about how AWL got greenlit relatively fast.
    Well, just to make you jealous, I’ll point out that Crawl got greenlit in ONE FREAKING DAY!
    Guess Valve is trying to clean up Greenlight as fast as possible.


  • LDG

    I mean, this is how to make a cross-platform HTML5 game:
    Step 1: Make a HTML5 game.

    Pretty much!

    About formfactors. screensizes etc… I think in your case you could easily switch to vector graphics and never have those problems again, thanks to @richtaur’s art style.

    Maybe! There’s a lot to it. I was playing Escape Goat 2 recently and the resolution of the sprites changes, it zooms in or out depending on the size of the level. Even this amount of scaling caused the game to run noticeably slower for me. There’s a bunch of different resolution options that made a drastic difference in performance for me. Vlambeer’s games are super pixelated but I’ve seen complaints that they look blurry instead of crisp on some platforms like Vita. Going SVG would be interesting but it’s not as tantalizing close as it seems. I draw with simple shapes and colors but I’m firmly rooted in Photoshop, and we’d have to redo our rendering stack (again!) to support SVG. Sooo many hurdles!

    I personally think the “magical number” is 5 people.

    I could see that! I think we’d add an art guy and a Whippering guy to get to 5 ;)

    Game jams are very cheap in terms of time

    They can be, but from our experience a 48 hour game jam ends up costing us about twice that time, after factoring in exhaustion and switching contexts. Considering we’re making zero money during that period, it’s probably about a $1k loss for us…

    I think at some point Geoff and Matt will be so far away from each other that their only communication will be Asana tasks.

    Haha, I hope not!

    I think the last time you mentioned Spelunky was two episodes ago or so…

    It has been {{ ZERO }} episodes since we mentioned Spelunky :’(


  • Patron

    @richtaur said:

    About formfactors. screensizes etc… I think in your case you could easily switch to vector graphics and never have those problems again, thanks to @richtaur’s art style.

    Maybe! There’s a lot to it. I was playing Escape Goat 2 recently and the resolution of the sprites changes, it zooms in or out depending on the size of the level. Even this amount of scaling caused the game to run noticeably slower for me.

    Well, but SVG resising is hardware accelerated, so…
    IDK.

    Going SVG would be interesting but it’s not as tantalizing close as it seems. I draw with simple shapes and colors but I’m firmly rooted in Photoshop, and we’d have to redo our rendering stack (again!) to support SVG. Sooo many hurdles!

    There’s probably a script for exporting vector layers from Photoshop into SVG files out there.
    Yeah, re-doing the rendering code again would probably be painful XD

    Game jams are very cheap in terms of time

    They can be, but from our experience a 48 hour game jam ends up costing us about twice that time, after factoring in exhaustion and switching contexts. Considering we’re making zero money during that period, it’s probably about a $1k loss for us…

    Well, as you said before, switching contexts can actually be benefitial.
    But I mean, a game jam is like training, training all those things (like launching a game) which you rarely have the opportunity to train normally.
    How much would you pay for that?
    Also, game jam games can turn into hit successes, like Evoland or Mcpixel.
    And the amount of fun you can have participating in game jams is just ridiculous.

    I think the last time you mentioned Spelunky was two episodes ago or so…

    It has been {{ ZERO }} episodes since we mentioned Spelunky :’(

    Spelunky has been burned into your brains.


  • Tiger Hat

    The thing that bugs me about the “HTML5 vs Native” performance/cross-platform type discussion is that people seem to think that just because my device can run HTML5, it should also be capable of running anything that HTML5 can spit out. Which of course, simply isn’t true. No one would ever say, “my device is capable of C++ and OpenGL so surely I can run Skyrim on my phone right?” I think that on a high end gaming desktop there is no reason that you couldn’t use HTML5 technologies to develop a AAA game like Skyrim. Of course, there are prolly a lot of reasons that you wouldn’t, like wasting the time to make Skyrim work that way when you have perfectly good C++ tools (and presumably a really good UI). I have contributed to an HTML5 MMO called Ironbane and many players want to play on a Chromebook and complain about choppiness, etc. Well, that’d be great if it worked, but most Chromebooks simply don’t have the hardware to be game machines. This doesn’t mean that ChromeOS can’t handle it, just means that if you have crappy hardware you can’t expect things to run smoothly. That’s why I really like the idea of using node-webkit to deliver HTML5 games to desktop systems, just being in an actual browser gives people the impression that it should work wherever the same browser does. Somehow I think the good ol’ minimum system requirements stuff needs to find it’s way into HTML5 games.


  • Patron

    @Warspawn said:

    The thing that bugs me about the “HTML5 vs Native” performance/cross-platform type discussion is that people seem to think that just because my device can run HTML5, it should also be capable of running anything that HTML5 can spit out. Which of course, simply isn’t true.

    Hum…
    What bothers me is that there’s no way to know if your computer can run a game which is in the Chrome Webstore for example, without buying it.

    I have contributed to an HTML5 MMO called Ironbane and many players want to play on a Chromebook and complain about choppiness, etc. Well, that’d be great if it worked, but most Chromebooks simply don’t have the hardware to be game machines.

    I think showing a “your computer is too crappy to run this game” message would solve that, at least partially.

    That’s why I really like the idea of using node-webkit to deliver HTML5 games to desktop systems, just being in an actual browser gives people the impression that it should work wherever the same browser does. Somehow I think the good ol’ minimum system requirements stuff needs to find it’s way into HTML5 games.

    I don’t think so.
    I really like the idea of playing stuff on the browser and not having to install anything. The only point I see in using node-webkit for game is allowing the game to be distributed on other platforms, like Steam.
    Also, I think that people who are smart enough to know what “minimum requirements mean” will probably understand that their computers can’t run all web games smoothly.

    I think that on a high end gaming desktop there is no reason that you couldn’t use HTML5 technologies to develop a AAA game like Skyrim.

    Sure, you could probably port Skyrim to HTML5 using ASM.js, but Javascript is probably too slow to run Skyrim.
    That was my point: WebGL is not the problem, Javascript is.


  • LDG

    Yeah @warspawn I hear your frustrations. IMO, HTML5 is great for 2d desktop games. Mobile is currently just too slow across the board… HTML5 is probably OK for 3d desktop games too but I don’t have much experience there, and from the games I’ve played there isn’t much performance wiggle room.


  • Tiger Hat

    I suppose I’m just old school then, I don’t mind installing things. I also don’t mind having to install a Unity plugin for example. Everything still gets copied onto your hard drive and cached, in effect it’s really installing stuff anyway, so I’m not sure what the difference is. Plus hard drive space is dirt cheap, why not store copies on your computer and not have to download them again at some point? For smaller web games sure, there’s not a whole lot of difference if there is not a whole lot to load. But just think if AWL was web, you’d have much longer loading screens (or downsampled graphics).

    I guess my point about Skyrim is along the lines of HTML5 not meaning “lives in browser” with node-webkit for example you should be able to compile native libraries and link them up to JavaScript so that whatever slowness from js is done in c++ instead. I’ve actually thought that sometime in the future there would be not Web Browsers, but Game browsers instead. Much like node-webkit operates with much lower security restrictions. Overgrowth has sort of taken advantage of that as well, only partially. I believe they use Awesomium for the UI (which basically gives them HTML5 in a “real” game), and then Ogre3d for the hard core engine. All that Awesomium is, is Chromium that targets and OpenGL render surface (or maybe DirectX as well) since Chromium is open source. So, I think that (maybe this is a Blue Ocean) with the right tech, that instead of making just a game, you could make a Game platform, or Game Browser. The difference from the OS (tho prolly like ChromeOS there’d be some GameOS) is that it would have the one API for games, like HTML5 with a DX11 core or something.

    anyway… that got rambly :D


  • Patron

    @Warspawn said:

    I suppose I’m just old school then, I don’t mind installing things. I also don’t mind having to install a Unity plugin for example. Everything still gets copied onto your hard drive and cached, in effect it’s really installing stuff anyway, so I’m not sure what the difference is. Plus hard drive space is dirt cheap, why not store copies on your computer and not have to download them again at some point? For smaller web games sure, there’s not a whole lot of difference if there is not a whole lot to load. But just think if AWL was web, you’d have much longer loading screens (or downsampled graphics).

    Yeah, I don’t mind installing plugins, because I only have to install them once and update them once in a while.
    But having to download a installer, or worst, a zip file, opening it, selecting where to save things, deleting the installer after you used it…
    Argh!
    I just wanna type a url and play the game!
    With that said, it would be super cool if games could install themselves automatically once you open them in the browser.

    I guess my point about Skyrim is along the lines of HTML5 not meaning “lives in browser” with node-webkit for example you should be able to compile native libraries and link them up to JavaScript so that whatever slowness from js is done in c++ instead.

    I’m pretty sure that the only slowness about Javascript is Javascript itself.
    I mean, input lag is probably minimal, as it’s done in native code, WebGL is as fast as OpenGL, because it IS OpenGL…
    Maybe the WebAudioAPI has a slightly higher latency than using OpenAL directly, but I don’t think it is a major issue.
    The only thing that is holding devs from making a HTML5 Skyrim right now is that Javascript is just slower than native code.
    ASM.js is already way faster than Javascript, that’s why a HTML5 Unity is possible now, but it is still 1.5 times slower than native code.

    I’ve actually thought that sometime in the future there would be not Web Browsers, but Game browsers instead. Much like node-webkit operates with much lower security restrictions. Overgrowth has sort of taken advantage of that as well, only partially. I believe they use Awesomium for the UI (which basically gives them HTML5 in a “real” game), and then Ogre3d for the hard core engine. All that Awesomium is, is Chromium that targets and OpenGL render surface (or maybe DirectX as well) since Chromium is open source. So, I think that (maybe this is a Blue Ocean) with the right tech, that instead of making just a game, you could make a Game platform, or Game Browser. The difference from the OS (tho prolly like ChromeOS there’d be some GameOS) is that it would have the one API for games, like HTML5 with a DX11 core or something.

    I hope that doesn’t happens.
    I mean, having lower security restrictions would be great, but I don’t wanna have to download ANOTHER thing to play games.
    The browser market is already too fragmented.
    I don’t want to have to switch browsers if I want to play a game which doesn’t supports Chrome.

    anyway… that got rambly :D

    Keep it up! >:)


  • Tiger Hat

    Well, maybe it’s a crazy idea, or maybe it’s one of those ideas where I’m describing something that already exists. But what if say for example, Steam wasn’t just cloud hosting your games, but also providing developers with an advanced HTML5/Javascript/C++ type API to build against. Maybe it’s more like what ChromeOS does, I think you can likely access a lot more native features there (tho I haven’t looked heavily into it). Really the “cross platform” problem of fragmentation is already and always going to be there as long as we allow competition (and of course we should). I mean, what’s different than having to install Chrome or Firefox or IE or any of the other browsers which are just applications. I suppose maybe Windows 8 already has some of this as you can build “native” apps for the app store using HTML5. I guess what it would really take and mean is that the big game developers would have to start dealing with the browser fragmentation issues that the web developers have to deal with. Although they already have to deal with the OS fragmentation issues in a similar manner. Maybe what I’m thinking of is just another console type platform, like OUYA (admittedly I know little about that and could be way off) or maybe this is all possible with ChromeOS. Again, I don’t see what could be slow about JavaScript itself, bindings to things inside the browser yes, but if you’re just wiring up V8 directly to your C++ app as a scripting language then I would think it’s hardly slower than Python or LUA or TorqueScript, etc that other engines use.

    ok, there’s some more ramblings :)



  • @richtaur said:

    HTML5 is probably OK for 3d desktop games too but I don’t have much experience there, and from the games I’ve played there isn’t much performance wiggle room.

    I can maybe see the justification for 3D HTML5 game due to cross platform capability. However, if you developing exclusively for desktop, HTML5 doesn’t seem to make that much sense to me when you can make use of the faster, lower level, system languages available. I guess I haven’t really explored the realm of 3D HTML5 games either, but that is just my interpretation.


  • Patron

    @Warspawn said:

    Well, maybe it’s a crazy idea, or maybe it’s one of those ideas where I’m describing something that already exists. But what if say for example, Steam wasn’t just cloud hosting your games, but also providing developers with an advanced HTML5/Javascript/C++ type API to build against. Maybe it’s more like what ChromeOS does, I think you can likely access a lot more native features there (tho I haven’t looked heavily into it).

    I think the last thing we need is another freaking API.
    What I think should be done is giving acess to low-level OS APIs, if the browser detects (somehow) that the application is safe.

    Really the “cross platform” problem of fragmentation is already and always going to be there as long as we allow competition (and of course we should). I mean, what’s different than having to install Chrome or Firefox or IE or any of the other browsers which are just applications. I suppose maybe Windows 8 already has some of this as you can build “native” apps for the app store using HTML5.

    Yeah, sure, we will always have fragmentation problems with browsers, but at least they try to implement the same API.
    It wouldn’t bother me having to install a “game browser”, what would bother me would be having to switch between browsers because a certain game doesn’t supports the browser I’m using.
    That’s what’s, for me, the biggest feature on Steam: Unification. All games I want are probably there.

    Again, I don’t see what could be slow about JavaScript itself, bindings to things inside the browser yes, but if you’re just wiring up V8 directly to your C++ app as a scripting language then I would think it’s hardly slower than Python or LUA or TorqueScript, etc that other engines use.

    Actually, no.
    “Bindings to things inside the browser” is pretty 'effing fast. There’s probably no big performance differences between using Webgl on Chrome and hooking up Javascript to OpenGL using Node.
    Sure, if you have a C++ game engine and you’re using Javascript just for scripting entities, performance shouldn’t be a problem.
    But if you’re hooking up Javascript directly to OpenGL/OpenAL/SDL or something, it will be certainly waaaay slower than native code.

    Javacript is probably as fast as Python or Lua, or even faster in some cases.
    But no scripting language is at par for hand optimized C or C++ code, in regards to performance.

    ok, there’s some more ramblings :)

    Love it!


Log in to reply