Independent game dev, formerly: @GoogleDevExpert (UX + Product), PM at Voodoo games, QA at Electronic Arts. Tweets about games, WebGL. Working on a new game.
@curtastic Sorry for the late reply, honestly after four years I'm wondering about this myself.
I'll think about it and come back in another four years.
@Jakob_Wahlberg@Sjupp I started suspecting that even real players have their nicknames edulcorated for moderation purposes, I've rarely seen so many "clean" nicknames in a multiplayer game.
@DarkNemo We received over 200 applications for a 3d artist job and we're a 2-people studio making web games, 70% from one of the largest schools in France. Our intern from the same school told me he joined the school because it promised an apprenticeship. I wonder how many actually get it
@gpt3_bot @00davehill00 Also a deal-breaker for me was always the inability to move the view around while applying transformations in Blender, a big issue for precise level design.
@00davehill00 There are many projects to make some type of middleware to ease all this, but afaik all in very early stages. The landscape is still a wild jungle you have to cut through with a machete.
@00davehill00 Here's what we do (not recommended): Sketchup for faster level design with naming conventions for everything (triggers, base colliders, impostors, mesh visibility, AO bake inclusion).
@00davehill00 I would gladly get rid of Sketchup and all the file format headhaches if only Blender had a better inferencing and snapping system. As @gpt3_bot mentioned userData would be a better choice than naming conventions.
@00davehill00 At some point we make an atlas of material diffuse colours and textures into a single image: 1 material per macro-block = 1 draw call. A custom webgl shader takes care of offsetting the UVs to tile texture quadrants from the atlas.
@00davehill00 After some file format hoops, we de-instantiate components and merge mesh groups in JS through puppeteer, a headless Blender runs a python script to merge vertices, create a second uv set, smart unwrap and pack with uv packmaster. AO is baked for each macro-block.
@bentossell Someone said that studying successful startups is less useful than studying failed ones. Whereas success stories tend to be relatively unique and hard to reproduce, reasons behind failures fall into common categories that can be spotted and avoided.