Return to “[Archived] Daily Dev Logs, 2012 - 2015”

Post

Week of July 21, 2013

#1
Sunday, July 21, 2013

Summary

(SIGGRAPH: Day 1)

Only attended the "preview" session today, where we got a sneak peak of all the cool technical papers that will be presented throughout the conference. Although there was a lot of cool stuff, I must admit that I'm a bit disappointed compared to last year. I feel like there was a lot more rendering and procedural-related stuff last year (which are my two main areas of interested at this conference).

Still, looking forward to diving in tomorrow :D

Also, it's really nice to see my fellow nerdy friends once again :ugeek:

PS ~ I expect dev log times to be later and more erratic than usual this week, both because I'm now on Pacific time, as well as the fact that my schedule will vary greatly from day to day. Sorry if it is bothersome to anyone :shifty:

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#2
Monday, July 22, 2013

Summary

(SIGGRAPH: Day 2)

Went to a few talks today, but the most interesting was, naturally, about ray tracing! I learned a lot about parallel algorithms for fast tracing, which was pretty cool to me since I've had an interest in ray tracing ever since the rendering competition. Unfortunately, nothing was really relevant to LT / my work in general. But tomorrow has a lot more real-time rendering stuff going on :)

Tonight at a social mixer, I got to meet a few of the CCP (EVE Online) people in-person!!! Ok, maybe it isn't that much of a big deal to most, but it sure felt cool to me. Getting to talk to a few of the guys behind those beautiful swaths of space...I just felt like a kid in the presence of childhood heroes. Cool guys :clap:

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#3
Tuesday, July 23, 2013

Summary

(SIGGRAPH: Day 3)

Awesome day! Really enjoyed the real-time stuff that went on. Among the things that I heard about today were techniques for lighting, depth of field, motion blur, bokeh, water rendering, and...a lot more! :D I was particularly excited about the water and lighting talks. The lighting talks got me thinking a lot more about lighting in LT and how I might implement a real lighting system. So far I have really not tackled this problem. To be fair, I really didn't think I would need a soluation for LT. Lighting in space is pretty simple. Still, it would be really nice to have, for example, lighting feedback when firing weapons or engaging thrusters. I learned some new techniques today that I think could actually be good solutions for LT. I'll let it propagate through my mind a little more then...we'll see :)

I was also really interested in the talk about water rendering. Not actually because of the water, but because it reminded me so much of the techniques that I was using when I first started exploring terrain rendering!! The guys giving the water talk introduced a "polar meshing" concept, which was exactly the first thing that I ever tried way back when I built my first terrain. I ended up running into the exact same problems that they described in the talk - but they managed to find some clever solutions that I never thought of! So it's possible that this technique could actually end up being relevant to planetary rendering in LT :thumbup:

Although it wasn't my favorite part of the talks, the presentations on DoF / bokeh and motion blur were given by one of the big engineers at Crytek!! It wasn't the best talk ever, but just getting to hear from someone who has contributed so much to CryEngine 3 - which, as you know, regularly causes me to burst into tears - was a great experience.

All-in-all, it was a really fun and informative day. I'm not expecting the rest of the days to be as cool, but one can always hope :)

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#4
Wednesday, July 24, 2013

Summary

(SIGGRAPH: Day 4)

Didn't really see anything that looked too interesting today. Luckily, I managed to meet up with Andy (since he lives in LA, which is just a hop, skip, and a jump away). He and another graphicsy friend and I sat around talking about procedural stories and AI for quite a while, which was pretty interesting. Andy and I also got to discuss a lot of interface-related concepts (in person, for once). As I said yesterday, I expected Tuesday to be the best day, and I think it was and will remain so. Still, it was totally worth it :D

Tomorrow I head back to the coding cave...I'm more eager than ever to get back to the real work :thumbup:

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#5
Thursday, July 25, 2013

Summary

Back home!! I just got in an hour ago, but I'm already gearing up for what I hope will be a great late-night / early-morning code-a-thon! :D

This time I actually had a bit of elbow room on my flights, so I managed to accomplish some coding..err..well...code ripping, that is :oops: I came to the conclusion that it's time to stop treating things like Persons and Factions so distinctly from entities. These things should be entities as well - they should use the component system (for example, a person having a name is no different from a ship having a name...it's the exact same concept, hence, should leverage the exact same code) and generally behave the same as other game objects. With this unification, I get to cut out a good bit of code :thumbup:

The coming days are going to be really intense (I hope). On the one hand, I'm soooooo close to having abstract simulation running. A few all-nighters might be all it takes to get me there. On the other hand, I have soooooo little graphical material to demonstrate for this month's update video that I'm really tempted to spend some time doing fast feature prototyping so that I can have some shiny stuff. Not sure which way I'll end up heading, but I know that, either way, the coding is going to be thick :geek:

And now, with this post, I declare the short, boring, and untimely dev logs to be 100% over! Thanks for your patience during that little SIGGRAPH-induced drought :)

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#6
Friday, July 26, 2013

Summary

Back to a solid day of work! Feels good :)

I spent roughly half of the day completing the ripping that I started yesterday, writing persons and factions in terms objects. I feel even better about this change than I thought I would! Creating a few new components and leveraging the component system allowed me to simplify a lot of the complexity behind the implementations (read: got to burn moar codezz!! :twisted: )

The most important result of all this work (which I failed to recognize yesterday when I started it) is that it will provide a unified mechanism for abstracting persons and factions! In fact, I think almost all of the abstraction work can be moved into the components, which means everything is very clean, well-defined, and automatic! So much better than my original architecture :oops:

I then moved on to writing the "real deal" AI planning engine to replace the placeholder one that I used to have. This is going to be fun :) The algorithm should be pretty straightforward. Making it run fast enough to achieve a decent level of intelligence...that will probably be the real issue.

No progress on graphics / new features today...maybe I'll do a feature sprint for the remainder of the night? We'll see. Many hours worth of coffee left before dawn breaks :thumbup:

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#7
Saturday, July 27, 2013

Summary

Bah...not what I was looking for from today. I hate saying that, but sometimes it's just the truth :oops:

Continued working on the AI algorithm, to the point that I am almost finished with what I think is a legitimate planning AI.

Attempted to start planet surfaces from scratch again, but for some reason I just wasn't in the graphics zone, and several things were bothering me - like not having the machinery for handling planet surfaces as clean as I'd like. It's really the same machinery as what is used for station interiors, but the dirtiness was bothering me, so I spent some time cleaning it up :geek:

Towards the end of the day, I switched over to the powerful desktop in hopes that the big screen and big GPU would inspire me a bit more than ye-olde-Macbook. Indeed, feeling the power under the hood and seeing LT on the big screen again lit a few fires in my heart, but these were quickly extinguished by some strange framebuffer-related GL bug that has (apparently) been recently introduced, and is causing ship textures to fail to generate on this particular machine/OS combo. And, since we're still encountering GL issues this far into development, you know what that means...

It's time for the weekly GL rant!!! Ok. Seriously folks. If you've followed the dev logs, you know how frequently I have struggled with this mythical "framebuffer" concept. Here's why: it's an artificial absurdity that has almost no real meaning or purpose - at least no meaning or purpose when exposed to the client. In my humble opinion, this is the largest mistake in GL's design (texture coordinates having inverted y clocks in at a close 2nd). It's an absolute pain. And the thing is, you can't say that it serves a purpose - because D3D (DirectX's 3D API) does it all without bothering the client with this. In D3D, render targets and what's being rendered to is an absolutely intuitive concept, even in the presence of multiple render targets. Where were you on that one, ARB?

Sorry about today :problem: Pushing through the night until this sucker is smited, after which I hope the graphics inspiration will flow :)

[ You can visit devtime.ltheory.com for more detailed information on today's work. ]
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Week of July 21, 2013

#8
Week Summary

General Sentiment

SIGGRAPH! What more is there to say? That magical time of year when graphics nerds across the world come out of their caves to share knowledge :D I didn't get as much out of this year as I did last year, mostly because there was a lot less rendering and real-time emphasis than I remember seeing last year. I wonder what next year will hold? ;)

Major Accomplishments
  • Attended SIGGRAPH 2013
  • Unified Persons and Factions with Objects, refactored them to use the component system
  • Rewrote the AI planning algorithm
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford

Online Now

Users browsing this forum: No registered users and 2 guests

cron