At 2/7/17 10:06 AM, MSGhero wrote: Graduating with a masters this semester, (very) slowly still working on a procgen roguelite whatever game. We're closeish to a public demo, but it's currently hard to find the time.

I'm also on the job hunt. President Trump's order to stop hiring for government positions kinda fucked several of my prospects, so I need to apply to more. Haven't heard back from any yet, but I'm hoping I get a flood of calls soon.

Nice! Was your degree related to your hobby (game development?) or something totally different? I do remember you posting about making an RPG game w/ an element of randomness a few years ago if I'm not mistaken, I'm guessing it's that? Be sure to tell us when the demo does hit, haven't played any good Flash games (actually any at all) since 2008!

Yeah I'm in the same boat WRT to job hunting, looking for an internship this sem but most are unpaid. I think I'm better off just not doing an internship since I doubt I'll ever be able to find one in my current field (economics) and instead going on exchange. I'm still torn between the two: work experience or actually being able to travel for the first time in my life. How did you find doing your masters?

At 2/7/17 12:28 PM, GeoKureli wrote: Congrats, both of you!

I went to a startup right away, and I loved it, easy entry and organic growth in responsibilities. After 3 years I hated it, though, and left. There's a site specifically for startup job listings

This was the advice I was given by a mid 30 y/o working in the IT field. His advice was to pretty much go SanFran and get a job in a startup because I don't have much to lose in my 20s. I've been doing a double degree in economics and comp sci and was thinking of doing my honours in economics but I'm not certain it'd help me out in any way, does it actually help boost my employment prospects in either of my fields or is it just there for people who actually like research?

At 2/11/17 10:09 PM, slugrail wrote: Nice! Was your degree related to your hobby (game development?) or something totally different? I do remember you posting about making an RPG game w/ an element of randomness a few years ago if I'm not mistaken, I'm guessing it's that? Be sure to tell us when the demo does hit, haven't played any good Flash games (actually any at all) since 2008!

Yeah I'm in the same boat WRT to job hunting, looking for an internship this sem but most are unpaid. I think I'm better off just not doing an internship since I doubt I'll ever be able to find one in my current field (economics) and instead going on exchange. I'm still torn between the two: work experience or actually being able to travel for the first time in my life. How did you find doing your masters?

Nah, I'm in mechanical engineering. I have no time to work on my game though ughhh.

My masters has been a roller coaster of doing nothing and doing all things. It'll be worth it in the end, but it's a struggle because the foundation of research I'm building off of is bad. There's a hard problem that people literally ignore that I'm tackling, and it's just weird that only one other person has even mentioned it, much less dealt with it.

Clearly, I watch far too much YouTube.

I discovered something quite magical a couple days ago that I think you all might find interesting.

I spent way too much development time on my DisplayObject class. I finally simplified it.

Things are starting to come along, but still a long ways to go. I've been wrestling quite a bit with the engine I'm using which is never fun. I have had a really annoying performance issue again. This time related to the line of sight. Specifically with rendering it. I've stopping tinkering with it for now, but at some point I will need to revisit.

At 2/26/17 09:29 PM, PrettyMuchBryce wrote: This time related to the line of sight. Specifically with rendering it. I've stopping tinkering with it for now, but at some point I will need to revisit.

I'm curious. Wouldn't LoS be calculated with two rays, then a TileMap would simply render the tiles needed? Is the bottleneck the tile map?

At 2/27/17 12:30 PM, egg82 wrote: two rays

Two is totally sufficient. Said nobody ever.
I suppose it depends on the number of rays, but honestly raycasting is pretty cheap.

At 2/27/17 10:10 PM, egg82 wrote:

At 2/27/17 12:30 PM, egg82 wrote: two rays

Two is totally sufficient. Said nobody ever.
I suppose it depends on the number of rays, but honestly raycasting is pretty cheap.

I do a fair number of ray traces in my AI. Theta*, which is a modified A* obviously, lets you move any angle from one point to the next, but it requires a raytrace/convex cast to ensure the path is actually clear. I thought I would have a lot of issues with it, but it runs fine and is better than smoothing after pathing.

At 2/27/17 11:06 PM, MSGhero wrote: I do a fair number of ray traces in my AI. Theta*, which is a modified A* obviously, lets you move any angle from one point to the next, but it requires a raytrace/convex cast to ensure the path is actually clear. I thought I would have a lot of issues with it, but it runs fine and is better than smoothing after pathing.

You can also always, always make any sort of pathfinding or ray tracing faster with QuadTrees. Either way, can't imagine that's the bottleneck. Has to be the way rendering is done.

After this last month or so, I'm reasonably confident in my ability to optimize things.

At 2/28/17 02:01 AM, egg82 wrote: You can also always, always make any sort of pathfinding or ray tracing faster with QuadTrees. Either way, can't imagine that's the bottleneck. Has to be the way rendering is done.

My traces are connected to the information within the grid, so quadtrees wouldn't do anything for me since I'm just using array access and checking for overlap with that one tile. "Not in the same quad" doesn't really matter bc that tile is being checked to find a path. Then again, it's kinda implicitly using quadtrees where each one is a single tile...

At 2/27/17 12:30 PM, egg82 wrote: I'm curious. Wouldn't LoS be calculated with two rays, then a TileMap would simply render the tiles needed? Is the bottleneck the tile map?

It's not the algorithm on the CPU. I am using recursive shadowcasting.

https://github.com/ondras/rot.js/blob/master/src/fov/recursive-shadowcasting.js

The problem is the rendering. The game engine I'm using is all WebGL based, but it's all abstracted away from me, and my ability to write my own shaders using their API is limited. So using their API to do things like masking, drawing graphics, is fine for a while, but as the masked area grows it becomes frustratingly slow. Ultimately I may have to switch over to my own rendering engine if it gets bad enough. The engine is pixi.js, btw.

At 2/28/17 09:39 AM, MSGhero wrote: My traces are connected to the information within the grid, so quadtrees wouldn't do anything for me since I'm just using array access and checking for overlap with that one tile. "Not in the same quad" doesn't really matter bc that tile is being checked to find a path. Then again, it's kinda implicitly using quadtrees where each one is a single tile...

I guess my point wasn't QuadTrees so much as what they do. Reducing the size of any searching increases performance, period. So, yeah, basically the same thing. Smaller array = better performance. Always using the smallest search space in any given situation = best performance.

At 2/28/17 11:34 AM, PrettyMuchBryce wrote: It's not the algorithm on the CPU. I am using recursive shadowcasting.

This lead me to a really good article on FoV algorithms if you haven't seen it yet. Was actually incredibly interesting.

The problem is the rendering. The game engine I'm using is all WebGL based, but it's all abstracted away from me, and my ability to write my own shaders using their API is limited. So using their API to do things like masking, drawing graphics, is fine for a while, but as the masked area grows it becomes frustratingly slow.

Do light and shadows need shaders? You could write a new lighting engine. It's what I'm doing, at any rate :x

On a completely unrelated note, I feel really smart for choosing GCloud over AWS right now.

At 2/28/17 08:23 PM, egg82 wrote: On a completely unrelated note, I feel really smart for choosing GCloud over AWS right now.

I was pretty happy that use cdnjs.com as the CDN for all the sites I work on. Had I been using AWS, pretty much every site I've made would be a broken mess because that's where I serve jQuery and Semantic UI from. (And would have had many pissed off clients.)

Also, who's the retard at Amazon who thought it was a good idea to host the service that checks the status of S3 on S3?

At 2/28/17 08:23 PM, egg82 wrote: On a completely unrelated note, I feel really smart for choosing GCloud over AWS right now.

At 3/1/17 03:54 PM, Diki wrote: Also, who's the retard at Amazon who thought it was a good idea to host the service that checks the status of S3 on S3?

Hate myself for making this

At 3/1/17 04:22 PM, Sam wrote:

At 2/28/17 08:23 PM, egg82 wrote: On a completely unrelated note, I feel really smart for choosing GCloud over AWS right now.

At 3/1/17 03:54 PM, Diki wrote: Also, who's the retard at Amazon who thought it was a good idea to host the service that checks the status of S3 on S3?

Hate myself for making this

Even without any dependencies, your site would have gone down if it was hosted on AWS. (And a shitload of them did.)

Yes, I am taking a meme seriously. Come at me.

At 3/1/17 06:05 PM, Diki wrote: Even without any dependencies, your site would have gone down if it was hosted on AWS. (And a shitload of them did.)

Yes, I am taking a meme seriously. Come at me.

I just wanted to stay relevant with the youth

At 3/1/17 06:09 PM, Sam wrote:

At 3/1/17 06:05 PM, Diki wrote: Even without any dependencies, your site would have gone down if it was hosted on AWS. (And a shitload of them did.)

Yes, I am taking a meme seriously. Come at me.

I just wanted to stay relevant with the youth

Get off my lawn!

At 3/1/17 03:54 PM, Diki wrote: Also, who's the retard at Amazon who thought it was a good idea to host the service that checks the status of S3 on S3?

That was my favorite part of the whole thing. The irony was delicious! I also liked the part where everything was "increased error rates" - yeah, the error rates were increased. To 100%.

At 3/1/17 06:05 PM, Diki wrote: Even without any dependencies, your site would have gone down if it was hosted on AWS. (And a shitload of them did.)

Anything hosted in particular regions (yeah, multiple) would have gone down. So, the solution would be to have multiple copies of the site spanning across different regions. Possibly even different providers, but that's damned tricky.

At 3/1/17 07:48 PM, egg82 wrote: Anything hosted in particular regions (yeah, multiple) would have gone down. So, the solution would be to have multiple copies of the site spanning across different regions. Possibly even different providers, but that's damned tricky.

This is one problem that IPFS.io could potentially solve

At 3/1/17 10:00 PM, PrettyMuchBryce wrote: This is one problem that IPFS.io could potentially solve

Granted, I only watched the video, but what in the world is it supposed to be? I was shown an example of it working, but I have no idea what it's supposed to be or how it's supposed to work. Is it like TOR? It can't be, but maybe it borrows ideas from it? Is it entirely P2P? What about domains and IPs? How does it resolve files, and how well does it work with browsers?

What have you lead me to? @.@

At 3/1/17 11:38 PM, egg82 wrote: What have you lead me to? @.@

This should answer a lot of your questions:
https://github.com/ipfs/ipfs/blob/master/papers/ipfs-cap2pfs/ipfs-p2p-file-system.pdf

In terms of the browser story, there are some projects that are working toward integrating it with WebRTC. I think the ideal plan would consist of integration by the browsers themselves. (Supporting an ipfs:// protocol or some such). I believe the other questions you had are answered in the above paper.

At 3/1/17 11:38 PM, egg82 wrote: Is it like TOR? It can't be, but maybe it borrows ideas from it?

Not really. A closer comparison is BitTorrent with a dash of Git: instead of serving media from one place, it pulls bits from several (like BitTorrent) and it uses cryptographically secure hashes to keep a history record to prevent data loss (like Git).

At 3/1/17 11:38 PM, egg82 wrote: Is it entirely P2P?

Yes.

At 3/1/17 11:38 PM, egg82 wrote: What about domains and IPs? How does it resolve files, and how well does it work with browsers?

Domains and IP addresses would work the same as they currently do; serving media just wouldn't be reliant on HTTP. Seems to work fine with the Wikipedia logo and jQuery. But those are still ultimately using HTTP, probably because browsers don't natively support IPFS, but the idea is there.

I only really know the basics, but it's essentially an attempt to replace HTTP for serving hypermedia (which HTTP sucks at) and make websites more stable by decentralising them.

Heyo,

im not sure if this is the right place to ask this, but I was wondering why the Flash Ads for my newest game do not work.
I get the income from the Flash portal Ads for that game, as well as Flash ad income for the other game I have monetised, but not the Flash Ads for the newest one.

I dont depend on the few moneys the Flash Ads would generate, I just think its weird and dont know why its not working :)

At 3/2/17 01:24 PM, StuffedWombat wrote: Heyo,

im not sure if this is the right place to ask this, but I was wondering why the Flash Ads for my newest game do not work.
I get the income from the Flash portal Ads for that game, as well as Flash ad income for the other game I have monetised, but not the Flash Ads for the newest one.

I dont depend on the few moneys the Flash Ads would generate, I just think its weird and dont know why its not working :)

I see the ad loading so it seems to be working on that end - the rest comes down to CPMStar reporting back numbers, which you would think something would have been reported back by now but the ads are largely performance-based / low performing nowadays so it's possible they just haven't tracked anything to report there.

Let me know if nothing shows up in the next 48 hours though. You're sure you aren't using the API ID from a different project, right? Like if you had copy and pasted the ad setup from another game you made?

Oh thanks for the quick answer!

Let me know if nothing shows up in the next 48 hours though.

yeah, will do!

You're sure you aren't using the API ID from a different project, right? Like if you had copy and pasted the ad setup from another game you made?

Just checked, everthing is correct. I will just wait :)

At 3/2/17 12:23 PM, Diki wrote: I only really know the basics, but it's essentially an attempt to replace HTTP for serving hypermedia (which HTTP sucks at) and make websites more stable by decentralising them.

Faster multimedia, sorts like what Google does with putting YouTube servers directly into ISP's networked areas. Interesting idea, me-likey.

Also, I love documentation.

So today I learned that MySQL literally has a command-line flag called "--i-am-a-dummy" that prevents you from accidentally nuking your database.

Came across an old bit of example code I wrote to show to my friend to confuse him. I challenge anyone to make sense of this.

I was doing something very similar in real code once, but with around 150-200 lines of code between the two for loops, and that was a nightmare to debug. (I was writing code for work; I would never actually choose to write PHP.)

At 3/3/17 06:18 PM, Diki wrote: Came across an old bit of example code I wrote to show to my friend to confuse him. I challenge anyone to make sense of this.

StackOverflow to the rescue.