[QUOTE=Karmah;47454792]I mostly finished up my animation editor.
[vid]http://i.imgur.com/EWv6U1q.gifv[/vid]
Aside from bug fixing, all I have left to do is implement manipulation of multiple keyframes at once, as well as think of some way of handling different interpolation modes (linear/cubic/etc)
[B]Edit:[/B]
Bah, the gif to web thing doesn't work for me[/QUOTE]
In case this is rendered directly:
I think it would be a lot nicer if you manipulated the fragment depth of the draggable nodes to be closer proportionally to the distance (squared) from their centre, and pick accordingly.
That way it would be easier to reliably grab nodes at a certain position, and very close nodes would still both have a decent visible area.
(Alternatively a "click through" method to cycle through the nodes layered at a certain position would be useful.)
[editline]4th April 2015[/editline]
[QUOTE=Darwin226;47455560]Is there no way then to have some static analysis determine if the program causes a stack overflow?[/QUOTE]
Not if you run it while parsing (and/or have an "eval" method) :rolleyes:
Today I migrated my game engine/game from using Box2D for collision to a proprietary system I coded up that uses quad-trees/SAT collision checking. Couple of days of coding without any testing, so I expected to have to fix a hundred problems when I eventually did run it. Nope, it ran smooth as butter the first time and everything went well, this is like winning the coding lottery.
[url]https://github.com/RickDork/Sora[/url]
[QUOTE=RickDork;47458407]Today I migrated my game engine/game from using Box2D for collision to a proprietary system I coded up that uses quad-trees/SAT collision checking. Couple of days of coding without any testing, so I expected to have to fix a hundred problems when I eventually did run it. Nope, it ran smooth as butter the first time and everything went well, this is like winning the coding lottery.
[URL]https://github.com/RickDork/Sora[/URL][/QUOTE]
That is the single most satisfying feeling in programming for me, coding for days without testing sounds like a recipe for disaster though ha.
[QUOTE=Socram;47458437]That is the single most satisfying feeling in programming for me, coding for days without testing sounds like a recipe for disaster though ha.[/QUOTE]
I hate it when you can't test stuff properly because loads of other systems depend on it, so you need to get it all right the first time.
Makes it that much more satisfying if you pull it off though.
I love what Xamarin do but this annoys me
[IMG_THUMB]https://s.paste.ninja/2015-04-04_23-16-09.png[/IMG_THUMB]
[QUOTE=ben1066;47458614]I love what Xamarin do but this annoys me
[IMG_THUMB]https://s.paste.ninja/2015-04-04_23-16-09.png[/IMG_THUMB][/QUOTE]
I'm not sure if you are a student, but if you are I recently (like... 3 days ago) finished the process of getting a student indie license, and it was relatively painless.
If you have a .edu email all you need to provide is one additional proof of enrollment with an expiration date. If you don't have a .edu, you need two proofs with expiration dates.
It looks like they process them manually, so it did take a couple of days to get everything taken care of though.
[QUOTE=Socram;47458765]I'm not sure if you are a student, but if you are I recently (like... 3 days ago) finished the process of getting a student indie license, and it was relatively painless.
If you have a .edu email all you need to provide is one additional proof of enrollment with an expiration date. If you don't have a .edu, you need two proofs with expiration dates.
It looks like they process them manually, so it did take a couple of days to get everything taken care of though.[/QUOTE]
That's precisely what I have, however, you now get Visual Studio with the Starter license, and lose it with the Indie license. Additionally, if you want to use C# 6 it seems that currently you either need to be, on OS X or use Visual Studio as Xamarin Studio on Windows uses MSBuild 12 and when I try and use Mono instead I get
[CODE]Build failed. Unexpected binary element: 0
Build: 1 error, 0 warnings[/CODE]
[QUOTE=ben1066;47458798]That's precisely what I have, however, you now get Visual Studio with the Starter license, and lose it with the Indie license. Additionally, if you want to use C# 6 it seems that currently you either need to be, on OS X or use Visual Studio as Xamarin Studio on Windows uses MSBuild 12 and when I try and use Mono instead I get
[CODE]Build failed. Unexpected binary element: 0
Build: 1 error, 0 warnings[/CODE][/QUOTE]
I don't know if you can get the newer versions yet but I got Visual Studio 2012 professional from dreamspark back before they required you to link a microsoft account. Only needs an .edu email.
[QUOTE=helifreak;47458869]I don't know if you can get the newer versions yet but I got Visual Studio 2012 professional from dreamspark back before they required you to link a microsoft account. Only needs an .edu email.[/QUOTE]
I mean that Xamarin gives you access to their VS integration with their Starter license, but not Indie. I'm using VS2015.
Just starting to get somewhere with the level editor for some contract work I am doing.
[img]http://i.imgur.com/YwmbpRg.png[/img]
The new Unity UI makes stuff like this so much easier to get working, if anyone hasn't looked into it yet you really should.
Some awesome news I want to share, there is a very good chance this contract work is going to turn into a full time position in Austin! And I currently have a really nice job offer here where I currently go to university, but my fiance and I are less inclined to stick around.
Might get emotional here for a bit, sorry if this is a little preachy. Just to be clear, I'm not bringing all that stuff up to just to brag or whatever, and I know there are plenty of people in here getting paid to make the awesome stuff that gets posted every day. I'm just really excited (and quite honestly blown away) that I get to do something I have loved to do for years as a hobby as my career. I wanted to thank all of you guys for being a huge part of the reason I stuck with game development over the years and always being an amazing and supportive community. Honestly, if it wasn't for gmod and facepunch I probably never would have gotten into programming in the first place.
On that note I also wanted to encourage all of you, regardless of your stage in life or game dev experience, to take what you do here seriously. I know so many people who completely bailed on their "dumb hobby" (games, art, music, etc) and ended up pursuing something mundane and safe, or something they thought would make them rich the quickest. A shocking number burned out, and a lot of the rest just hate what they're going to do, which they're kind of stuck with for the next 40 years of their life. If you play your cards right and keep at it, game development as a career is a very real possibility for everyone here. Demand is high and ever growing, and the pay ain't bad to boot. Whether you're just starting to understand hello world or you have a couple games under your belt, [B]just keep at it.[/B]
You guys all kick ass, thanks again.
[/emotionalCrap]
[QUOTE=Dr Magnusson;47457879]What's going on here?[/QUOTE]
What's going on [I]anywhere[/I]
What even is a bananaLemon
[code]
if 12.433432530290641 then
return BananaTangerine
[/code]
This is something I'd expect the Deep Thought computer in The Hitchhiker's Guide to the Galaxy to respond with
I'm working on fleshing out an idea I had that utilizes a scaling dimension in addition to spatial dimensions to potentially solve the problem of only being able to have objects similar in size without having large objects be floaty and tiny objects plain impossible
I'll probably end up writing a lot of special physics myself, because it's gonna take some funky fractal spatial partitioning and special cases to make the interactions between objects of very different sizes more realistic. Just from my really basic formulation so far, I'm seeing some cool possibilities for a different kind of collision detection popping up, which in 3D will involve projecting tetrahedrons in 4D or something. But basically you can tell if a big thing is imminently going to collide with something smaller
I fixed some bugs and got the now playing working to an extent. There are some other minor bugs I need to fix before I put out another update.
Screenshot from my OPO:
[img_thumb]http://i.imgur.com/adPp0Xm.png[/img_thumb]
I wanted to switch gears a little and try to get some of the Icons and UI straightened out because it doesn't look good at all.
Last night I spent some time trying to create an Icon but I wasn't very pleased with the results:
[img]http://i.imgur.com/nmYkMvD.png[/img]
Besides not looking very good, it doesn't really fit Android's design principles so I'm giving it a second shot in Inkscape. I haven't really got anything on the canvas yet but it's starting to look like guideline hell..
[img_thumb]http://i.imgur.com/TBKx0zB.png[/img_thumb]
So I've been messing around with neural nets again, I revisited something I was working on for an assignment last year, which was stock value prediction.
The way is works is that you train a neural network to associate the values of a particular stock each day over the past, say, 6 days with the value of the stock on the next day, day 7 in this case. It thereby learns to predict what the next day's value will be, given the values over the past 6 days.
Last night, I got really excited when I discovered I could accurately predict whether the FTSE100 index would rise or fall:
[IMG]https://dl.dropboxusercontent.com/u/10518681/Screenshots/2015-04-05_02-59-33.png[/IMG]
(red is actual FTSE100 index, blue is what the neural net predicted)
I didn't believe it. With this accuracy, I could genuinely make millions.
So, I set out to prove myself wrong. Common sense told me that this shouldn't work as well as it did. I asked my friend to check over what I was doing in my MATLAB scripts, and he didn't spot any mistakes. We racked our brains and eventually came to a tentative conclusion that we've got something we could use to make a decent return on the stock market. We spent a good deal of time discussing trading strategies and figuring out what our expected returns might be if we invested over the year.
We talked for a while, discussed our potential fortunes and dreamt of the yachts we'd buy. I went and made some dinner, and I came back to my desktop with a fresh mind. It's at this point I spot my fatal error. All my dreams I'd built up over the past few hours came crashing down as I realize: I've been plotting my predicted values one day in advance of what they should be. [IMG]https://dl.dropboxusercontent.com/u/10518681/reply/emot-sad.gif[/IMG]
Instead, my prediction values were [I]actually[/I] corresponding to a day later than what I thought. I.E, my predictions were offset by 1 day (or if you want to think of it another way, my predictions were based on future data that I wouldn't have available to me if I was predicting the stocks for real)
So instead, my prediction looks like this:
[IMG]https://dl.dropboxusercontent.com/u/10518681/Screenshots/2015-04-05_02-53-05.png[/IMG]
(again, red is actual, blue is predicted)
Interestingly, the neural network is weighting the value of the previous day significantly higher than any other days in the 7-day sliding window. What it's effectively saying is "I reckon tomorrow's value of the FTSE100 will be pretty much the same as today!". Sounds like a naive prediction when you think about it, but actually it's a pretty safe bet, especially with something like the FTSE100 where changes on a day-to-day basis are typically fractions of a percentage. So, if you asked me what the value of the FTSE100 would be tomorrow, I could tell you it will be the same price as it is today and I wouldn't actually be far from the truth.
So anyway, moral of the story: if it's too good to be true, it probably is. And also, [URL="http://garry.tv/2013/03/15/oil-your-brain/"]oil your brain[/URL]. I believe that is what eventually led me to discover my dream-crushing error.
[QUOTE=Trumple;47459769]stock prediction wizardry[/QUOTE]
So, just shift the graph back one day and you're good! :v:
[QUOTE=Trumple;47459769][...]
So anyway, moral of the story: if it's too good to be true, it probably is. And also, [URL="http://garry.tv/2013/03/15/oil-your-brain/"]oil your brain[/URL]. I believe that is what eventually led me to discover my dream-crushing error.[/QUOTE]
You couldn't have done it anyway with a single day of prediction because afaik there are pretty harsh regulations on day-trading that require a lot of money up front.
It might have been worthwhile if it could predict about a week in advance, but with less time you probably can't just waltz in there and constantly buy and sell things as if you're in a vacuum :v:
Was doing some work on multithreading for a CS assignment (introduction to threads, they're pretty cool)
We were to make a method that summed data in an integer array using threads (partitioning the array, and using individual threads to sum those parts, and returing the partial sum).
We didn't have to, but I ran 20 tests each on random arrays with varying threadcounts / array sizes, and exported the data to an Excel spreadsheet:
[IMG]http://puu.sh/h1WUi/0dea8bfa31.png[/IMG]
I discovered that the overhead caused by the creation of threads (at least in the JVM) is actually slower than the concurrency it provides. However, that concurrency starts speeding things up when you get into larger data sets!
Just wanted to share my results with you guys (:
[QUOTE=BackwardSpy;47455985]I taught myself Ruby yesterday and today I wanted to reinforce what I had learned, so I wrote a simple IRC bot.
Can anyone who uses Ruby have a look over my code and point out any style problems (going by general opinion) or things I could have done in a more idiomatic way? Thanks!
[url]https://gist.github.com/BackwardSpy/eeceeedf0f79ab5a3894[/url]
I'm enjoying the language so far, it seems very readable and quick to write.[/QUOTE]
Commented ;)
I started writing the software for the microcontroller portion of my CarPC project. I'm using a Teensy 2.0 board, which seems to be a lot more flexible than a regular Arduino when it comes to emulating input devices (With the exception that the emulated keyboard media keys don't work on Windows... For some reason...).
I wanted to emulate my car stereo remote, since I may have the touchscreen positioned in front of my head unit and I didn't want to adjust the screen every time I needed to make the volume louder or toggle the bluetooth setting. Luckily, I had some parts laying around that let me load up some example code for the IRRemote library that spat out the Sony remote codes as I pressed the buttons on the remote. I saved those and hooked up an IR LED, which happily played the codes back. I haven't tested them in my car yet, though I programmed my Flirc dongle with the car stereo remote and it seems to like the emulated codes just fine.
It took a while to finally figure out an easy way to write/read serial commands using C#. The examples I found were way more complicated than they needed to be, though I ended up finding a simple one on Instructables of all places. I got the basics down, though I need to add some error handling and command verification before I'll be completely happy with it.
Later, I'll hook up a couple temperature sensors and get them updating the values in the C# application.
[t]http://i.imgur.com/Ae6EeiL.jpg[/t]
Just wondering, but how long does it take you guys to load Sponza as an .obj?
I'm redesigning the major tool tab of my level editor, and settled on this as a design:
[t]http://i.imgur.com/cMQje8F.png[/t]
Anyone have any ideas for it?
[QUOTE=Ac!dL3ak;47460097]Commented ;)[/QUOTE]
Thank you, that is very helpful!
Tbh I never understood what kind of psycho wants to be a game dev for life. The field is the hell of IT. Absolutely unstable (you never know if the game is gonna shoot or blow), hard to get into, tight circles, geographically almost US and Britain only, full of terrible business practices, plus all the problems of a B2C thing.
Like, you can be a web dev or software dev, make more money with more stability and just as interesting tasks, find a job anywhere in the world So you can switch to a new country every year (I know people that do), easily get into the startup scene if you are into it, easily requalify any day, easily move to management. And make games in your spare time of course.
[QUOTE=Socram;47458765]I'm not sure if you are a student, but if you are I recently (like... 3 days ago) finished the process of getting a student indie license, and it was relatively painless.
If you have a .edu email all you need to provide is one additional proof of enrollment with an expiration date. If you don't have a .edu, you need two proofs with expiration dates.
It looks like they process them manually, so it did take a couple of days to get everything taken care of though.[/QUOTE]
Sadly, this happens with a student license as well :(
[QUOTE=Trumple;47458963]What's going on [I]anywhere[/I]
What even is a bananaLemon
[code]
if 12.433432530290641 then
return BananaTangerine
[/code]
This is something I'd expect the Deep Thought computer in The Hitchhiker's Guide to the Galaxy to respond with[/QUOTE]
Jokes aside, generating type-wise valid programs is super hard. Generating SEMANTICALLY valid programs is probably venturing into machine learning and what not.
Fortunately, the parser only cares about grammar so that's not a problem.
[QUOTE=Berkin;47459821]So, just shift the graph back one day and you're good! :v:[/QUOTE]
Berkin, with this brilliant idea we'll [del]lose[/del] make millions! I'll be sure to send you some of the resulting [del]crippling debt[/del] fortunes.
[editline]5th April 2015[/editline]
[QUOTE=Tamschi;47459932]You couldn't have done it anyway with a single day of prediction because afaik there are pretty harsh regulations on day-trading that require a lot of money up front.
It might have been worthwhile if it could predict about a week in advance, but with less time you probably can't just waltz in there and constantly buy and sell things as if you're in a vacuum :v:[/QUOTE]
Had a quick look into this, I had no idea there were regulations on such a specific form of trading. I can't find any relating to the UK but I'm sure there are equivalent rules to the ones in the US.
Thanks for the heads up anyway. I was going to have a go at predicting a week in advance today so I'll let you know how it goes!
So this random generation of arbitrary data is really freakin' neat so I implemented a similar thing in C#. Is uses a lot of functional stuff in the background ([URL="https://gist.github.com/LukaHorvat/c63af608a28fac017977"]it's C#, I swear![/URL]) but it's really simple to use and (so far) completely type safe.
Here's how you generate a list of dice rolls.
[img]http://i.imgur.com/hDAFO01.png[/img]
[editline]5th April 2015[/editline]
How about random sentences?
[img]http://i.imgur.com/9YgfcdL.png[/img]
The random number generation is currently very bad (a lot of 'm's), but that's an easy fix.
I decided to get moving on with some GUI as I was starting to get very bored reading and watching courses involving console programs. so far, I'm having loads more fun as the result is a lot more noticeable than a console program.
as a test go, I made a program which opens a window with a single button that leads to a picture of a clown.
[IMG]http://pred.me/pics/1428250946.png[/IMG]
[t]http://pred.me/pics/1428250961.png[/t]
[url]https://github.com/predme/Honk-For-Clown-JavaFXTest[/url]
super basic stuff, but still a little complicated as syntax and so on doesn't quite sit in the muscles yet. on the topic of JavaFX, do you guys know of any good sources to learn from? all I was able to find which was more tutorial/course oriented than documentation is TheNewBoston's JavaFX tutorials which I personally think isn't too easy to follow.
Update on my VConsole for Source 1:
[t]http://i.imgur.com/7Ea0NFW.jpg[/t]
Switched to use C++'s ListView for better presentation, sadly it doesn't support per-item coloring :/
Hi facepunch, I did an LED thing for my RPi
[vid]http://a.pomf.se/ufefuz.webm[/vid]
The code is an absolute mess but I'll try my best to make it presentable :v:
The song is [URL="https://www.youtube.com/watch?v=MGKFLGrmjfw"]C2C - Down The Road[/URL] for the curious
Sorry, you need to Log In to post a reply to this thread.
