Jump to content

  • Log In with Google      Sign In   
  • Create Account

Banner advertising on our site currently available from just $5!

1. Learn about the promo. 2. Sign up for GDNet+. 3. Set up your advert!


Member Since 05 Oct 2002
Offline Last Active Aug 23 2014 01:22 PM

Topics I've Started

Can you help my game analytics startup win a competition?

15 April 2012 - 11:14 AM

Hey everyone!

First of all I want to apologise for the spammy post, if this is against the rules or otherwise a "no no" thing to do, mods, feel free to close or delete or whatever it is thats done in those situations. With that said...

My partners and myself have entered our (game dev related) startup in an Irish startup competition and we need all the votes we can get! The competition was basically to enter a 100 word pitch and get as many votes as possible before 18:00 GMT tomorrow (Monday 16th). At the moment, we're about 250 votes behind, so we need help to climb back to the top before the competition closes tomorrow. You're help is very greatly appreciated!!

My startup is Prescience AI

You can vote here: https://www.facebook...12044075473841/
(you need a Facebook account)
You can read the pitches/intros here: http://on.fb.me/HEwXvv

For convenience, I have copied our pitch below.

PS: If our product is something that would interest you as a game developer, feel free to sign up on the website or Facebook page or PM me here. We're still in the early stages, but the game developers we have spoken to so far are very positive about what we have done and are doing!
We're also looking into developing a free or really cheap package for indies and hobbiests, though, again, its too early for that still.

Thanks very much!!!


website and Facebook page
Our 100 word pitch:

Hi all, the advent of Prescience Actual Intelligence (PrescienceAI) marks a new frontier for the gaming industry. We provide a solution for social/mobile/web game developers to significantly increase their in-game revenues. Our solution enables developers to monitor the behaviour of successful players that have purchased items and feed this information back to new players in real-time. The information is fed back into the game as hints, dynamic content changes, dynamic AB testing, and gameplay tweaking (eg difficulty level, number of enemies/obstacles, availability of powerups). These on the fly game changes increase purchases made by providing the gamer with an enjoyment optimised gaming experience Posted Image

AI/Analytics middleware for online games - I need some feedback/advice

03 March 2012 - 02:29 PM

Hey fellow gamedevvers!

I've been a long time member of gamedev (been lurking since about 2000 or early 2001), though I've never been a frequent poster. I am currently at Startup Weekend in Dublin. For anyone not familiar with Startup Weekend, the idea is that people form small teams and basically launch a startup - that is, come up with a concept, validate it, build a business model, build a proof-of-concept product, market it and pitch it to venture capitalists - all in 52 hours!
My group is working on a game AI/analytics middleware project for competitive online games and I was hoping some people who either work in game development or are developing games as a hobby might have some input for us before we pitch the startup to VC's tomorrow evening.

Our goal is to build a platform which would integrate into an online game to provide more human-like AI by analysing gameplay data from actual human players and generating data to help AI's in their decision making logic. We would include both a system which determines what the most human-like action would be for a given scenario in the game (this is what we are prototyping this weekend) and an analytics system which records gameplay data from the games - this would be used for both the AI and also to provide the game designers with feedback on how the players play the games.

We've already spoken to a few developers (including the founders of Demonware) but would love some more ideas and feedback.

Basically, what I would like to know is
  • do you or have you used middleware for AI and if so, which?
  • what were/is your experience with AI middleware - did it leave a generally good impression?
  • if you were building a multiplayer online game, would you be interested in middleware for AI?
We have also set up a landing page and facebook page, in case anyone is interested. If you like this project, I would love if you could "like" the facebook page, as it helps us during the team evaluations tomorrow. We also have a questionnaire that we would love people to fill out if anyone has a few minutes of time.

Website: prescienceai.com
Facebook page: www.facebook.com/PrescienceAi
Questionnaire: www.surveymonkey.com/s/823KXFK

Any and all feedback greatly appreciated!

- Dan.

Show gamedev: I just finished a custom MIDI controller project

12 May 2011 - 07:39 PM

WARNING: shameless plug post. Please ignore if you don't like.

Since its not in any way game related, I don't think this is worthy of a "Your Announcement" thread. I just wanted to share this project I've been spending a lot of time on (though my gamedevving has suffered because of it Posted Image), perhaps some of you will find it interesting?

For the past three months (including down time when we accidentally fried the main board and had to wait 4 weeks for a replacement) myself and my brother, who's a DJ, built our own custom MIDI controller. Well, semi-custom: its a hacked and modded Midifighter (schematics and firmware of the Midifighter are open source). I did all the firmware programming, my brother assembled all the components, did the case and is working on Traktor MIDI mappings for this thing.

This morning, after a 12-hour all night hackfest, we finally completed it! I will be releasing the code within the next two or three days and we will eventually get around to blog about how we did it. In the meantime, here is a few preview photos and a very brief video showing the controllers "demo mode" (basically just cycling LEDs). The original announcement that we were working on this project can be found here on the DJ Tech Tools forums.

Besides a lot of solder, the project consists of the AT90USB162-based Midifighter, which was programmed in C, and a PIC24HJ32GP302, which I programmed in PIC24 assembly. This is being used as a simple LED driver, communicating with the AT90 over SPI.

If anyone is interested, feel free to fire me any questions and I will try to answer them, if I can. Please let me know what you think, I'd love any feedback I can get (good and bad Posted Image). What details would you most be interested in when I get around to write a blog post or two?

Brief youtube preview video here. Some pics below and a few more pics in the DJTT forum thread above.

Posted Image

Posted Image

Posted Image

Double-buffered VBO vs updating parts of a VBO

09 May 2011 - 09:36 AM


I am currently working on my rendering code (using OpenGL 3.2, in case it matters) and I am wondering which of these approaches would be best (or if any better alternatives exist).

In both methods I store data about each object (position, colour etc) in a VBO, lets call it objectsVBO.

Method A, double-buffered VBO:

objectsVBO[0] and objectsVBO[1] are the two VBO buffers.
unsigned int buf = 0;
updatedObjects is an array of all objects, after having been updated by gameplay & physics code

every frame:
    sort updatedObjects back to front, remove objects not in view
    glBufferData to put updatedObjects into objectsVBO[buf]
    glDrawArrays(objectsVBO[buf], 0, num_objects_in_view);
    buf = 1 - buf;
EDIT: Looks like this method may be a better way to accomplish the above.

Method B, updating part of VBO:

objectsVBO is a VBO storing all objects
updatedObjects is an array of objects that have changed, after having been updated by gameplay & physics code

every frame:
    glBufferSubData to put updatedObjects into objectsVBO (possibly only objects that are on screen)
    indices = index of each object in objectsVBO that is on screen, sorted back to front
    glDrawElements(objectsVBO, indices);

Method A means only objects to be drawn ever need to be pushed to the gpu and since its double buffered, hopefully the next frame can be prepared before the first frame is fully rendered.

Method B means only objects which have actually changed need to be pushed to the gpu while unchanged objects are still there from the previous frame but I do not have double buffering.

Both methods only render whats on screen and render back to front as required. Static geometry will be kept in a separate VBO in both cases.

Any help or ideas appreciated as I don't have an awful lot of experience with 3D rendering yet.

A pragmatic language for the future - 5 years later

19 April 2011 - 04:54 PM

After having recently reread the entire A pragmatic language for the future thread, I am interested in hearing ApochPIQ's thoughts and experiences on the subject.
I am also interested in hearing if/how the views of the other participants of the thread have changed over the past five years, if at all, especially in light of the more recent additions to the programming world (for example, languages from the Emerging Languages Camp: Go, Mirah, CoffeeScript, Clojure, Frink, Factor - yes, I know some of these existed five years ago too).

For anybody who missed the discussion five years ago, ApochPIQ's language, Epoch, can be found here.

I am especially interested in the following areas:

  • General experience of developing a language aimed at being pragmatic.
  • What went right/wrong and what you would do differently were you to do it again?
  • Are there any language features that you have since learnt about that you have incorporated or plan to incorporate? If so, which ones and why?
  • How successful do you think the Epoch syntax is? It was a specific goal to make the syntax be somewhat C++-like, do you think this has been beneficial or a hindrance?
  • Some topics that were hotly debated in the original thread were macros, homoiconicity, variant types (algebraic data types) + pattern matching and exception handling. Are you happy with your decisions in these matters? Are there any alternatives you would consider if you were to do it again? (Eg, I like Go's defer/panic/recover mechanism as a replacement to exception handling)
  • Are you happy with the concurrency/parallelism/multicore support that Epoch has (or has planned), or do you think it should be much tighter coupled with the language and its syntax?
  • You were very clear that OOP should be built into the language, rather than as a library like Nathan Baum was arguing. Do you still feel this way and for what reasons? Do your feelings on OOP here apply to other major paradigms?
  • Regarding OOP, looking back, if you were to design Epoch again, would you choose class-based, prototype-based, duck-typed, message-oriented or something else as the object model?
  • The type system was a big deal in the original thread, especially types with constraints/validation. Unit/dimension types (like J.P.Morrisons "Smart Data", or Frinks units of measure) were also mentioned. How did this work out in practice, both from an implementation and from an end-user (programmer) point of view? To what extent did you take the "the compiler should be able to statically verify as much as possible, and runtime assert the rest" philosophy?

I know those questions are pretty specific to Epoch, but I would like to hear the answers to similar questions applied to the languages that other people have developed/are developing.
For example, I know that Telastyn was working on a programming language (Tangent).

I am also interested in other peoples views on those matters, on Epoch and on programming languages in general, regardless of if they have or are working on a language or not.