Clark Quinn's Learnings about Learning
(The Official Quinnovation blog)

15 August 2008

Game Development Tools

Clark @ 8:09 AM

The last topics in our 2 day game design workshop for the Guild (great group of attendees, great experience) were evaluation, production, and organizational issues.  On the production issue, the perennial topic of tools came up.  In thinking about it, I realized that we needed a map, so I started coming up with one (a diagram, of course :) ).  I ran it past Jeff (Johannigman, my co-conspirator on the workshop) in our taxi to the airport, to his general approval.


The two dimensions are complexity of the scenario (only covering branching and model-driven), and the power (e.g. complexity) of the tool.  It’s a pretty linear map, and realize that small distances aren’t significant (so the clusters are roughly equivalent).

The impossible dream is that tool that everyone wants that makes it easy to develop model-driven interactions.  Sorry, I’m convinced it can’t exist, because to be flexible enough to cover all the different models that we’d want to represent, it’s got to be so general as to be essentially just a programming language.  QED (Quinn Ephemeral Decision).

This is a first stab, so feedback welcome.  If desired, I can create it in Gliffy and we can collaboratively develop it (though my first effort with that was underwhelming in participation…).  Thoughts?


  1. Yeah I came up a similar diagram a while back:
    See also

    It would be nice to have something visual and as simple to use as supercard/hypercard, yet powerful enough to make modern applications (applets, desktop apps, games, web applications…).
    I actually have experience now working on parsers and compilers so it’s not purely a realm of fantasy (just vaporware :).

    But you might also look into other tools like labview that have a visual building blocks syntax.

    Comment by Doug Holton — 15 August 2008 @ 5:52 PM

  2. Great map!

    The “scenario complexity” label could be added to the y axis with “simple branching” at the bottom and “model driven” at the top, just as it is now.

    The x axis could be labeled “user experience” or “interactivity level” with “simple”, “complex” and “AI” as the scale values.

    We might be loosing the fact that creating complex scenarios with artificial intelligence becomes highly technical but this might be a point to make elsewhere.

    If you ever put it on Gliffy, let me know. Thanks for the Gliffy pointer, I didn’t know about this site.

    B.T.W., I do agree with your impossible dream zone.

    Thanks for the post,

    Comment by Raymond Bissonnette — 15 August 2008 @ 6:03 PM

  3. This is a great diagram.

    In the arts what has developed is visual programming languages such as Max (http://www.cycling74.com/products/max5) These are great because they allow visual artist’s and musician to build complex models and systems without having to code, in a traditional way.

    Over the last year or so I’ve been working on a Flash based engines for simulations that can be modified just by changing XML files. What has been interesting, is that to make the system usable for most educators I’ve had to move towards a branching system. What I’ve decided to do is keep the engine/player open source so that models and behaviors can be added.

    Because most of the industry is involved in custom development do think this stops the development of toolset to make development easier ?

    Comment by Robin Petterd — 17 August 2008 @ 2:15 PM

  4. […] will be near impossible for one tool to satisfy all requirements. Clark Quinn makes this point very well with reference to model driven ILS – these are inherently so different […]

    Pingback by The coming era of Rapid Sims « Thinking Worlds — 23 August 2008 @ 5:34 AM

  5. Thanks for posting – we don’t get the chance so often to attend your workshops this side of the pond, so this content is appreciated.

    I agree that one tool will not cater for all needs on ILS development. As the sector matures we are seeing distinctions between different types of ILS that are appropriate for different types of learning objectives, audience and business requirements. The taxonomy of ILS needs more specification but I can see a time where different tools provide rapid development for specific different types of ILS. They become specialised tools that simplify development in these narrower areas.

    Comment by Chris Brannigan — 24 August 2008 @ 10:01 AM

  6. Chris, you’re right, I should figure out how to get over there and offer a workshop or so. Hmmm…

    As to tools, the taxonomies I’ve seen as yet are pragmatic, not principled, which doesn’t give us much leverage. But it certainly is possible that there will be some framework that provides conceptual clarity and the ability to tailor tools to categories. Fingers crossed.

    Comment by Clark — 25 August 2008 @ 4:16 AM

  7. Hi All,

    Just a quick comment towards the impossible dream. It can happen and it is happening. A real tool for real development without the need of programmers.

    All RD here, so not much I can say other than it is the future of development.


    Comment by Prof. — 1 October 2008 @ 1:33 PM

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress