Comments for Immersive Intellegence Colleagues

Comment on SQL versus Hadoop versus SciDB by Dan Graham

Dan Graham — Tue, 11 Oct 2011 17:31:12 +0000

A substantial amount of scientific analytic processing is done every day using relational databases. Nonetheless, there is a clear need for the scientific community to begin collecting and managing data in ways specific to scientific workloads. One-and-done research has to give way to ongoing research which means capturing data and managing it. Repeatable job execution and incremental algorithm improvement is needed.

RDBMS vs Hadoop vs SciDB debates are a waste of time. While Hadoop and RDBMS can be made to perform any algorithm, they are often not the right tool.

SciDB is the right direction. Open source is a good fit for experimentation and the way researchers think. Hopefully, the SciDB vision of reliable data access for scientific algorithms will motivate a wider community to contribute to the code base.

Having said that, the SciDB organization must actively solicit knowledge and contributions from the RDBMS community who long ago solved the complexities of massively parallel clusters. It is no trivial feat to get dozens or thousands of CPUs to behave as a single application. Just having a reference “to be” architecture for clusters would help the SciDB open source community immensely. Hadoop may offer some primitive starting concepts but is 10-20 years behind the RDBMS community.

Good eye Richard!

Comment on Stanford AI Course in the Global Village by David Fliesen

David Fliesen — Mon, 10 Oct 2011 16:01:02 +0000

This looks to be an exciting course Richard. Can’t wait to see how things go and hope to learn a lot. You can also find the textbook (although it’s not required for taking the course) on Kindle for a lower price.

Comment on Playing with Large Meshes of Mammoth Cave by Don’t Be Afraid of the Dark « The Accidental Cootchie Mama

Don’t Be Afraid of the Dark « The Accidental Cootchie Mama — Tue, 13 Sep 2011 10:02:45 +0000

[...] skipped through the historic cave tour like a pro. Nothing at all to fear. Because we enjoyed the heck out of the first tour, we decided [...]

Comment on Gamification of Data Analysis by Anders Gronstedt

Anders Gronstedt — Mon, 05 Sep 2011 00:08:38 +0000

Great book review Dick, Mihaly Csikszentmihalyi’s chart illustrating the balance between boredom and anxiety is a great tool to show that there is a trade off between making games simpler to use, which is what clients and new users always push, and less interesting to play in the long run, which is a requirement that is rarely voiced in the development of educational games.

Comment on Gamification of Data Analysis by richardh

richardh — Sat, 03 Sep 2011 01:24:25 +0000

Several excellent points, Andy, such as: “system that understood the level of motivation and ability of each person”.
Could you give an example of a “heuristic persuasive gaming engine” that has or could of this ability, especially as it relates to data analysis?

Comment on Gamification of Data Analysis by Andy Hewitt

Andy Hewitt — Thu, 01 Sep 2011 18:45:40 +0000

The problem that is outlined by Richard is not as simple as he has outlined. The real crux of the issue is actually about interpretation (read ability), relevance (read motivation) and in turn the effort (vs incentivisation) that leads to hit and miss results.

I believe that ‘flow zones’ where you can obtain points, levels, badging, on boarding, quests etc are relevant but fall short because they assume that everyone has the same ability, motivation, and will respond consistently to the incentivisation. We know from our indepth experience and analysis here at Global Dawn that is simply not the case.

My main issue is that the suggested solution really falls short for me as does most gamification. What it recommends is layering techniques over the task and solving the engagement issue for some but not for all in a consistent meaningful way. It deals with users understanding and hopes the ‘fun’ will influence the motivation of the task and in turn results.

I think a more relevant and successful alternative would be to create a
system that understood the level of motivation and ability of each person at each point of interacting with the behaviours required to be performed against the data.

This would mean you could tailor the facilitation of support, messages and rewards at each point to ensure that the activities are completed, by all based on them as individuals. You dress/layer this alternative up in fun and playful ways but the core is more responsive and full proof. My vision for true engagement is a heuristic persuasive gaming engine which tailors the way people experience information and that had a direct impact of what was delivered to them via a tailored dashboard based on relevance.

Comment on Playing with Large Meshes of Mammoth Cave by Creating Unity3D Scene to Roam Mammoth Cave | Immersive Intellegence Colleagues

Thu, 01 Sep 2011 01:21:34 +0000

[...] a previous blog, I discussed the Mammoth Cave project and the need for a lite-weight scene in Unity3D as a [...]