April 1, 2018

What is MS Fool for $1000, Alex?

Want more Jeopardy! answers in question form? Check out the J! archive.

There is a recurring insider reference among Very Serious Computer Users regarding using Microsoft products to perform sophisticated computational tasks [1]. While most people tend to think of these programs as not computationally sophisticated, programs such as Excel [2], PowerPoint [3], and even MS Paint can do some pretty powerful computing. One such example are 3-D Models enabled by object/shape manipulation in PowerPoint and Paint.

Around every April 1, Carnegie Mellon University students participate in an annual tongue-in-cheek conference (sponsored by the satirical Association for Computational Heresy) called SIGBOVIK (Special Interest Group on Harry Questionable Bovik). Not sure what where Harry Q. Bovik got his credentials, but if you enjoy the Ig Nobel Awards, this should be right up your alley.

SIGBOVIK often features highly interesting quasi-research [4] projects based on the aforementioned Microsoft program suite. But there are other groups creating programs and computational artifacts because they can. Here are a few examples of this collected from around the web:

1) Using MS Paint to create high-quality works of art, courtesy of Hal Lasko and "The Pixel Painter".

2) Tom Wildenhain's SIGBOVIK 2017 lecture on Turing-complete Power Point.

The (non-) Turing Completeness of PowerPoint. One of many computational artifacts that are Turing incomplete. COURTESY: Tom Wildenhain YouTube channel and HackerNoon.

3) Try out this version of Doom programmed in Excel, courtesy of Gamasutra. The game program runs on a series of equations that requires VBA to run. There are several files you need to download, and the blogpost goes through the full set of challenges.

Rasterized Graphics with Underlying Linear Equations [5]. Examples from the Excel Art Series. ARTIST: Oleksiy Say

4) The final example returns us to SIGBOVIK (2016), where David Fouhey and Daniel Maturana bring us ExcelNet. This is an architecture for Deep Learning in Excel, and has gone through the paces of the Kaggle MNIST competition. Another example features Blake West and his implementation of a deep learning architecture in Google Sheets.

[1] This is similar to advertising membership in the cult of LaTeX, touched upon in this discussion of LaTeX fetishes.

[2] the first spreadsheet (Autoplan) was developed in the form of a scripting language, which is a more general version of more formal programming languages (e.g. Java versus Javascript).

[3] presentation software can be pretty diverse. But is any of it Turing Complete (TC)? If you work out your design process using Wang Dominoes, then perhaps TC-ness can be realized.

Example of a Wang tile.

[4] definition of quasi-research: research projects that do not produce useful results directly, but often as a side-effects of the research process itself.

[5] by combining the world of image rasterization with interlinked linear equations, there are some exciting (but conceptually and technlogically undeveloped) opportunities in this space.

No comments:

Post a Comment