michaelgalloy.com

In this recent IT Conversations interview, Jon Udell discusses topics from Greg Wilson’s recent talk “High-Performance Computing Considered Dangerous”. Wilson places a new emphasis on testing and usability over more power and speed.

The interview focused on the difficulties of learning techniques in High-Performance Computing and, in general, about the split between scientific computing and “commercial” software development. Tools have to improve because there is no more time in the science curriculum to teach science graduate students more about software engineering practices. In particular, I thought the comments about version control aiding in the reproducibility of computations were interesting (version control as a “software lab notebook”). He also mentions that tools like “Matlab, IDL, and Mathematica” are scoffed at by the HPC gurus, but are the tools that many scientists are using to solve their problems (he put Excel in a category of popularity above these). Our focus should be adding easy to use HPC techniques to these platforms.

David Fanning made some fixes to the Brewer color tables addressing the issues reported in the comments of the last update by Philippe Le Sage. The update includes the following:

1. Fixing values in sequential and diverging color tables.
2. Fixed up names of diverging color tables
3. I also realized that it was unnecessary to include multiple versions of each of qualitative color table because the one with the most colors is a superset of the other versions.

See the last update for routines to access these color tables.

Thanks David and Philippe!

I have always been interested in Processing, the open source project that, according to its website,

…is an open source programming language and environment for people who want to program images, animation, and interactions. It is used by students, artists, designers, researchers, and hobbyists for learning, prototyping, and production. It is created to teach fundamentals of computer programming within a visual context and to serve as a software sketchbook and professional production tool. Processing is developed by artists and designers as an alternative to proprietary software tools in the same domain.

I read Ben Fry’s blog and his book is on my Amazon.com wishlist.

Now, John Resig has created a Javascript API for Processing. Expect a lot of more use of Processing on the web.

I modified one of the demos to use some IDL example data.

Information Design Patterns is a site created from a master’s thesis that catalogs many common chart types. Many types are listed along with some useful ways to browse through them. Each chart type consists of description, a sample layout, a real world example, and a (mostly not active yet) implementation sections.

There’s an XQuartz update. Remember to keep the package around to re-install after doing updates from Software Update.

I’m in Seattle for the rest of this week and part of next week. See you next Wednesday!

I just received a copy of IDL Primer by Ronn Kling. (Full disclousre: I received a free copy because Ronn used my Periodic Table of IDL Operators on the back cover.) It’s a pocket sized “quick reference” that provides a quick introduction to the main functional areas of IDL: the IDL Environment (DE is covered, get Kling’s Navigating the Workbench for coverage of the Workbench), syntax, array operations, program flow, 2D and 3D graphics, image processing, object graphics, file I/O, mapping, and animation. I have found it to be a quite handy reference, particularly the list of “Important Routines by Function” at the back (similar to the Functional List of IDL Routines in the online help).

The deadline for the NSF visualization contest is coming up on May 31. Winners will be published later this year on September 26.

Check out last years winners (and back to 2003) to see the level of competition. Of course, the descriptions of the winning entries concentrate on the science on display, but I wish there were also some information about the tools and process used to create the visualization.

XQuartz has an update for X11 on Mac. Remember to re-install this package after further updates to Mac OS X.

UPDATE: it’s been updated to 2.2.0.1 already.

I have finished the line-integral convolution (LIC) implementation as a DLM and cleaned it up a bit (see other articles for background). More work is needed, but it can give useful results now. The movie at the right took about six minutes to generate (300 calls with 300 by 300 images).

The API for the call is pretty straightforward. There are two required positional parameters u and v which are m by n float arrays (they must be float arrays right now). There is one optional keyword texture which can be set to a byte array the same size as u and v. It is useful to set texture to the same array for frames of a movie; there will be a lot of jitter if you don't.

Installation is slightly more complicated than for IDL code. Grab the zip file, unzip it, and place the contents in the IDL path and the IDL DLM path. Start IDL and type:

You should be ready to go as long as your !make_dll system variable is set correctly (which is probably the case on normal Unix platforms, but only if you have installed the Developer Tools on Mac OS X or Visual Studio on Windows). I would be interested in people's experiences trying to install this since I haven't distributed a DLM on this site before.

I am going to package some visualization routines together with the vis prefix, but I'm waiting until I have a critical mass of useful routines before distributing the whole thing.

UPDATE: by the way, if you use vis_lic and can share your results, I would like to see any images made with it.