Git commit messages can be important for browsing the commit history of a repo. This article gives seven basic rules for creating good commit messages:

  1. Separate subject from body with a blank line
  2. Limit the subject line to 50 characters
  3. Capitalize the subject line
  4. Do not end the subject line with a period
  5. Use the imperative mood in the subject line
  6. Wrap the body at 72 characters
  7. Use the body to explain what and why vs. how

There are explanations for why each of these rules is a good idea.

The Advent of Code for 2024 has started! Each day there are two coding problems — solve the first one to unlock the second one. I took the opportunity to try out the new IDL notebooks in VS Code that I just learned about yesterday. It was a nice environment to work through a problem.

IDL 9.0 was released recently[1]. Some notable new features:

  • Apple Silicon Mac support — most of my IDL use is from a terminal window on my Mac to a Linux, but occasionally I do some IDL development directly on my Mac. The improved speed is probably not a factor for me, but it is good to know NV5 is still going to support MacOS.
  • HttpRequest class — I have some routines that I use to make requests, but this should be a nice addition to the standard library.
  • IDL Extension for VS Code — I wrote about this already, but I missed the IDL native notebooks (not Jupyter notebooks). This looks very interesting, enough to install VSCode for a while to test them out.

There are a few more features, as well as updates and library updates. Details in What’s New in IDL 9.0.

  1. I learned about it today the “Redefining Image Analysis with ENVI 6.0, IDL 9.0, and the ENVI Ecosystem” webinar, but maybe it’s been out for awhile? ??

APOD had an atypical image last January (I haven’t been writing a lot recently!) of a periodic table for the elements that showed where each element is normally created. The description lists some examples:

The hydrogen in your body, present in every molecule of water, came from the Big Bang. There are no other appreciable sources of hydrogen in the universe. The carbon in your body was made by nuclear fusion in the interior of stars, as was the oxygen. Much of the iron in your body was made during supernovas of stars that occurred long ago and far away. The gold in your jewelry was likely made from neutron stars during collisions that may have been visible as short-duration gamma-ray bursts or gravitational wave events. Elements like phosphorus and copper are present in our bodies in only small amounts but are essential to the functioning of all known life. The featured periodic table is color coded to indicate humanity’s best guess as to the nuclear origin of all known elements. The sites of nuclear creation of some elements, such as copper, are not really well known and are continuing topics of observational and computational research.


Jason Kurth took his custom hydrogen alpha solar telescope to the desert in Utah to film the recent annular eclipse in 8k. The result is on YouTube. Amazing! There are also some other great images on his site and Instagram.

via kottke

I have not heard the term “infoposter” until I read Connie Malamed’s post describing them and contrasting them with infographics.

The infoposter, a category name that seems to fit this type of graphic, is not yet defined in Wikipedia. Yet it is probably safe to say that the infoposter is a graphic that conveys multiple segments of information typically using words and numbers to represent quantitative data. Infoposters generally use iconic-type graphic elements for visual design appeal and are typically vertical in orientation, similar to a wall poster.

By the way, this was a reference in Building Science Graphics by Jen Christiansen. I’ll have more to say about it when I finish reading it.

I often need to combine several vectors into an array of structures where the i-th element of the array of structures corresponds to the i-th elements of each of the vectors.

For example, if I have three vectors of differing data types:

IDL> x = indgen(3) + 1
IDL> names = ['Mike', 'George', 'Bill']
IDL> values = 100.0 * randomu(seed, 3)

And I want to combine them into a single array of structures for printing, I can do:

IDL> print, mg_zip(x, names, values), format='(%"%02d. %-15s: %6.2f")'
01. Mike : 3.22
02. George : 43.89
03. Bill : 97.14

This part of a growing collection of routines I have to manipulate tabular data, centered around an mg_table class. The code for mg_zip is in src/tables directory of the mglib repo.

More to come on tables…

There is an official VSCode extension for IDL! NV5 Geospatial mentions it in a blog post:

TLDR: We have an official extension for Visual Studio Code called IDL for VSCode! It is open source on GitHub, you can see some of the features we have planned, and there’s a place to ask questions and provide feedback. See below for a story about our journey to get here, our inspiration for the user experience, and our goals for community engagement.

The extension seems fairly useful. It’s more than just a syntax highlighter. It can give linting-type information: finding syntax errors, give recommendations (for example, you are missing a compile_opt statement), warnings like unused variables, etc. And, of course, everyone’s favorite editor feature — auto-completion.

There is a, so far, quiet, GitHub Discussion for discussing the extension.

The Datawrapper blog has been publishing some data visualization posts. This article gives a an overview of ways to make your legend colors more accessible to viewers:

You designed a useful (and fun!) color key for your data visualization to explain to your readers what your colors mean? Great! Now you’ll want to make sure people don’t forget those associations between colors and categories. This article explores a few ways to do so:

  1. Keep your color key always close by
  2. Reuse your colors in annotations
  3. Make your color key interactive
  4. Reuse your colors in tooltips
  5. Reuse your colors in text around your visualization

Better use of color keys

This great, one-purpose website displays an interactive map of the path of the 2023 eclipse over North America.

Eclipse path near Dallas, TX

This allows you to find a place that you can get to in the path of totality on April 8, 2024. I’ll be watching near Dallas.

Pardon my snobbery, but if you see only the partial phase of this or any other eclipse, you really did not see the eclipse!

I agree with that. If you’ve only seen a partial eclipse, don’t think a total eclipse will be anything like that.

older posts »