What functions should be moved to basic MATLAB?

What are some of the functions that you think should be moved from a toolbox into basic MATLAB?

The function should be one that you have found is used in a variety of situations including outside of purposes related to the toolbox it is found in.

Keep in mind that if a function is a major feature of its respective toolbox, then moving it into basic MATLAB is not likely to happen.

This is not a discussion of which toolboxes should be merged into basic MATLAB... that is a very different discussion!!

Answer 1

[Image Analyst]

My choices (so far):

1. medfilt1, a moving 1-D median filter, in the Signal Processing Toolbox.
2. medfilt2, a moving 2-D median filter, in the Image Processing Toolbox
3. insertText, a function to burn text into an image, in the Computer Vision Toolbox
4. pdist2, a function to calculate distances from every point to every other point, in the Statistics and Machine Learning Toolbox.

[Walter Roberson]

I know that I have suggested more than once that MATLAB would benefit from a "render to buffer" interface -- but in the meantime, moving insertText (and insertShape) into main MATLAB would be a good step.

You are right, pdist2() gets used in many contexts.

Answer 2

[Blue]

Hi Walter,

Im tempted to say shaperead from the Mapping Toolbox. Yes, I know shaperead is a big part of the Mapping Toolbox but shapefiles are used in a wide variety of fields and its really hard to live without that function. Otherwise I would nominate boxplot from the Statistic Toolbox. Its a small function that could easily be in basic Matlab.

[Walter Roberson]

Interesting choices, and very much in the spirit of the question!

[Paul]

FWIW, base Matlab has

doc boxchart

as of 2020a. I don't think it has quite all the bells and whistles of boxplot(), but it might be close enough for what you need.

Answer 3

[Walter Roberson]

I would nominate buffer() from the Signal Processing Toolbox https://www.mathworks.com/help/signal/ref/buffer.html .

buffer() takes a vector and a length, and rearranges the vector as columns that are that long; if the last column is short it adds padding.

buffer() does get used a lot in communications contexts, but it also has uses anywhere that you do array processing, including image processing, and including feature creation for neural networks or Statistics toolbox.

Answer 4

[Walter Roberson]

wrapToPi() has obvious uses outside the Mapping Toolbox https://www.mathworks.com/help/map/ref/wraptopi.html