ImageJ should be defended

I would like to draw your attention to a case where ImageJ is being blamed for its “bug” / “glitch” that causes unintentional removal of band in gel, causing the retraction of a paper.

I think this is not right.

7 Likes

Agreed! This is an example of (at best) carelessness and improper usage. They say they pressed backspace to delete an unwanted ROI and didn’t notice it that it cleared the area to 0. As in fact it is supposed to do (user guide).

They claim in the video that this is a “bug” in ImageJ!

Thank you @Kota for noticing this.

2 Likes

It seems @Wayne has responded to this and made a new daily build (see this RW reply)

Basically, it isn’t a bug (I’m pretty sure almost everyone here would agree with that), but he idiot-proofed it by disabling backspace once 1+ lanes are selected.

3 Likes

This is beyond comprehension. Since we are discussing glitches let me mention one glitch that makes the authors reasoning moot:
In that unexplainable video-evidence provided by the authors to retraction watch, the backspace key does two things: it clears the ROI AND dismisses the ROI (presumably, the reason why they did not notice the black box glitch).

The backspace key is the shortcut to the Edit>Clear command. The Clear command clears (i.e., fills with background color) area ROIs. There is nothing in the source code that will dismiss the ROI after it has been cleared. In other words: If you try to recapitulate the steps in the video, you will notice that pressing the “backspace” key, only clears the pixels inside the ROI, it does not dismiss the ROI. The ROI is only dismissed once you run “Edit>Select None” (shortcut: Shift+A) (or equivalent command), or click elsewhere in the image canvas.

So: either they have overridden the default behavior of the backspace shortcut key, or have a faulty keyboard in which the keys Shift, A and Backspace are fused…

@Wayne: please, please do not create an exception to the backspace shortcut command. Its functionality is well documented and this exception only degrades user experience as it introduces unexpected behavior. The status bar documents the action performed by the backspace key, and ImageJ will warn users when closing a modified image. I see no reason for the gel analyzer to disable key shortcuts.

Modifying the “backspace” key behavior is acknowledging that the glitch exists which is not the case. What happens when someone next takes a Western blot image (typically dark bands on a lighter background) and presses F, the shortcut for Edit>Fill: will ImageJ have an exception for that case too?

If someone is to blame here, I would blame the idiots who wrote the documentation. They obviously are doing a disservice to the community.

8 Likes