Because “equalize” is not a single universally defined operation in photo software. In the strict histogram sense, it can mean remapping tones toward a more uniform distribution, but editors may optimize for different goals instead.

Programs can equalize based on different metrics, such as:

• global pixel-value histogram
• local contrast in different image areas
• perceived brightness or perceived contrast
• constrained/limited equalization to avoid harsh artifacts

There are also multiple established methods, including adaptive histogram equalization, contrast-limited adaptive histogram equalization (CLAHE), dynamic histogram equalization, and others. These do not necessarily produce a flat final histogram.

So if one program gives a near-uniform distribution and another gives a curved result, that doesn’t automatically mean one is wrong—they may be using different equalization algorithms or adding tonal compression/expansion to preserve appearance and avoid losing edge contrast or creating unnatural results.

In short: “Equalize” is an umbrella term, not a guaranteed promise of a perfectly flat histogram.