[Critical UX Issue] Unwanted Screenshot Annotations from Drawings in Vertical Chinese PDFs (Request


Dear Zotero Team,

I'm writing to report a significant usability problem in Zotero 7's built-in PDF reader that severely impacts workflows for East Asian language researchers.

Core Problem
Any graphical annotation (lines, rectangles, highlights, arrows, or freehand drawings) is automatically converted into a screenshot-based note. This behavior:

Creates extremely long, unreadable screenshot images when used in vertically typeset documents (e.g., Classical Chinese/Taiwanese Traditional Chinese PDFs), as the screenshot captures the entire column width.

Generates visually cluttered and bloated notes that cannot be searched, edited, or meaningfully organized.

Wastes storage space with redundant image files.

https://s3.amazonaws.com/zotero.org/images/forums/u10517527/51iv2fomtni10r7uqxd5.png

Why This Matters for East Asian PDFs
Vertical text is inherently wide (1 character height ≈ 10–20 horizontal characters).

A simple line/rectangle annotation forces a screenshot spanning 1000+ pixels in height but only 50px in width.

These "thermal receipt-like" screenshots dominate the notes panel, requiring excessive scrolling and offering zero textual value.

Requested Solution
Please add an option to DISABLE the automatic conversion of graphical drawings to screenshot annotations. Ideal implementations could include:

☑️ Preference setting: "Save graphical annotations as vector overlays only (do not create note items)".

⚙️ Tool-level control: Per-tool toggle (e.g., disable "Save to Notes" for Rectangle/Line/Highlight tools).

♻️ Alternative: Save drawings as text-based markups (e.g., [Rectangle at position (x1,y1)-(x2,y2)]) instead of images.

Why Text Annotations Aren't Enough
We deliberately avoid drawing tools, but accidents happen:

Misclicks on crowded toolbars

Unintentional drags while selecting text

Lag-induced cursor jumps
→ One slip instantly creates a useless screenshot note that must be manually hunted down and deleted.

Additional Context
Reproducible: 100% consistent in vertical/chinese-japanese-korean (CJK) PDFs.

Plugins aren't solutions: Workarounds (Better Notes, Zotero PDF Translate) help manage notes but don't prevent screenshot generation.

Prioritization plea: This makes Zotero's annotation tools nearly unusable for scholars working with pre-modern East Asian texts.

Thank you for considering this critical refinement. Supporting non-Latin script workflows demonstrates Zotero's commitment to global academia.
  • Any graphical annotation (lines, rectangles, highlights, arrows, or freehand drawings)
    Please don't use AI to generate bug reports. Lines, rectangles, and arrows aren't tools in Zotero, and highlights don't create images.

    Your screenshot shows the image annotation tool. The entire point is to create an image annotation. If you have some other example, you can provide it, but this is working as intended.
  • Hello,

    My apologies if my previous explanation was not sufficiently clear. I would like to elaborate on a specific challenge I am facing with PDF annotations.

    The issue arises from how PDF software handles annotations generated by drawing tools (e.g., lines, rectangles). When I use these tools to mark a section of text, the program automatically creates a corresponding annotation in the sidebar. This annotation typically includes an image preview—a screenshot—of the area I marked.

    For most users who read standard, horizontally-oriented texts, this functionality is perfectly acceptable. The screenshot generated is usually of a reasonable, proportional size.

    However, for scholars like myself who work extensively with vertically-oriented East Asian classical texts, this implementation presents a significant aesthetic and practical problem. When I mark a short phrase in a vertical column of text, the resulting screenshot preview in the annotation pane becomes excessively long and narrow. This makes the entire annotation list look cluttered, unappealing, and difficult to navigate, as I illustrated in the image I shared previously.

    Therefore, the ideal solution would be for the software to make a clear distinction between a "Marking" and an "Annotation".

    Here is what I mean by this distinction:

    A "Marking" should be defined as a purely visual, on-page indicator. When I perform an action like drawing a line or a rectangle, the shape should appear directly on the PDF content itself. Crucially, however, it should not generate a new, separate entry in the annotation sidebar. Its purpose is simply to be a visual cue on the page.

    An "Annotation", in contrast, would be an action explicitly intended to create a comment, a note, or an extractable piece of data. Actions like adding a "sticky note" or highlighting text with an associated comment would fall into this category, and they would logically appear in the annotation sidebar for management.

    In essence, I am proposing the ability to use drawing tools for simple, non-intrusive markup on the page itself, without having these actions populate and clutter the formal list of annotations.
  • Again, there are not multiple "drawing tools" in Zotero. There is one drawing tool, an ink annotation tool, and it's not what you showed in your original post. If you'd like to show a problem with the ink annotation tool, you can, but that's not what you showed. What you showed is the image annotation tool. The entire point is to create an image annotation. (The tooltip for that tool is "Select Area", which is somewhat leading, but what it creates are image annotations.) It would not be correct to use that tool to create a rectangle around horizontally oriented text either.

    If you're selecting text, you should use the highlight tool. If that's not working properly on vertical text, you should report that and link to an example PDF where it happens.
  • Text selection functionality is erratic or non-operational in documents featuring vertical Traditional Chinese typography. Attempts to highlight text result in erroneous behavior, such as the selection of a few non-contiguous characters, or a complete failure of the selection tool. Despite having processed these documents with Optical Character Recognition (OCR) software, the text selection issue has not been resolved. This is the example .
    https://s3.amazonaws.com/zotero.org/images/forums/u10517527/k84dwwt8fpz0l807lku2.png
  • Please link to the file or email it to support@zotero.org with a link to this thread so we can take a look.
  • Thank you!!!!!
  • All support happens here. The email address is just for sending files. You don't need to explain the problem again in the email, but you do need to send an actual PDF, or a link to one, not just another screenshot.
Sign In or Register to comment.