Skip to content

Commit a88202c

Browse files
committed
doc-range
1 parent d8e8f7b commit a88202c

File tree

3 files changed

+28
-3
lines changed

3 files changed

+28
-3
lines changed

.idea/workspace.xml

Lines changed: 12 additions & 3 deletions
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

site/faq.md

Lines changed: 16 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -46,3 +46,19 @@ only perform upward traversal of the AST of the first gaze and record the hierar
4646
as `same`. This approach significantly reduces storage space. In a previous debugging study, we set the eye-tracking
4747
device's sample frequency to 60Hz, and during the 20-minute experiment, the eye-tracking data amounted to only about
4848
40MB.
49+
50+
#### Q7. Which eye gazes can be analyzed, and can gazes on the UI be understood?
51+
52+
In CodeGRITS, we only analyze gazes within the code editor. For example, in the following figure, only gazes within the
53+
red rectangle are mapped to the source code tokens and performed the upward traversal of the AST. For gazes outside the
54+
code editor, such as those on the file explorer, menubar, tool window, console, etc., we only record their raw
55+
information and add a remark `Fail | Out of Text Editor`. CodeGRITS cannot understand the semantics of gazes on the UI.
56+
57+
<div style="text-align: center;">
58+
<img src="../static/range.png" style="max-width: 100%; width:900px; height: auto;"><br><br>
59+
</div>
60+
61+
However, if certain UI elements within the editor are triggered, such as the list for auto-completion or the inline
62+
definition display, CodeGRITS mistakenly interprets the gaze as being directed at the code content below the UI (when it
63+
should actually be on the UI). We have not yet found an appropriate method to resolve this issue, which may lead to
64+
inaccuracies in some gaze analyses.

site/static/range.png

140 KB
Loading

0 commit comments

Comments
 (0)