add overview figure

paper/paper.md

@@ -39,16 +39,17 @@ The **anywidget** project provides a specification and toolset for portable and
 reusable web-based widgets in interactive computing environments. First, it
 defines a standard for widget front-end code based on the web browser's native
 module system. Second, it provides tools to author, distribute, and execute
-these modules across web-based computing platforms. Since its release a year
-and a half ago, anywidget has steadily gained adoption. Nearly 70 new widgets
-have been created with or ported to anywidget and published to the Python
-Package Index (PyPI), along with numerous standalone scripts and notebooks.
-These tools cover general-purpose visualization libraries [@jscatter;
-@Heer2024-rr] as well as notebook integrations for applications in biology
-[@gos; @vitessce; @viv; @cev], mapping [@lonboard], astronomy [@ipyaladin], and
-education [@drawdata]. It has also been integrated into popular visualization
-libraries like Altair [@altair], enhancing interactivity in notebooks and
-deepening user engagement with visualizations and code.
+these modules across web-based computing platforms (\autoref{fig:overview}).
+Since its release a year and a half ago, anywidget has steadily gained
+adoption. Nearly 70 new widgets have been created with or ported to anywidget
+and published to the Python Package Index (PyPI), along with numerous
+standalone scripts and notebooks. These tools cover general-purpose
+visualization libraries [@jscatter; @Heer2024-rr] as well as notebook
+integrations for applications in biology [@gos; @vitessce; @viv; @cev], mapping
+[@lonboard], astronomy [@ipyaladin], and education [@drawdata]. It has also
+been integrated into popular visualization libraries like Altair [@altair],
+enhancing interactivity in notebooks and deepening user engagement with
+visualizations and code.
 
 # Statement of need
 
@@ -77,6 +78,16 @@ interactive visualizations across computational notebooks.
 
 # Overview
 
+![Overview of the anywidget project. (a) Anywidget Front-End Module (AFM) is a
+specification for widget front-end code based on ECMAScript (ES) modules
+(middle). Platform hosts load the module and call lifecycle methods with
+required interfaces (top). The `anywidget` Python package adpats JCPs into
+AFM-compatible hosts, allowing Jupyter Widgets to be authored and distributed
+with AFM. Other AFM authoring tools help developers write their widget
+front-end code with popular frameworks (bottom). (b) A project template CLI
+allows bootstrapping new anywidget projects that are ready to publish to PyPI.
+\label{fig:overview}](overview.png)
+
 ## A standard for widget front-end modules
 
 The **Anywidget Front-End Module (AFM)** is a specification for widget
@@ -115,7 +126,7 @@ module, either inline or in a separate file, without a build process
 UIs, developers can introduce a build step targeting AFM to utilize advanced
 tools (\autoref{fig:afm-and-anywidget}a, bottom).
 
-![AFM and anywidget Python package overview. (a) AFM can be authored in
+![Authoring a custom Jupyter Widget with anywidget. (a) AFM can be authored in
 web-standard ECMAScript (top) or with a front-end framework using a bridge
 (bottom). (b) The anywidget Python package allows authoring custom Jupyter
 Widgets with AFM (top), usable across various JCPs (bottom).