I can understand why it would not be clear which one to place your content on, given their similarities. Both of them run in separate processes and have many similar configurations. The key difference between BrowserWindow and the web browser is that BrowserWindow is a window on the platform, and web browsing is an element on the web page. It may be a little obvious, superficial difference, but most of their differences and customs stem from it.
Most of the reasons why web browsing exists is to include untrusted content in your application. If you read web browsing examples, many of them indicate that BrowserWindow has full access to the Node API by default. Hosting untrusted content inside it is providing this substantial access to your system and presents a security problem. However, web browsing does not include Node integration, so it protects your application and platform from hosted content.
However, this difference is a small red herring, since Node integration can be disabled in BrowserWindow and can be enabled in the webview element. That is, you should be able to safely place untrusted content in the Windows browser, taking away access to Node and trusting the trusted content of the host in the web view and giving it access to Node.
The key to web browsing is that it allows you to embed untrusted content on a web page / view in your application. If on the same view / page you would like to have some content that is trusted with full access to the Node API, as well as some untrusted content and limited access to the Node API, then this can only be done with the webview element. This is a way of splitting and blocking a portion of a webpage hosted by the Windows browser, allowing the rest of it to open.
Besides introducing untrustworthy content, the only case I can think of using webviews through BrowserWindows is to open and view several separate processes in one window. An application can choose to create 10 different windows for 10 different processes and have a platform descriptor layout, focus, etc., Or it can open 1 window with 10 web views for 10 different processes and handle the layout, focus, etc. Inside this window.
(Edit) To proceed to editing the question:
In both cases, I suggest using webview.
In the first browser scenario, you mentioned “tabs”. There is no simple and cross-platform method that I know for creating tabbed applications using multiple browsers, since windows are created by the native OS. However, you could achieve this by creating a tabbed application on one web page, on each tab containing a web view. In this case, you want Node integration to be disabled in webview (it should be the default), since downloading content from the Internet is usually untrusted.
The second scenario, an editor with rendered HTML, is not so clear. You can use web browsing, iframe, or render content directly in a div. Rendering directly to a div may be the best option for something like Markdown or small HTML snippets if you don't need custom css or want to run JavaScript. Otherwise, it makes sense to use webview or iframe. The difference is that the webview is launched in a separate process and can have integration with Node integration or bending, while the iframe is loaded in the same process as BrowserWindow, and, I think, blocked security. Regardless of how to view side by side without another window, you need to use an HTML element like webview, not BrowserWindow.