My idea is that using custom scrollbars, if remain in the scope of aesthetics, i.e. look like just changing the skin of the default scrollbar, is not a bad idea. Users immediately understand other custom parts of a web page like custom buttons, custom checkboxes, etc. Do you have problem understanding the clickable attribute of vote buttons in stackexchange sites? Or do you have problem interacting with custom input text boxes for tagging questions in stackexchange sites, or Yahoo mail? Scrollbars also can be fast-recognizable by end users, as long as they don't change the layout, or the expected functionality of the browser, for example scrolling in the reverse direction, or creating a circular scrollbar. You're good to go as long as you only change the skin.
However, there are some problems with custom scrollbars that you should consider:
- They usually need a considerable amount of time to become cross-browser
- They might not be forward-compatible, i.e. by releasing new versions of browsers, they may break functioning.
- You have to manually implement scrolling behavior. For this item, consider HTML buttons. When you add
disabled='disabled' attribute to them, browser simply won't accept clicks or enters anymore on them. However, when you use a
<div> element to mimic the behavior of a button, then you should implement disabled behavior on your own. Custom scrollbars also should implement scrolling functionality, which is already implemented by browser.
In spite of the things we say, I recommend you start a hobby project (not production) to experience using and implementing custom scrollbars. See this page for good examples.