(This was a failed experiment.)

Since nobody's quite sure what the configuration and scripting interface should look like or needs to do, let's collect some use cases.

Feel free to add use cases, and expand on ones that already exist. The idea is to make them specific, so that they do not suggest a solution on their own.

Many of these things are already supported, or could be with the appropriate scripts. The goal here is not to provide solutions to these scenarios, but to determine what people are doing, what they would like to do, and where scenarios share similarities.

“I” here is not meant to be a particular person; the person who added a use case may have no interest in it themselves. “I” is just easier to type than “some guy”.

Not all use cases can or should be supported by uzbl (but we might as well collect them anyways).

I would like to set the User-Agent header. It would be nice to be able to put the version of webkit, my OS, etc. in it.

I would like to be able to set the value of arbitrary headers (for example, Accept-Language).

When a page has finished downloading, I would like an external script to get the contents of the page, process it, and send it back to uzbl for display.

When uzbl fetches a resource that has a video/* Content-Type header, I would like mplayer to launch with the URL.

(This is not specific to video/* or mplayer, I just want to make this scenario distinct from download handling)

When uzbl fetches a resource that has a Content-Type that isn't supported, I would like it to be saved to my local filesystem.

I would like an external script to know when a download has completed (for notifications, etc.).

(The solution to this is probably the same as the solution to “Downloading unknown Content-Types” and maybe “External handler for specific Content-Types”. But maybe not, so let's keep distinct scenarios.)

My computer crashes a lot. I would like to maintain a list of the URLs I have open in uzbl so that I can restore it when I need to.

I would like to keep a log of some or all of the commands that I execute.

I would like to bind actions to items in the right-click menu.

I would like to be able to define keybindings and other actions (right-click, etc.) based on the current URL.

I think it's a good idea to have several daemons: to handle cookies, remove ads(more friendly then privoxy, with interface like adblock), on-fly transform content(greasmonkey like) etc.

I would like to be able to pipe all text from within specific html tag to another application or file. For example I am browsing a long config file on line. I could right click on the text and select “send to …”