On 15 April 2011 03:48, Rob McBroom <[email protected]> wrote:
> On Apr 14, 2011, at 5:36 PM, Eric Doughty-Papassideris wrote:
>
>> This could have the further advantage of exploring the current web
>> page with another set of plugins. Say I'm looking at a specific photo
>> in flikr, it would be pretty cool for QS to go through the HTML get
>> it's hand on the "content" of that web page instead of just it's URL.
>
> I’ll have more to say on this thread when I have time, but just FYI, with a 
> URL in the first pane, you can right-arrow into it (or use /) to get a list 
> of links on the page. It’s almost impossible to think of something that Alcor 
> not only thought of first, but already implemented.

I am aware of that, but since there's no way to set the parser used
per url (that I am aware of anyway), it is impossible to retreive
smarter things than url+title or the links therein, and you end up
with links to menus and ads. I'd like to explore ways of interpreting
the content within the browser, if a plugin is written for browser
mediation it should be possible to trust "html+url is available",
which makes it possible to get only the search results for a google
page, or to make a google maps page "image compatible", or an event
from what you're looking at your calendar. I'm just wondering whether
plugins can benefit from deeper understanding of the browser's
content, which would need some abstraction. The reverse might be
possible for actions like injecting javascript.

Reply via email to