With html documents, Quicksilver can either scan all the text lines of
the document (i.e. treat it like any other plain text document), or
all the links in the document.

Right now you seem to be scanning all the text lines of the document,
which is why you are seeing the raw html code. This is the expected
behaviour in such circumstances. I doubt that scanning the html links
would be of much use to you, since that is not the information you are
interested in.

To make your definitions appear cleanly in the QS interface, you will
need to store the info in a plain text document somewhere. If the data
on that webpage is fairly static, you might just copy and paste it
into a .txt file, and add that .txt file to your catalogue. Assuming
these definitions are only rarely updated, and that it is done
manually by you personally, it wouldn't be much work to copy over any
changes from the web page to your plain text documents at the same
time.

If it is important to you that the data QS is accessing is at all
times the very same data being displayed on the webpage, and that data
is changing often (such that manual updates are not feasible), you
could use any of various html tricks to save the data in  a plain text
file and have the page load it on demand. A simple iframe pointing to
the .txt file could do this quickly and easily, but not very prettily.
More useful, perhaps, would be some CSS styling or something like a
php script that reads the .txt file and presents the information in a
nicer way.

On Mar 13, 6:44 pm, Pete Siemsen <[email protected]> wrote:
> I maintain some webpages of definitions for the people at my site, for
> example
>
> http://www.cisl.ucar.edu/nets/topics/s.shtml
>
> I'd like to use QuickSilver to look up the terms in the files.  Even
> better would be if I could highlight a string in Terminal, Safari,
> Mail etc. and hit a hotkey to see the definition, perhaps by opening a
> Safari window onto the webpage.
>
> I implemented functionality like this with a Mac OS X Service that
> executes a Ruby script that throws up a dialog box.  It seems that
> QuickSilver could do a similar job better. I tried adding the HTML
> file to my QuickSilver catalogue, and specified the content as "text
> line", and it kindaworks - QuickSilver displays the ugly raw HTML
> containing the search term from the web page.
>
> Any suggestions welcome.  I could make minor changes to the web pages
> if that would help, but they serve a useful function in their present
> form.  I'm a QuickSilver newbie, so the answer might simply be "get
> plugin XYZ".

Reply via email to