[ANN] GDeskRSS: RSS Feeds from the Google Desktop Searche tool

[James Britt]

Google has made available an application that runs on a user’s PC,
searching and indexing files much as the Google Web site does for the
Internet. The application uses a Web browser for its user interface,
which means that search results are also available to applications that
know how to make HTTP requests.

The Google search app appears intended for Internet Explorer 5 or
higher, running on Windows. However, it seems to work with Firefox, too.
But still, only for Windows.

GDeskRSS takes a set of keywords, runs a query, and parses the resulting
HTML. The extracted data are then reformatted as an RSS 0.9 feed.

There are a few ways to run the code. You can call the main file,
gdesk_rss.rb, from the command line:

ruby gdesk_rss.rb <your keywords here>

The resulting RSS feed will spill across the command window, of dubious
value, but you can at least see if that much works.

You can also run the code via WEBrick, using gdesk_brick.rb

ruby gdesk_brick.rb [port_num]

If you omit the port number, the WEBrick instance runs on some default
port other than 80 (likely 8090, but check the code in gdesk_brick.rb).

Another way to run the program is to use Catapult. Catapult is simple
Ruby WEBrick app that runs code by dynamically loading classes and
instantiating objects as specified via a URL. I wrote it as a way to
easily run local Web services on my PC without having to create a new
WEBrick instance for each one, or rely on Apache.

See http://www.jamesbritt.com/code/catapult/ for details

GDeskRSS is mostly proof-of-concept, written to see 1) how hard it would
be (not very), and 2) how useful RSS feeds tracking local file content
might be (remains to be seen).

Hopefully others will find it useful and/or entertaining

See http://www.jamesbritt.com/code/gdeskSearchRSS/ for details and
downloading version 0.1.0.



James Britt

 

[Francis Hwang]

GDeskRSS is mostly proof-of-concept, written to see 1) how hard it
would be (not very), and 2) how useful RSS feeds tracking local file
content might be (remains to be seen).

Cool stuff. One possible use: So that people can monitor file changes
and additions on a shared file server. Say, John in marketing wants to
be notified of every file (strategy doc, project planning document,
email) that mentions the word "Java" or "Ruby" or whatever. (Of course,
as with many other apps like this, the privacy issues are
considerable.)

F.

 

[James Britt]

Cool stuff. One possible use: So that people can monitor file changes
and additions on a shared file server. Say, John in marketing wants to
be notified of every file (strategy doc, project planning document,
email) that mentions the word "Java" or "Ruby" or whatever. (Of course,
as with many other apps like this, the privacy issues are considerable.)

Privacy is a big issue. I ran across an article (sorry, lost the link)
that points out that the search tool is really meant for single-user
machines, as it does not distinguish among user browser caches. As far
as I know it will index shared drives as well as local ones (don't have
a shared drive right I can test this on). But there is no way to select
what drives or folders get filters, other than some high-level choices
by document type.

Conceivably, one could restrict the search to only MSFT Office docs, and
offer feeds from that search base. Then you could monitor documents in
an office across shared drives.

James