Reading Web Pages in Raw Mode
Continuing from the introduction page, we send the message 'news.google.com' to web@db4.us and get back the following:
The above screen is the textual translation of the news.google.com page. Normally you don't pay attention to the top of the news.google.com page, but here you will need to scroll down to interested page area, such as the headlines.
In theory, web@db4.us can translate any web page into text.
In practice, web pages of mostly textual contents will work better, because
you would not have to scroll around so much to see the interested contents.
Mobile sites like
m.usatoday.com,
and mobile version of news sites like
mobile.sfgate.com or
m.latimes.com work much better.
The hyperlinks of the web page have been labeled with integers starting from 1.
The above screen is showing the first 9 hyperlinks (in parenthesis) of the message.
If the page consists of any image, the image link will be surrounded by braces "{}" rather than parenthesis.
To achieve the clicking effect of a browser,
we send another message with the link numver.
The above command will retrieve the page corresponding to the 8th link, the shopping.google.com page.
After the page shopping.google.com is returned, if we see another page that we are interested in viewing, say with link 78, then we can keep sending more link messages:
Here is a video demo of using the click option.
Go from one page to another...
Given the following news.google.com text page:
Google News
[(1)Click here for ARIA enhanced Google News.]
[(2)Web]
[(3)Images]
[(4)Maps]
News[(5)Video]
[(6)Gmail]
[(7)more]
[(8)Shopping]
[(9)Groups]
For instance, from the google news page, the "Shopping" link
has been labeled with 8. To achieve the effect of clicking
on the "Shopping" link, the following text message can be sent:
8
78