<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.32.2">
</HEAD>
<BODY>
Boring "Real-World Details":<BR>
<BR>
<BLOCKQUOTE>
So, we are planning a summer trip to Glacier National Park in Montana. <BR>
<BR>
We would really like to camp at Many Glaciers Campground in the park. However, at present, all the sites are already reserved. They are reserved through recreation.gov, starting 6 months to the day from today. <BR>
<BR>
But we have a really strong impression that people initially book a long stretch, then later either reduce the duration once they get more specific plans, or cancel. <BR>
<BR>
SO, we really want to detect if/when sites become available over the next 6 months and jump on it before someone else does. <BR>
<BR>
There is someone who offers this as a service for $40 a reservation, irrespective of whether they are successful.<BR>
<BR>
</BLOCKQUOTE>
Exciting Technology Application: <BR>
<BR>
<BLOCKQUOTE>
Initially I looked at the HTML for their search page, with the thought of using "wget" to simulate the reservation request. That increasingly looks like a fool's errand, assuming that they may have session cookies related to sign-on and other magic handshake crap that would be difficult to simulate. And what happens when they alter their data fields? <BR>
<BR>
Then I thought: All I want to do is: <BR>
<BLOCKQUOTE>
Setup a browser window on our Centos 6 desktop, any browser that understands https;<BR>
Run that browser through the responses to get it to the search window on this campground, and put in all the dates and related input. <BR>
Then: <BR>
<BLOCKQUOTE>
Run SOMETHING that will automate: <BR>
<BLOCKQUOTE>
Hit the Search Submit button;<BR>
See if the resultant page contains "No Suitable availability"<BR>
IF Not: Email me<BR>
Sleep 15 minutes<BR>
Rinse, Later, Repeat<BR>
<BR>
</BLOCKQUOTE>
</BLOCKQUOTE>
</BLOCKQUOTE>
This sounds to me like a very elemental application of a test/control manager for a GUI interface. If I can automate an existing browser, we can eliminate all the complexities of trying to fake out their web server. <BR>
<BR>
Since this just sits on my desk in the basement, I can live with hard-coded screen coordinates. <BR>
<BR>
What tools exist in Linux to do this? <BR>
<BR>
</BLOCKQUOTE>
regards, <BR>
<BR>
Neal Rhodes<BR>
MNOP Ltd<BR>
<BR>
<BLOCKQUOTE>
<BLOCKQUOTE>
<BR>
</BLOCKQUOTE>
</BLOCKQUOTE>
</BODY>
</HTML>