<br clear="all">I've tried to create a twitter runner today for the sake of knowing how to make one from the ground up. <br><br>The twitter engine has several interfaces, but I only wanted to use the send information "interface" (or should I call it hack?).<br>
It seemed pretty straighforward but I got a message "Authentication required". <br>As I have the twitter applet up and running I suppose that each DataEngineManager::self()->load("twitter"); will do exactly that and create a new engine. That needs to be configured. <br>
So it should be perfectly possible to have several applets with different users logged in. What is a good thing.<br><br>But as it needs to be configured with a password the runner doesn't know, we need some solution:<br>
<br>- Send the data from runner to the applet and let it manage it. The applet has all the information of configuration needed. But if we have several twitter applets loaded what applet will we call?<br>- The engines should store the configuration, not the applets. The applets will read and write configuration in engines. It will be always possible to run an engine without configuring it (default configuration), of course the default configuration may be done by the user but once done it is default can can be used automatically.<br>
- Create a GUI to configure this runner. This GUI will be launched from the GUI that eventually will be done for whitelisting runners, prioritizing, etc. <br>- The engines have their own configuration gui. The configuration gui of the applets can call the configuration gui of the engines. A way to launch its own GUI if still unconfigured and someone tried to use it for engines.<br>
<br>Maybe none of these ideas are good solutions. Or maybe they are desirable for some point in the future. <br>It may be a solved or trivial problem. It is 4.30 am here and can't really think clearly :P<br><br><br>-- <br>
Jordi Polo Carres<br>NLP laboratory - NAIST<br><a href="http://www.bahasara.org">http://www.bahasara.org</a><br>