Search This Blog

Monday, February 1, 2010

Sikuli - Does that mean it's "sick" AND"cool"?

Image-based Scripting
OK - so what does that heading mean?  Well, over the years (decades) there have been many ways - mainly proprietary - to automate the things we do on our PCs.  These solutions are quite often command line interfaces that do not work well in today's ultra-graphical user interface (uGUI - yep, that's mine, just coined it!)

So what do we need?  We need something that works across the PC operating systems like Linux, Mac OS X and Windows.

Well the bright ones at MIT have a project called Sikuli that provides a new way to approach scripting or rather automating tasks based on graphical or image matching.  You capture the graphical area on the screen then paste that after an action like "click" in the Sikuli editor.  For instance, say you want a way to script your username/userid and password for any graphical application INDEPENDENT OF THE BROWSER YOU ARE USING - then you could do that with Sikuli.  You can fill in forms, automate menu traversal, etc.

But, don't be limited by my explanation, have a squizz at their demo.

This means you can save scripts (or should this be albums/galleries?) of automated tasks like opening up your email, setting admin attributes on/off, etc.

If have just downloaded and installed the Linux version and will update this post with comments after I have taken it for a test drive - stay tuned...

1 comment:

  1. Update: Well I cannot recommend Sikuli under Ubuntu 9.10. I could not get it to recognise any images. I do not think the creators have given Linux much focus - all their examples are Mac & Windows...

    I have posted a question in their site, I will update when I get a response.

    ReplyDelete