Suggestions For Enhancing KNIME

I have a few suggestions for some ease of use enhancements. Perhaps this will generate additional ideas as well. Comments are of course welcome.

Much of my time is spent manually opening a node's outport table, maximizing it, inspecting it then closing it. It would be nice to have a window that is always visible (for example next to the Console tab) that would automatically display the outport table(s) of the selected node. (If multiple nodes are selected then no table tabs would be shown). The same could be done for view outports.

Similarly, the inspection of workflow variables would be much easier with an IDE-like 'Watch'
window (again maybe next to Console tab) that allows the user to select the workflow variables to watch (or alternatively display them all) and the system would display and auto-update the values.
Perhaps it would even allow the user to specify a condition upon which the currently running nodes would pause execution. Example: if workflow variable 'currentYear' == 2010, then pause execution of all running nodes. This would be useful for troubleshooting workflows.

Finally, a button on the toolbar that would distribute the selected nodes equidistant along the horizontal axis between the left-most and right-most selected nodes would be useful. A similar button for the vertical axis would also be useful.
 

Cheers, -Don

Hi Don,

you may appreciate images of a prototype that we put together during last week's KNIME retreat - I'll leave it to you to guess what the little view on the right does ;-)

Cheers, Michael

PS: To appear on a desktop near you in summer 2012...

Wow, that looks great! I see you're even developing a breakpoint node. This will make workflow development and troubleshooting much easier.

Thanks KNIME team!

P.S. It looks like the second screenshot is intended to show the currently selected node's data table, right? (although the tab is labeled "KNIME Variable Monitor"). I'm sure this will be appreciated and will make folks more efficient in KNIME.

Cheers,
Don

...good point - the name is misleading. It started out as a Variable Monitor with the idea of having different types of such "monitor views". But we then moved to making the monitor more powerful, allowing to look at various aspects of a node (output table, variables, configuration)... Lots more needs to be done but the general framework is there.

Michael

PS: I am glad you like the Breakpoint node. Bernd hates it ;-)