It pains to have to say that I have all but given up on convincing my colleagues to use Knime. Even through i presented my work with Knime in group meetings several times (both as powerpoint presentations and “live”), I recently noticed that even our very computer-literate IT support guy, who has also seen my presentations, still has not grasped the basic Knime principles. I wonder why this is, apart perhaps from me being not a very gifted presenter.
One of the reasons, I think, is how well spectators are able to follow what is happening on-screen when I click around in a workflow. And I think here is a major issue here for people unfamiliar with Knime: when I right-click on a node to check the output, I get a menu with no fewer than 20 complicated options, none of them with the obvious label “Node output”. An option corresponding to “Node input” is totally absent. I think this is a problem for newbies.
Suggestion: replace the usual right-click context menu with a simpler one containing only “Node output”, “Configuration” and “Execute” options, perhaps add a “Node input” option. Make all the other options appear only when i use ctrl-right-click.
I’ve always thought it was a bit awkward that the action for viewing a node’s output was tucked away at the bottom of the context menu. Maybe simply moving the output view options to the top of that menu would help? Or having say a ctrl-double-click on the node icon open the output view (maybe with a tabbed viewer window for multiple output ports).
Personally I think the KNIME window with the explorer, workflow coach, node repository, console and description panes all crowding around the main canvas can seem a bit overwhelming. Maybe there’s something that can be done to improve the focus on the workflow itself, without making those other necessary and useful panes too hard to discover and access. I realise that the Eclipse platform seems to lend itself to these pane-heavy designs though :-/
Hi @Aswin , I think a lot of people live in a “I don’t have time” environment, and for them to want to learn something, it has to be worth their time.
People will usually spend time learning something if it’s bringing value to them, so if you want to have them wanting to learn and use Knime, you need to show them what Knime can bring for them.
You don’t need to go into details about how Knime works, at least not at the beginning. Your first step should be to demonstrate how Knime is doing what they need, and how it is to do it. Once that interest is planted, they will be willing to learn more about it, at that point, they will look into how to do things in Knime on their own.
Indeed, I agree that it is not obvious sometimes how to look at the output/results, but you can just explain that in most of the case, the last item in the right-click menu is what allows them to see the results. And that’s one of the things that I like, I can see the results at any step of the workflow, so I can work forward or backward if I want to pinpoint an issue by looking at the results of the nodes.
The other advantage that I find is that you can fork the workflow, that is you can have the output of a node go to as many nodes as you want, so if you want to experiment with something, you don’t have to go and modify your current workflow, you can just fork to a different workflow without touching your original one. Two things here: 1) You don’t risk breaking your original workflow; 2) since you have 2 different ways, you can compare both results.
@bruno29a I have explained the advantages that you mention and others (such as the tight connection between data and settings/algorithms) so many times, but it somehow doesn’t “click” with my audience. One half prefers to stick with Excel/Graphpad, the other half with R/Rstudio. As for explaining Knime in a presentation, in my opinion showing works much better than telling; but it needs to be obvious for the audience where they should look, and overcrowded menus don’t help.
@Tom_Hawkins perhaps Knime can introduce a “Newbie perspective”, with only the panes necessary to make a simple workflow (only canvas + node repo + description) and a context menu containing only output + config + exec… This perspective can then also be used by experienced users whenever demonstrating something to an audience that is unfamiliar with Knime.
KNIME Junior Analytics Platform
One more idea: open the node output by double-clicking one of the black output triangles (or output squares). The UI already shows a tooltip with table info when hovering over these UI elements, so having the table appear with a double-click would be quite intuitive, especially at higher zoom levels.