What is the general consensus?
Better with labels, or without?
…or even if no label is entered, the node collapses and displays no label…
More importantly, is Nick still thinking about making the link
nodes the same size as regular nodes? Last year we had a nice (long) discussion on Slack, where we discussed the merits of changing the link node shape/size/behavior. Unfortunately, the Slack history does not go back that far anymore.
Talking about the size of the nodes, it would be great to have the ability to create “icon-only” nodes like the “link node” for custom nodes …
Would it be possible to keep the link node the same size as it is now, except if a label has been entered by the user, then it becomes full size?
ie ' cake & eat it' !!
Well - You mostly would want to add a label as it makes it easier to pick from the list of other links to join to… but may or may not want that to show.
My own wishes:
- Yes to labels - actually “Names” to fit in with other nodes.
- Allow them to expand horizontally to fit the name to the same degree as other nodes but keep the default width as-is because that helps clearly identify them in flows. This would allow names like “1”, “2”, etc for those who want to keep things minimalist.
So, if left blank, would the list name default to the node ID - same as now.
I think this is right answer for the long term.
Near term (i.e., I do this now) the link node icon can be modified easily to give results like:
With the new icons in the appropriate directory (/usr/lib/node_modules/node-red/public/icons in my installation), they can be selected in the node edit dialog from the node-red drop-down list of icons under "node settings." My experience has been that only a few of these are needed in most flows.
well currently yes - but I’d like to be able to add a label for the “list” - without having that shown - as I also like the minimalist style that keeps them identifiable.
(though I could live with it)
… but no to overwriting the icon !
I’m confused (as usual). Isn’t it already the case that you can name a link node and have that name appear in the list of possible connections? So the proposal is to have that name appear in the flow window as well? I think that a long, descriptive name for the link is more valuable than trying to keep names short enough not to eat up space in the editor.
OK Dave, I get the aesthetic objection to messing with the icons, but are there other reasons not to go this route? Do I risk breaking something on my system?
I like the small size of link nodes but would like to able to label them (using Name) but only have them resize to the smallest size possible to include the name - something like...
from
to
A, B, 1, 2, etc would not be my choice as the labels are meaningless IMO.
I would prefer to be able to label them;
Temp, Humidity, Voltage, etc, so it indicates what the flow contains.
@drmibell - No - and indeed you are indeed free to add custom icons and set them as you do… nothing should break - but either making that the default way for users to do this - or for a single char name to automatically be placed on top… no thanks
Thanks. I haven’t changed any of the defaults – just made a few new icons available if I want them. I like being able to see where connections go without having to select one of the nodes. Also helps when I have saved or printed out a screen shot.
@Paul-Reed It’s partly a matter of what you’re used to. In electronics, schematic diagrams often use this trick to show where connections go in a circuit, especially when it would be too confusing to show all the links or where they extend across multiple pages of a diagram. I think that functional names belong on the nodes that do the work. In NR, the wires have no names or properties other than their endpoints, and logically speaking link nodes are the same thing.
Yes I get that. As a retired electrical technician myself, I can see the comparison… but it would be good for the wider community to be able to choose whichever link identity they are comfortable with.
Link nodes also do the work, they direct the flow in specific directions, so why shouldn’t they have an identity?
@Paul-Reed One of the great things about Node-RED is that it is mostly self-documenting. This discussion is about one of the few cases where it falls a bit short, at least in the view of some of us. Even so, there are lots of ways to keep track of what flows do and (with the project feature) what we do to them. I’m all for adding new ways to do documentation, and I trust the developers to keep things consistent with the basic architecture so that changes add clarity and prevent confusion.