Hello. I'm an extensive user of the original NR Dashboard, and have only today downloaded Dashboard 2.0 to start playing with it, and I have some questions about its future and development.
On one hand I am very excited by the new possibilities and already I like it.
On the other hand, within only a few minutes I found quite a few bugs / unfinished features. I understand it's sometimes good to get a product out into the public but was surprised the release version is 1.x.x as that would imply it's mostly ready for production.
Please correct me if I'm wrong here, or if my expectations are amiss.
Initial reflections:
-
I immediately started using Switch node. It works exactly the same as the old one, so I'm really happy. I also noticed there's a "group switch" node which creates multiple switches in a group for toggling between options which I think has a lot of possibilities for dynamic automations e.g. light scene selection
-
Next I brought in some "number input nodes". I prefer the UI here compared with the old Dashboard, however it's simply unusable as when I set the range (min/max/step) it just didn't work. So I was setting some heating target temp, and I when I clicked the up arrow I'd expect it to show me the number at the beginning of the range I selected. Instead it showed "1". Then I clicked down and was able to set my heating to -1 degrees. Having viewed the Github issue mentioned above I'm left wondering if this is a bug, or if it's just something that hasn't actually been implemented yet. I was also disappointed there's no "value format" field as in the previous version. So when I set a temperature, there's no obvious way to show to the user that it's measured in Ā°C.
-
Next I added a gauge, and spent a while setting up the "segments" (these are the different colours for different thresholds of the gauge). Also I set up the range for the gauge. I enjoyed spending time choosing the segment colours and thresholds. Unfortunately when I deployed, the segment colours just didn't show at first. Then I changed another field and re-deployed, and suddenly they showed. Then the problem happened again, so I changed the gauge type, and immediately lost the threshold colours, threshold levels, and range, all of which I had carefully set and planned to duplicate later. These custom thresholds just disappear whenever you change the gauge type, which is not well thought through and very frustrating for the end user. It turned me off using it altogether.
-
Also in the old "gauge" widget, it is possible to set the "Value format", which means two things: (a) you can display a value with a given format, and (b) you can plug this directly into a node that outputs an object full of different data, and choose which one you want to appear in your gauge. Example: {{msg.payload["AM2301-12"].Temperature}}
would mean the gauge shows the temperature type from the object. Unfortunately this just isn't supported in the Dashboard 2.0 version.
-
Next I looked into using UI templates within a subflow (newly supported by the revised framework thanks to NR 4.x.x). When I looked into this initially it seemed I couldn't leverage this new NR functionality with Dashboard 2.0. I have just spent a lot of time converting a lot of my subflows to integrate template UI nodes, where previously I had to use a lot of workarounds which were really messy. So I looked into this and found an issue was opened in Dashboard 2.0 back in March of this year, and basically there's no update and it's still open. [edit: not yet sure, but perhaps this is resolved but the issue remained open. Will look at this more closely.]
-
Finally I wanted to try something small: to see if there was a UI LED (little indicator) that I could use to replace the one I use for the old dashboard. I was glad to see there was one. I installed it (@flowfuse/node-red-dashboard-2-ui-led
), added to my dashboard, and ... it just doesn't work. Neither the little circle nor any labels are even displayed on the dashboard.
It is not my intention to simply complain, but rather to provide the following constructive feedback. After all, I really think the original Dashboard is ready to be ditched for something more modern, specifically I have issues with the lack of flexibility in layouts, theming, and generally feeling it's dated. So I'm overall positive about Dashboard 2.0. But if Dashboard 2.0 is going to be even slightly successful then in my view it should start by being far more transparent about the fact it's not ready for prime time. Could it instead be sold as a work in progress? Change version numbers to 0.x.x and not 1.x.x which heavily implies that the majority of features that are clearly designed to work, actually do work. (For example, when you include a field in a node edit dialog, the user should absolutely not expect to be the one to discover that it doesn't even work.)
This doesn't mean every feature must be supported, but it does mean the thing is not full of bugs and "discoveries" which the user needs to find out himself, which lead to the user basically becoming disheartened. I'd suggest this is critical to managing the expectations of users who start to use a product.
In my view if you don't manage these expectations from the beginning, people will very quickly form strong opinions of a product that it's unreliable, and those opinions will never go away. This is in stark contrast to Node-RED itself, which has a very well-disciplined approach to this, which is exactly why it's so successful.
The experience has led me to ask the following questions. And I suspect other users in time may wonder the following:
-
is Dashboard 2.0 over-subscribed? I mean, is it trying to do too much, without enough contributors to succeed quickly enough? When I say "quickly enough" there's a sense that it must at least be able to keep up with the development of Node-RED itself.
-
how can I be sure it's worth investing my time to migrate my extensive dashboards over to 2.0 if there are issues that are opened then simply left for a long time with no update (e.g. this issue opened in March)? This makes me feel very insecure that there's a good future for a product.
-
does this product rely very heavily on community development? If yes then why isn't this made far more clear from the outset? The way it's presented on the flowfuse website is like it's a well-polished product.
-
do the developers wish for any level of consistency of user expectation between Node-RED itself and Dashboard 2.0? Will Dashboard 2.0 become the new official dashboard for Node-RED? Have any comments been made by Node-RED core developers about the future of Dashboard, or did this 2.0 come about out of frustration by lack of comments on this by core NR developers? Is this 2.0 accepted by NR core developers as a viable replacement, and do they approve of the way / direction it's being developed?
[edited for tone]