We won’t be changing anything right now
The system restrictions around changing this field means that we have no realistic options for changing the tag recommendation text in a way that meets our goals. Our most practical option is to lock one of the tags to [dnd-5e], our primary suspect, but we will be unable to measure whether that has any practical benefit.
So, given the lack of good options, right now we won’t be changing anything.
We can’t change much except maybe making a tag suggestion static
Here is our current tag prompt when asking a new question:
It follows this format, with each tag randomly determined:
e.g (tag1 tag2 tag3)
According to two SE staff members, Catija and Shog9, changing the prompt can be done with developer assistance. However our options for making this change are very limited:
- We can’t change anything outside of the parenthesis
- We can make some/all of the tags determinate/static instead of randomly generated
- We can have a combination of static text and tags (static or random)
No viable options for changing the text
Syntactically, that doesn’t really leave us any way to clearly indicate that users are expected to include a system tag if it is needed for their question and that there are a variety of systems to choose from. And that was the major goal of the change. So, there’s nothing we can do here to accomplish that.
As far as I can tell, the only thing we might find helpful to do would be to lock in one or two of our most troublesome system tags and hope that it helps people see to use them. However...
We need data for this change, and there’s no way to get it currently
Before considering any change, we need to find a way to measure the current issues surrounding missing-system-tag questions so that the effects of the change can be measured and so that they can be reverted if the change is not working. However, we’ve not yet found any way to make these measurements.
This is important because pinning a system comes at a cost of other tags that might have been displayed and a more philosophical cost of perhaps appearing to favor those systems as site-wide focuses.
Without a data source to measure success/failure and with no clear options for changing the text to accomplish the goals, we have decided the best course of action is to make no changes right now.
If anyone has any suggestions for measuring this issue (or for some solution we may have missed), please don’t hesitate to send it our way.