It’s very mercurial!!
It was doing this yesterday morning but by the afternoon typing, “input” brought the node to the top.
Right now, less is more. 
3 Likes
I love how the search is different between every version of Dynamo, every person using Dynamo and even what time of day you are using Dynamo.
Here is Dynamo 3.1.0 (sandbox) at 7:47am MST
Here is Dynamo 3.3.0 (sandbox) at 7:50am MST
2 Likes
Call it AI and now it’s a feature.
4 Likes
I honestly think the search really should’ve been a Dynamo package.
Dynamo installs come with whatever version, but this would let users update it.
I’ve said it before, but I will say it here again:
People getting different search results between minor versions of Dynamo is unacceptable. Imagine being in a beginner training class and your Dynamo search results look different than the person next to you. The presenter assumes it is all going fine, so now you are lost and feel like you did something wrong!
4 Likes
I’ve always assumed it’s a memory/performance thing. May be anecdotal, but I feel like I tend to see DLL nodes show up higher than packaged DYNs. Seems like the search function returns results based on how fast it finds them and not necessarily how strong of a correlation there is.
1000000001%.
1 Like
A note on John’s example above, the Input and Output nodes should be suppressed in a workspace environment, meant to only surface in search when in a custom node environment (so it showing showing up in 3.2 at 8:06MST a different gremlin).
Something like ‘string’ should be consistent though… and it isn’t (source/proof not needed).
I read this while I was in the office… a minute or two later someone came over to my desk to ask if I was OK because they could hear me laughing so much.
2 Likes
The lone improvement is the feature ‘autocompleting’ to search for input componenets to those you’ve placed. The simplest, selection/drop down lists that were impossible to find are a little easier to find so long as I have the right element already in the definition space.
Still unacceptable from the perspective of the broader software interface but, I guess it is something.
What is this hahahaha 
1 Like
@Alien Thanks for bringing this to our attention. The variance in the search results by time, smells buggy and we are looking into it.
Thanks to everyone else on this thread for the feedback on the varying search behavior by version. The Dynamo team has been actively working on improving the performance and accuracy of Lucene search, since it was released in Dynamo 2.19. We want the search experience for Dynamo users across the board, both new and experienced users to be consistent and reliable.
While we have some significant challenges in getting Search improvements across the line in historical versions of Dynamo, for many reasons that mostly amount to risk, we are actively doing our best to ensure the overall Dynamo Search experience, across every version of Dynamo, is the very best it possibly can be. Hence, these improvements have been delivered in smaller chunks in the recent versions, so that we can make the search experience comparatively better while we continue working on the next chuck of improvements.
We made performance updates to Lucene search in Dynamo 3.1, 3.2 and 3.3 releases. As you can see in the snippets below how when searching for “String” in the recent version, the input node for string shows up on the top of the list (in 3.2.1 & 3.3 Dynamo versions).
Note: We also fixed a bug in 3.2, so that the input node only shows up when you search for it in custom node workspace. This might have caused some confusion as highlighted by others in this Thread as this was not captured in the Release notes. We can do better and will have a more robust process to capture updates and bug fixes in the release notes. "
These smaller chunks of improvements over time to make Dynamo search perform how it is expected to function = the search behavior varies from version to version until we get to the finish line.
We understand how these regular updates can create an unfavorable situation for users as with every update, you are seeing a slightly different behavior. Thanks for sharing your feedback, we will explore ways to navigate the challenges across our multiple teams and investigate how we could make the updated Lucene search behavior replicate in older versions to maintain consistency across versions. Meanwhile, the team is continuously working to improve the performance of Dynamo search, to show the nodes on the top when the entire node name has been typed by user and fix the outstanding search edge cases like “Select Model Elements”, “Select Element by Category”,Etc…
4 Likes
Personally I wish Autodesk released Revit less often… The yearly updates feel rushed…
A sceptic would suggest it’s purely financial and not to the benefit of the customer.
Same goes for Dynamo… I’d much rather have smaller changes, less frequently that didn’t have so many bugs.
Maybe I’m alone in thinking this… But it really irritates me.
Also, 20+ years and the date format is ONLY American still… 
1 Like
It’s tough - as sometimes the only way to find bugs (Especially esoteric ones) happen when out in the wild. We test ourselves a lot and find and fix many bugs prior to release, but can’t stress test Dynamo nearly as much as all of you working with complex datasets on real-world projects. Our goal is actually to release more so that any of these discrepancies can be fixed earlier; same outcome as you desire, with knowledge and reproducible cases that enable us to do this 
2 Likes
Are these dates in Dynamo or Revit? If it’s Revit you’re giving the feedback in the wrong spot - https://forums.autodesk.com/t5/revit-ideas/idb-p/302
Guessing Revit as Dynamo’s only been out of beta for 8 years now…