This one has low key bugged me for a long time. I do a search to pull up models with the exact string “stealth press””… which should give the the two or three models that have that string in their name, right? Wrong. I’m like, my bad let me enclose that in quotes so you know it’s “stealth press” together as one string”. Which pulls up the exact same results as before. The models with those exact words don’t show up at all in my search. One single model in the search result’s list was related to a stealth press, but I didn’t see any others. The fun thing is it does show a lot of other creators presses, Dremel presses etc. it feels like the search is adapted to show you what makerworld wants you to see.
It feels like Bambu Lam has their search weighted to make you browse through a group of possible related models regardless of what you put in. I already get that on the main makerworld screen, on “related models” at the bottom of each listings page so I don’t need or want that when I am looking for a specific thing!. If nothing else, put the models that have both words I’ve used first in the results then you can do whatever you like below that.
To add put that extra little knife twist on top they claim that this is sorted by “relevance”. Ugh.
There are standards for search engines for a reason! Boolean searches being ignored is a problem. If I do a search for “common sense” and it does not find any exact matches to “common sense” in the set, it should show me that there is, in fact no “common sense” available to be had.
Bambu Lab should be embarrassed that I can get better results by leaving makerworld and using google. As an IT professional, this is some “late 90’s” level functionality and it’s just embarrassing.
I understand that part of the problem is that some (can I say as shole? Probably not) “unscrupulous” designers pad their mediocre level models of with every keyword under the sun, which highlights another “gap” in the system. Why do you allow it? If a person is using more than about 5 tags, they are trying to cheat the system. If you model an espressos press, a hot set socket, dremel, machining etc then “press” & “espresso” “coffee”, “puck” and “tamper” should lead anyone wanting a method of compressing finely ground coffee into a puck right to their door. Put a limit on that since we obviously can’t be trusted with that missing common sense.
I suggest going into Account Settings > Preferences and look at Model Preferences to see if there’s anything in there that may be weighing your search results.
If you ever tried searching for a news article on BBC News with the entire headline p, you will not find it. You must use google.
Unfortunately, this rarely works as it requires the designers implementing the feature.
Weighted, yes. Poorly, definitely.
It doesn’t help that they think they are adding things in the backend to stop the bad accounts from high jacking the results. They don’t appear to act on those accounts either, meaning the search feature can just search rather than act as a blocker.
I believe they think their “intelligence” is a bonus when it makes things worse.
If they did some enforcement of making sure the tags entered for a model at least matched the content, that would be a start.
But, until they want to fix search, it will nit get better.
For someone who can’t find a stealth press you sure do wanna condemn and judge others.
Looking for your search string brought up dozens and more versions of said press, and it’s accessories. Perhaps you should climb down off your horse and maybe take a look at what YOU are doing wrong before throwing stones.
That’s something I’ll look into next. If that was the issue with my search example, I’d be the first to thank you. But you’re missing the point.
Since this site started, we’ve seen waves of fraudulent or predatory practices to game the rewards systems. The company could implement simple measures to address these gaps. Enforcing their own rules would be a good start, as fraudulent behaviors violate the TOS.
My point wasn’t about that specific search. It’s about setting common-sense limits to curb some of the nonsense and predatory behaviors.
I’m glad the search isn’t broken. Those results were odd. Almost every search I did came up with the same, possibly due to a corrupted cache. I’ve been in IT since the early 90s and understand how most search engines work. I retired before narrow/LLM AIs became common and rapidly expanded. Thanks again for the sanity check!
Exactly my point, although I think your phrasing was better than mine. It’s such an obvious trend (along with “I usually don’t do this but…” at the top of every single model some people have uploaded. Dude, at that point it’s “I always / frequently do this…” Just own it, if you are going to resort to that. Some very successful / popular designers do it and they aren’t exactly hurting for downloads.
I had to make this longer than 25 characters so feel free to ignore everything past the comment on your phrasing. The rest is just opinions on soliciting to get rewards from people whether or not they intended to download, print or use their designs in any way. To me that’s the next hurdle… or just something to bitch about.
That could be, although I don’t typically use filters. I like to follow trends and look for talented people - just skimming in to find inspiration and ways to be a better modeler. I appreciate the tip, and will try to recreate the issue and post updates if I can find anything interesting. Thanks!
If you do not care about other people’s time, and only care about algorithmic search juice feeding you company scrip for engagement clicks: ALWAYS tag and embed text that will be seen by an opposition. “Jesus” tags on satanic figurines*, etc. The 10 comments you get from complainers will boost your visibility more than hundreds of dollars of ad spend. To be clear, this is a dark pattern, but it is Bambu’s choice in software that embraces and amplifies it.
