DQA votes disappearing

[jdmore]

@cthawley

One leading/more specific ID here (and I’m not criticizing the IDer, just the scenario) adding a more specific ID essentially wipes out all that previous ID effort (changes OID and removes DQA votes).

If I’m understanding the currently intended functionality, only a change in CID (not OID) should wipe out previous CID votes. If that’s not what’s happening, then this seems like a bug; otherwise this discussion should probably be in the General category instead.

@cthawley

I think this change tilts the balance unnecessarily in favor of IDers who are willing to make speculative/leading IDs.

At least under the intended functionality, this shouldn’t be the case. It would take several more agreements with the initial “speculative” ID to shift the CID.

But in any case, as long as identifiers are voting on something labeled “Can the Community ID be improved,” I think it is logically inescapable that when the Community ID changes, all of the previous votes become meaningless because they were based on a different CID.

[cthawley]

Oh wow, apologies, I did not know that this change was retroactive (ie, observations that had been RG at genus due to “No, it cannot be improved” would now be casual). That is even more of a mess. I agree with @spiphany that

@spiphany

Note that I am not suggesting that it is a good way to handle this. I very much think it is not and I have said so repeatedly.

and am also confused via situations like

@spiphany

Edit: Though I am a bit confused about what exactly triggers the DQA to be erased vs. staying and making the observation casual; is it possible this has changed in the last few days? I was looking at an observation with an essentially identical situation, where 8 days ago the observer added a species ID to an observation where the DQA had previously been used to make it RG at genus, and the observer’s new ID made the observation become casual without the previous DQA being erased.

Maybe all of the different changes with different timing have caused some weird indexing cases?

In any case, I agree with @jdmore that the initial report definitely isn’t a bug (so I marked jwidness’s post above as a solution) and the other situations discussed, including my own, might be bugs (?) but might just be something weird with how the system works that I don’t understand/understand incorrectly. With so many changes occurring not at the same time, it’s hard to parse. I’ll keep my eyes out for any patterns and make a Bug Report if there’s something I can pin down.

[yongestation]

@spiphany

Edit: Though I am a bit confused about what exactly triggers the DQA to be erased vs. staying and making the observation casual; is it possible this has changed in the last few days? I was looking at an observation with an essentially identical situation, where 8 days ago the observer added a species ID to an observation where the DQA had previously been used to make it RG at genus, and the observer’s new ID made the observation become casual without the previous DQA being erased.

My understanding is:

1. The DQA votes only get erased when the CID changes. Since the new species-level ID only changes the OID and not the CID, those DQA votes remain, leading to…

2. Though this DQA item is disabled when the OID and CID disagree, in this case the votes were placed when they did agree. Since a later ID made them disagree, it became casual.

This seems like, if not an oversight, a pretty undesirable consequence of the current system.

[matthew_connors]

Thanks everyone, I agree that my initial bug report was not actually a bug. However, there are a number of bugs/undesirable effects that have now been discussed here in the comments! Maybe it would be better to move this discussion to General and change the title? Forum moderaters, please feel free to do whatever you think is best for the ongoing discussion :)

[spiphany]

@cthawley

Oh wow, apologies, I did not know that this change was retroactive (ie, observations that had been RG at genus due to “No, it cannot be improved” would now be casual).

No. They become casual if the DQA has been clicked (e.g. so that it is RG with community taxon at genus) and the observation subsequently receives an ID (e.g., a single species ID) which results in the community taxon and the observation taxon becoming different while the DQA is still active.

At least, this is what has been happening since the new requirement that community taxon and observation taxon be the same. The example you gave makes me wonder whether one of the “tweaks” referred to in the recent blog post is to now reset the DQA when the observation taxon changes instead of when the community taxon changes. The only other explanation is that perhaps you misremembered and the DQA was never used on this observation?

[maxkirsch]

@matthew_connors

in any case where a single person’s identification can change the community ID (I think most commonly this would be when there are only two identifications at the finest ID level), anyone can completely erase the votes without any notification to anybody else. Based on the observations I regularly review, that sort of situation comes up much more frequently than you might think.

one thing that could potentially address this particular issue (if the removal of votes when the community taxon changes is here to stay) would be if the system could save votes that a particular community taxon could/couldn’t be improved behind the scenes if the community taxon changes, and then restore those votes if the community taxon changes back; not sure how feasible that is though