search

AAFC-BICoE / dina-planning

AAFC-DINA planning repository
3 stars 2 forks source link

Copy down determinations from parents #337

Open ssbilkhu opened 2 years ago

ssbilkhu commented 2 years ago

From June 16th 2022 Demo Meeting:

Discussion: Claudia - a button copy down determinations from parents would be useful A: we will revisit this topic

GIVEN I have accessed DINA as [a role]

WHEN [describe the action initiated by the user]

THEN [describe the response of the system]

AND [describe additional action(s) from the system]

dshorthouse commented 2 years ago

Would these still be called determinations, or might they lack features (eg a determiner who may never have looked at the children to make their assertion) & instead be called something like asserted identifications? Another way of asking: Is there a difference between an identification and a determination? Should these be modelled differently? Or, yet another way of asking: Would such a copy down determination function be a feature that facilitates data management (because the data model & software now seem to require it) or is it a feature that is required to best represent the intentions of the determiner?

jmacklin commented 2 years ago

This is tricky and may need broader consultation. Also, isn't the other direction relevant as well: children up to parents? If a determination is simply an assertion then it should not matter who/what made it (an expert vs a machine/algorithm). The issue becomes one of trust based on what is known of the determiner. In some cases this will be straightforward especially in research use cases where the data is close to the person doing the data management (typically CM) and who is making the decision about where to store an object based on more than one determination (filedAs). However in legacy cases, it may be less clear which determination to "trust". Since this use case is mostly driven by identification by sequencing, perhaps the real question is are we considering machines/algorithms sentient and are they then de facto "Agents"...?

dshorthouse commented 2 years ago

In addition to the motivations/provenance/intent, there is also a mechanical issue that makes this additionally tricky. Each determination is bundled in the concept of an Organism. And so, copying down/up any determination along a trajectory of parent/child material samples also requires some care in handling each bundle of organism. Which organism? All of them? A selection of them? Are they the same organism (= same internal UUID) or are they different? Is the parent origin material sample a "pure" strain or does it have a mix of organisms? Copying up/down determinations may make illogical presentations if we don't additionally show from where the determination originated prior to it being copied up/down.

banchinic commented 2 years ago

I want to add that I would like to have the option to copy the determination, not have it done by default. If I choose to copy the determination of the parent, in my case it is because I have verified that the culture is pure and that it is the same organism (well multiple spores of the same species...).SeqDB does this by default, all the specimen replicates have the same determination as the specimen they were created from. Which is nice and saves us time when it is a pure culture but not always my case. I'd say 80% of the time the children will be the same determination as the parent. That is why the option to copy would save us a lot of time during data entry.