dcramer / peated

https://peated.com
Apache License 2.0
64 stars 13 forks source link

Kill batch information #209

Closed dcramer closed 3 months ago

dcramer commented 3 months ago

Its frankly just duplicate data. Yes they can be different, and in some cases a lot different (though maybe call them something else), but e.g. Ardbeg Traigh Bhan different batches are mostly irrelevant. Those can go into the tasting details rather than bottle details.

Id rather focus on release years, albeit those are also mostly duplicate.

We're still going to have to clean this up at some point and have the system auto collapse bottles into each other, and then allow the variants to exist (e.g. Traigh Bhan is the bottle you're recording, and theres a variant thats Batch X, or Year Y).

dcramer commented 3 months ago

I may reconsider this but this is the biggest area of frustration right now. Its complicated if we keep every variant, its wrong if we don't.

dcramer commented 3 months ago

One last thought:

The biggest issue here is deduping variants. Its really tough to normalize the data (I'm still doing a bunch by hand), and when I search for "Traigh Bhan", I'd rather see one result. Maybe selecting that result would then show me all variants of it?

The question is, if we do that, and arguably thats the ideal experience, how do we actually dedupe like that? Do we also have this 'aggregate bottle' page? What if you don't know which year of release? Sometimes its not so obvious, is there a generic one?

The problem with keeping that bit of metadata is it complicates usage for most people vs experts. The problem with removing it, is releases can vary quite a lot, so aggregating ratings/flavors/etc in across them is not great.

dcramer commented 3 months ago

One more wrinkle: a lot of stores wont index on batch number, so maybe its best to collapse them and ignore this after all?

We could still bring back per-tasting "excessive details" if we wanted, and I think its probably ok if ratings/flavors/etc are blended across all variants of a bottle. If we added the structured data per tasting that could suffice.

There is an open question though in generla of how we should treat Single Cask whiskies. Those doesn't always imply the same thing as a batch, but I think Single Cask in a lot of cases are different, but not always. Probably a subjective decision. e.g. SMWS its always going to be single cask.

A bunch of other cask releases? In this world we'd bundle them together as its not useful detail.

dcramer commented 3 months ago

Closing in favor of #210