structured data adr #407
structured data adr #407
Conversation
docs/adr/ADR-01-STRUCTURED_DATA.md
Outdated
|
|
||
| At this point we need to introduce some more precise terminology: | ||
|
|
||
| TODO move or copy this glossary somewhere else? |
There was a problem hiding this comment.
would move the final documentation to SEMANTIC_CONVENTIONS.md and only link it here or explain it in short
docs/adr/ADR-01-STRUCTURED_DATA.md
Outdated
|
|
||
| Knowledge about markdown is important for rendering it correctly. | ||
|
|
||
| The format is normal [Markdown](https://en.wikipedia.org/wiki/Markdown). |
There was a problem hiding this comment.
I would use a different wording.
e.g. "We agree to use standard markdown according to wiki"
docs/adr/ADR-01-STRUCTURED_DATA.md
Outdated
|
|
||
| ### Lists | ||
|
|
||
| Knowledge about lists is mostly important for contains searches and search faceting. We cannot sort lists. (TODO can/should we?) |
There was a problem hiding this comment.
sorting according to list values should be done by client, not?
it would be more complex as we would also need some information about the sort order. maybe we have to implement this later when we find a use case for it in the publisher, but for now i would not
There was a problem hiding this comment.
yeah i also think we should not support this right now
docs/adr/ADR-01-STRUCTURED_DATA.md
Outdated
|
|
||
| Knowledge about numbers is mostly important to correctly sort them. | ||
|
|
||
| The format looks like: TODO define format and possible numbers here |
There was a problem hiding this comment.
format is just javascript number... or not?
There was a problem hiding this comment.
i would like it to be more specific, what do we want to support? only "normal" numbers like 2 and -3? floating point? hex values? e notation numbers?
docs/adr/ADR-01-STRUCTURED_DATA.md
Outdated
|
|
||
| #### Language Variant | ||
|
|
||
| Has the VariantName "lang" and the value is a two-letter language abbreviation, like "en", "de", "fr", ... |
There was a problem hiding this comment.
how would we denote the "original" language?
Or do we use some kind of indication of collected value vs enriched value?
There was a problem hiding this comment.
hmm
maybe an event already has two languages on it's source?
i think we should just annotate generated values extra, the rest should be as is
we could think of having no lang variant as the default?
closes #367