User:This, that and the other/unsettled policies
Fodder for future discussions and votes:
- Names
- What does it mean for a name to satisfy CFI's attestation requirement?
- At RFVE we have kept entries with (non-durably archived) evidence of three individuals sporting the name (Talk:Lambda) and deleted entries where this threshold isn't met (Talk:Dniprova)
- WT:COMPANY
- What we currently have in CFI is quite inadequate. We keep a proper noun sense at Microsoft even though it seems to objectively violate WT:COMPANY as written.
- Some rough ideas for clearer inclusion criteria: The name must satisfy one or more of the following:
- figurative use similar to WT:BRAND
- evidence that a wide variety of lexical terms have been derived from the company name
- surprising translations? (Does this ever happen?)
- unexpected inflections (in inflected languages)
- Presence of "the" at the start of entry titles
- Gets debated at BP occasionally
- Snowclones
- Appendix:English snowclones is a black hole. Content is poured in by editors but it is never seen by readers.
- Unless they follow one of the rare redirects from mainspace (orange is the new black), no casual reader will ever find these entries, because:
- (a) they're in the appendix, so they won't find them using our search box, and
- (b) the word snowclone is very obscure, so they won't find them using Google.
- The use of algebraic variables in entry titles makes the problem even worse: how long does it take you to work out what Appendix:Snowclones/X Y is X or Appendix:Snowclones/X are X refer to?
- perhaps we could use POS labels instead of random letters: "ADJ NOUN is ADJ" and "NOUNs are NOUNs"
- ...and then we could move them back to mainspace? That's the dream
- Otherwise at least give the appendix a name that people can understand, like "phrase templates"
- Set up proper attestation criteria for mainspace - perhaps 5 durably archived cites with different variations, none of which is the obvious "original version" in cases where one exists (Appendix:Snowclones/to X or not to X)
- For ease of discoverability, perhaps we need to be bolder about allowing the "original version" of these phrases in mainspace even when they wouldn't pass CFI normally, e.g. to boldly go where no man has gone before
- Then they will show up in the Wiktionary search box, and this might help Google too
- Although it might be sufficient just to create them as redirects to snowclone-space
- Headers
- EL currently has a list of allowed and disallowed POS headers
- But some languages have been flaunting this list for years, to the point where it is essentially useless
- Currently the list is "global" without any distinction by language
- I think we need to reorganise this list by language or language family, specifically allowing certain headers in particular languages only
- e.g. "Definitions" is allowed for Chinese entries only
- "Word" is allowed where it is unclear which of several parts of speech the term belongs to (compare kelaunikui)
- "Initial" and "Final" are allowed for certain North American languages (can't remember now)
- etc
- See User:This, that and the other/EL headers
- Good redirect? Alternative lemma forms
- Lemma forms which are likely to be searched by a person who is moderately competent in the language ("knows just enough to get themselves in trouble") upon seeing an inflected form
- e.g. Italian autonominare → autonominarsi (reflexive verb)
- e.g. Latin consum → coest (it's an impersonal verb, but how can you know that just from a single instance?)
- Lemma forms which are likely to be searched by a person who is moderately competent in the language ("knows just enough to get themselves in trouble") upon seeing an inflected form
- How to handle redirects clashing with real entries