User:Snthdiueoa/On notability
I thought that I would put down some thoughts on my notability criteria since I seem to spend a lot of my time in AfD discussions.
Overall, my position is approximately halfway between "inclusionist" and "deletionist". I agree that we should be fairly lenient about what we allow in Wikipedia; however, at the same time, it seems that there are far too many articles that scrape through AfD debates because they can demonstrate "non-trivial" (ie two sentences rather than just one) coverage in "reliable secondary sources" (ie two local newspapers rather than just one) and it makes Wikipedia look a bit of a joke. So, here are my exact criteria:
The minimum standard: the source test
In order to comply with Wikipedia's policies on verifiability, no original research and what Wikipedia is not, we need to set an absolute minimum in terms of sources. The notability guidelines say that "a subject is presumed to be notable if it has received non-trivial coverage in reliable secondary sources" but it is ambiguous about exactly what constitutes "non-trivial" coverage. The absolute minimum standard is therefore this:
Any article must have sufficient coverage in reliable, secondary sources, that when all statements not based on such sources are removed, and any additional available information is added from such sources, the remainder is more than just a stub, or a news item, or anything else listed in what Wikipedia is not.
Furthermore, I require that the subject be given more extensive treatment than the size of a stub in an absolute minimum of three reliable, secondary sources. These sources do not have to be in English, though I would expect newer sources to be available online unless it can reasonably be shown that it is unrealistic to make such an expectation.
Scraping by: when the minimum is (or isn't) good enough
Normally, I expect subjects to have more extensive treatment than the bare minimum. However, exceptions can be made if a case can be demonstrated:
An article may be kept with a bare minimum of sourcing if and only if it can be demonstrated that it belongs to a significant category where being able to find any coverage in reliable sources at all is unusual.
An example of such a subject is Colemak: this has received coverage in only a small number of reliable sources and only just passes the source test. However, it belongs to a significant category -- alternatives to QWERTY -- where very few alternatives are more than proofs of concept only used by the designer, and only one other example -- the Dvorak Simplified Keyboard -- has had more extensive coverage.
Conversely, the Verdurian language fell into a category -- constructed languages -- where it could be demonstrated that extensive coverage in reliable secondary sources can reasonably be expected, so the fact that Google News returned no results meant that it had to go.
One particular peeve of mine is AfD debates that end in "keep" or "no consensus" although all the coverage in reliable sources only concerns a single event, and in some cases (e.g. Wrong Planet) the main subject of the event was actually not the same as the subject of the article. The rationale for these !votes is usually given as "once notable, always notable" -- which is a misrepresentation of what Wikipedia:Notability#Notability is not inherited actually says. While it is the case that once something is notable, it may continue to be considered notable as long as reliable sources are available, notability first has to be established:
- It must also be borne in mind, however, that Wikipedia is not a news source: it takes more than just a short burst of news reports about a single event or topic to constitute evidence of sufficient notability. The Wikimedia project Wikinews is appropriate for topics receiving a short burst of present news coverage.
For this reason, if an article's only claim to notability is that it featured in a single event, I will vote for its deletion, unless the event itself was significant and the article concerns the primary subject of the event concerned.