Hydronium Hydroxide

Useful Pages

On Sourcing (mainly for AFDs)

If something is commonly written in a native form (including non-Latin scripts and Latin scripts with diacritics) then that needs to be searched for.
The notes at WP:INDAFD broadly apply not just to India, but to a lot of countries.
Given an English form of an Arabic word -- or even a word in Arabic script -- a non-speaker will find the searching painful (Semitic roots, etc). There can be multiple transliterations of even proper nouns (for instance see Muhammad (name)).
If sources are not apparent due to large number of false hits, try limiting the search to the country with "site=*.xx" where XX is the two-letter country code such as FR, IN, AU, CZ, etc)
For publically-listed companies, the company code should be searched for with the term "analyst"; this will tend to reveal reports.
BLP Obituary Notability Litmus Test: If the article subject were to die today, would they be likely to receive obituaries in non-local general or specialist sources?

Heresies and Articles of Faith

WP:5P1 should be "Wikipedia is a hyperlinked encyclopedia. This distinction means that more weight should be given to preserving marginal articles which (integrally) interlink with other articles than ones which are functionally dead ends and which can be consolidated (WP:NOPAGE) or deleted.
Those who find sources demonstrating sufficient notability during the AFD are and should be under no obligation to actually add them to the article (though it's nice when they do). Corollary: Those who nominated or voted to Delete without finding said sources should feel free to feel obligated to add said sources as penance.
Once AFD consensus is clear, additional !votes should only be made if they actually add value to the discussion. There is no value (other than running up stats) to adding an additional simple per-nom/not-notable Delete !vote to a nom with 2-3 properly-reasoned Deletes already and no counterargument. Similarly 3-4 properly-reasoned Keeps with only the nominator wanting different is sufficiently clear.
The problem with !voting Merge at AFD is that someone has to do the work of merging. Where a merge is the obviously right outcome, this can be -- and is best -- done while the AFD is going on (since the resulting text should be at the target anyway, and it's easier to !vote Merge when the outcome can be examined).
For articles of marginal notability, the implicit question at AFD should be "How can the information in this article be retained?". It may be that ultimately it shouldn't be. It may be that there's a future option for incorporation but the article doesn't exist, and so deletion is currently the right choice. But if there is a reasonable alternative to deletion -- WP:ATD -- then a Delete !vote is incorrect, and there are often enough AFDs where this should be immediately obvious (eg: non-notable song/album from a notable band) but that accumulate Deletes.
Where there is a class of articles, precedent matters even if it's bad, and Weak Keeps/Deletes on this basis are reasonable, and even preferable. RFCs (eg WP:SCHOOLRFC, WP:NSPORTS2022) are a better/cleaner way of standardising than chaotically trying to chip away using individual AFDs which will vary depending on who participates.
(For new articles) AFD should basically not exist, and WP:NPP and WP:AFC should not exist in their current form. For an article to enter mainspace, WP:THREE (or thereabouts -- NCORP is FOUR) should be clearly met. THREE is easy to confirm/debate. THIRTYTHREE is not, which is one of the reasons why near the bottom of the NPP queue there can be an agglomeration of articles of marginal notability with an excess of refs. Notability should be evident at the stub phase. Want to create an article on wikipedia and don't have the reputation (ie, a PERM - possibly below autopatrolled) where you can be reasonably expected to create in mainspace with sufficient RS? Identify your Credible claims of significance and provide only a limited number of refs clearly meeting SIGCOV (or that shows that NACADEMIC, etc, is met).