This would be a nice idea, but such a facility would have to be carefully executed through a combination of selectivity and management of processing constraints
Selectivity
Some of the checks are no-brainers (we’d like to know if there’s a Latex error rather than it simply not working) but some are in a gray area. For example:
I use the heading notation to represent edge types with colon as a metacharacter:
[[George#:parentof]]
in building a semantic network. The notation is benign to Obsidian; it just links to the whole note since the heading isn’t there. This allows me to do some analysis in Obsidian via graph search filters, and a lot of analysis outside of Obsidian. If we could not selectively turn on or off each check, all of my semantic network links would be flagged as an error.
Another good example:
I’ve noticed that a bunch of people in the community rely on backlinks as a form of tag or categorization:
[[recipe]] [[thyme]] [[basil]]
so that, in this example, the person can easily ask a question such as “Show me all of the recipes that can make use of basil,” and this would be the backlink list for the note “basil”. Yet, all of these notes are blank, and would be flagged as an error.
In a nutshell, some “errors” are not errors.
Management of processing constraints
The sum-total of these checks is very resource heavy, and the features could probably be separated into different contexts. For example, linting checks for “external” languages like Latex could be done when the note is visited. You can’t fix the problem until you visit the note, and since there’s no apparent problem until you do so, this seems to me to be the best time to check.
Edit: typo