behsoh
reading

Judging a Source by Its Distance From the Subject

Learners depend on tutorials, forum threads, and reviews written by strangers. A simple question, how close is this source to the thing it describes, sorts most of them.

Everyone learns from strangers

Almost nobody learns to program from one approved textbook. You learn from a search result, a forum answer, a blog post, a video, and the documentation, all stitched together. That is normal and it works, but it means you are constantly deciding whom to believe, usually in a few seconds and without noticing that you are doing it. It is worth making the decision a conscious one.

Distance is the first thing to measure

Historians have a practice called source criticism: before using a document, ask who produced it, when, and for what purpose. For a learner, the most useful single question is about distance. How close is the author to the subject?

Closeness cuts both ways. A source close to its subject knows the details: the maintainers of a library know exactly how it behaves. But a close source also has something at stake. A project's own site is the best place to learn how to use the project and a poor place to learn whether you should. A distant source knows less detail and has less reason to flatter. Neither is simply better. They answer different questions.

Interest is not the same as dishonesty

A conflict of interest exists when someone's position gives them a reason to prefer one conclusion, whatever the evidence says. Having such an interest does not make a source wrong, and it does not make the author a liar. It means that their praise carries less weight than their facts. When a vendor says its product is the easiest to use, that tells you little. When the same vendor tells you which settings the product requires, that is probably accurate, because getting it wrong would only create trouble for them.

So read interested sources for what they are good at: procedures, conditions, exact wording. Look elsewhere for judgement.

How community sites arrange themselves

An internet forum mixes two kinds of content, and a well-run one keeps them apart. There are guides: edited, stable pages that describe how to do something. And there are discussion threads: dated, personal, and often contradictory. A guide tells you the intended way. A thread tells you what happened to one person on one day. Both are useful, and confusing one for the other is a common mistake.

The pattern is not special to programming. For a plain example from a wholly unrelated field, take a Traditional Chinese review and forum site built around one gaming brand (夜色娛樂城論壇). It keeps play guides, an outline of a withdrawal process, and roundups of activities on one side, and discussion topics on the other. The brand's abbreviation is part of the site's own domain, and that is the first clue about its vantage point: it is written from nearby, not from a distance. That is not a reason to dismiss it, and it is not a reason to rely on it alone. It is a reason to read it as one source among several, strong on how things are laid out and weak as an independent verdict. The same reading applies to a programming forum hosted by the company whose framework it discusses.

Signals worth checking

  • Who runs it? Look at the address, the about page, and the footer.
  • When was it written? An undated tutorial about a fast-moving tool is a gamble.
  • Does it show its work? Code you can run beats claims you must accept.
  • Does it admit limits? Sources that mention drawbacks tend to be more careful in general.
  • Do others agree? A claim found in one place is a lead. Found in several independent places, it is closer to knowledge.

Independent means independent

Be careful with that last signal. Five pages that all copied the same original are one source, not five. Before counting agreement, check whether the pages share wording, examples, or an obvious common origin. Real confirmation comes from sources that reached the same conclusion separately: the documentation, a test you ran yourself, and an unconnected author.

The check you can always run

Programming has an advantage that most subjects lack: many claims can be tested in a minute. If a forum answer says a function behaves a certain way, write three lines and find out. A result you observed yourself is the nearest source there is, and it has no interest in the outcome. Use it whenever the claim is cheap to test.

Keep the source with the note

When you copy a technique into your own notes, record where it came from and when. Months later, when it stops working, you will want to know whether it came from the project itself or from a comment by someone guessing. That one line of provenance turns a pile of snippets into something you can reason about.

Working rule: trust close sources for how, distant sources for whether, and your own test over both.

Last revised 2026-11-20 · Wendell Arkwright