This essay contains the advice or opinions of one or more Wikipedia contributors. Essays may represent widespread norms or minority viewpoints. Consider these views with discretion.

It is a popularly-held belief that stubs can be defined precisely by the size of the article involved. Once an article hits a certain number of bytes, it is automatically deemed to be no longer a stub. Below that size, it needs a stub template.

This, sadly, is not the case. If it were, it would be possible to totally automate stubbing and stub-sorting, freeing up many editor-hours of work on Wikipedia.

There are several reasons why a long page may be a stub, and similarly there are several reasons why a short page is not a stub. The two simplest of these rules are the following:

Only articles can be stubs. Categories, disambiguation pages, infoboxes, user pages, talk pages... all of these have their own templates that can be used if they are in some way deficient. None of them can ever be stubs.

Stubs have very little text. Text here refers only to written sentences of information - it excludes images, infoboxes, navigation templates, lists of examples, external links, and any of the other items which may be found on an article, all of which are ostensibly there for the purposes of supporting the actual text of the article. Thus, what may seem to be a long article may in fact be a one-paragraph stub with "peripheral add-ons".

[edit] The "Croughton-London" rule

Both of these two rules might well be handled by a bot to some extent; the third — and in some ways most important — rule for deciding what is or is not a stub cannot be. That is the Croughton-London rule.

Consider an article that is slightly longer than your standard one-paragraph stub. It has maybe three short paragraphs, amounting to less than a screenful of information on your laptop. Is it, or is it not, a stub? The answer will often be given not so much by the article itself as by the topic of the article.

A small article on a relatively small or insignificant subject is far less likely to be considered a stub than the same sized article on a far larger or more important topic. There is simply likely to be far less that is noteworthy, say, about a small English village than about the nation's capital city. The article on Croughton, Northamptonshire at one time ran to only a handful of short paragraphs, but it was sufficient to regard it as a fairly comprehensive article about the village and therefore not a stub.

If the article on London were the same size as this article, it would be another matter entirely. London is a national capital, has been of major importance for at least two thousand years, and was for some of its history the biggest and arguably most important city on the planet. An article of only three or four short paragraphs on London would be a stub of embarrassing proportions.

Thus, a stub is a stub not just by dint of its length, even taking into consideration whether it is an article and how much of that length is text. It also has to be judged in terms of the relative importance of the subject of the article. And that, sadly, is both an arbitrary process and one that cannot be done by bot alone.