Julian Knodt

On useless blog plosts

Like this one.

Useless blog posts are a great way to get people to think you’re a hotshot. You put stuff on the internet, you know what you’re talking about because you said it to all these people. Especially in the community of web-development, hot shit(like node, express back in its heyday) servers, data-processing and ML fads, and the like, there is an abundance of blog-posts on random topics, which delve about as deep as a puddle on the sidewalk after it rains. The strange thing is, these blogs posts provide very little informative material, yet they seem to be extremely useful to the community. Why is that the case? Primarily, because they’re backed by a much better indexing tool than tools like readthedocs or ctrl-F on the readme of a project: Google. To put it clearly, you can find blogs with good SEO backing incredibly easily thru Google. Many tools, even those that are popular, such as react or pytorch lack workable examples, and often provide very bare-bones documentation.

Are these blog posts bad? Should you care at all? Somewhat. The issue with these useless blog posts is two-fold. First, in some cases, people often produce such things in order to promote their own knowledge. Given the structure of hiring today, this is unsurprising, and a reasonable way to promote yourself. But, it does lead to blog posts where the creator has no reason to maintain it. Second, it’s hard for someone to find the distributed docs, as there may be many different places where thorough examples are being tested. The programmer’s job thus becomes a lot more of “hunting down how to do X” rather than learning from one repository all the basic tools and seeing a variety of capabilities.

How might this be remedied? Clearly, there’s many different desires in play. People want shit explained, and people want credit for explaining shit. One possible option would be to lean into open-sourcing, by making it easier for people to add to documentation, and also give credit for contributions(possibly even by name). This leaves the maintenance up to the project maintainer so it won’t ever go out of date. Of course, then the project maintainer has to make it very clear how to add docs, which is work in itself. Some big projects make it incredibly easy to add documentation, such as python, elixir, or rust doctests, and also provide good methods for generating documentation from comments. Further exploration of these ideas, as well as having the intent of producing useful documentation and letting others contribute will probably help replace the duct-tape that is blogs.