A Bitter Guide to Open Source | Hacker News Reader

A Bitter Guide to Open Source

medium.com

90 pointsedward15 comments

That's the great dilemma. Either you write the actual fkin' code, or you do roundabouts all around "project management" tasks, which in itself can be seen both as a needed means for a project's success, and as the perfect procrastination tool.

If you have 100 hours to write something and do ALL that is noted here you'll end up not shipping anything at all. I'm currently the main developer and maintainer of the Kurento project [1], and I've been working hard on all these accessory but important details because my vision coincides with that of the article's author, but at the same time they feel like a time sink which distracts from doing actual work on issues and bugs.

[1]: https://www.kurento.org

The main thing all these guys do is to ship the most trivial thing that will work (is better than competitors) but leave it open for improvements.

Tests are written sometimes second, not first. Specifically regression only testing is pretty good and less intensive than full testing everything. But to know about regressions and problems you need reports and to do that, the software has to be shipped, in use and the report channels have to work well.

This "guide to OSS" doesn't tell you anywhere, but talks about creating a JavaScript library, resulting in advices like "your OSS project needs to have valid package.json", which reads very strangely. Could use some "for example"s here and there.

This is a very strange article from a universe where the only languages available are JS and TS.

this "Medium" style of writing kills me. Lots of words, little substance.

It's fine if you don't like his style, but the "little substance" part is hardly true. I contribute to OSS on a very regular basis and I was left with four or five solid takeaways after reading the article. The author even admits he isn't good at writing but a lot of people asked him to write the article. I'm glad he did.

Agreed. I'm about to release a fairly large JavaScript service as open-source (the Sidekiq to Kue's Resque, as it is), and I got a lot out of it.

gratuitous obscenity, speaking from on high, self-declared greatness. not saying the guy isn’t good at what he does and has done, but god he’s insufferable.

Agreed, but any writing is better than none, if new writers have good examples to follow.

Do you think this is specific to Medium? Why didn't it happen with Wordpress?

> Agreed, but any writing is better than none

This is not true at all. There are many examples of writing that is worse than nothing. Just look at many examples of clickbaity material [0], that would make the world better to not exist. This is also disproven by the truly dangerous stuff by extreme examples like Hitler, etc.

This doesn’t apply to OP’s post, but I don’t think the philosophy that any content is better than none.

I do think that personal writing is always helpful. Practice is really useful for writers. But most should stay private or at least unpromoted. When garbage is promoted actively to readers and then criticisms are shrugged off with “its just version 1.0” with no future edits or updates, we have a worse discourse. And I think this means we have a worse outcome than if no initial, shitty version is put out.

I don’t think this is to Medium per se, but their format promotes clickbait, low engagement articles with their applause and payment scheme.

[0] https://www.buzzfeed.com/sydrobinson1/times-pets-completely-...

I agree, I wish people would be more discerning and write less, tweet less, etc. and just wait until they have something really interesting to say. Then you have these marketing experts telling people to post at least once a day to keep their audiences engaged. Yuck...

I guess the world will always run by extroverts.

What criteria do you recommend for qualifying the unpaid writing of an individual, as ready for online publication?

I don’t think I have a criteria. It’s pretty hard first off, and even if I could, mine wouldn’t be very useful for others.

Probably the most useful would be the author honestly assessing their content. And second would be a good set of editors. Then you could trust the publication.

> Repo Prerequisites

> README.md

> CONTRIBUTING.md

> LICENSE.txt

> A valid, filled out package.json

Hmmmm...one of those is not like the others.

> Some a$#h@!& is gonna come in and whine about something

This is very true just like in our society (good or bad intention it may be). Nice article Ken Wheeler by the way.

How much time do you normally spend writing OSS?

Reply on news.ycombinator.com