Why Enterprise Software Sucks

You mentioned 'for HN' twice in the same sentence.

Your 2nd, 4th, and 5th sentence all try to assert the same thing - older comments are less relevant than newer comments - despite there being no evidence, anywhere, that this is the case. Your 3rd sentence is partly true, though there's some historical record aspect to HN comment threads that you're discounting.

Your second paragraph is a non sequitur. Sites that correlate points in time with timestamps are ubiquitous, and there's rarely a compelling case for using wildly different approaches to displaying same.

1. Windows search: It will sometimes not find what you are looking for, because it is not able to at least do a "string contains" check, if there are no other results found. How stupid is that? Glad I don't have to use Windows unless I am trying to solve a problem for someone who uses Windows.

2. Confluence wiki: Nope, sometimes you will simply not find that page, with some word in the title, which you input into the search field. Too stupid a software, to simple check for string contains. Oh and also sometimes your search does not work at all, because some service of Atlassian is offline.

3. Friend of mine did some work for an insurance company. They had a software to look up customers. You think you could easily find someone by looking for their first or last name? Nope, he had to enter an additional space after the first name, to be able to find the person, which of course was confusing.

So, I as a single person have coded in some projects search fields for applications, but some huge company with dozens or more developers cannot get a simple search field right, while I as a single person can write a search, that allows incremental searching and is functionally complete with "not", "or" and "and" operations on objects. Somehow I doubt the capabilities of the engineers, who wrote the logic for those search fields in the examples I mentioned.

There is also other things wrong with Confluence Wiki:

A while ago they claimed to now support Markdown in their editor. Well, it was a not even half-assed support of Markdown, as most of the things I tried to type in Markdown did not convert to the respective rendered elements. To this day, the editor they have just sucks at Markdown and at times behaves unpredictably. I despise having to put anything in the wiki, because of it's buggyness, sluggishness and bad support for Markdown. How difficult can it be to give me a simple editor and let me type Markdown or another sensible format?

My personal blog has a better Markdown parser, than Confluence Wiki and I did not even have to write it myself.

My self-written wiki uses reStructuredText files for pages, simply put in a directory tree, which gets me document internal linking and cross references to other pages of the wiki and the ability to simply put the wiki into a git repository and have everything version controlled. Oh btw. version control is also F*ed up in Confluence Wiki. Also I can simply "export" any page, by copying its file, while in Confluence Wiki, you cannot even get a proper export of a page. This is of course intentional, to keep customers stuck in their Wiki. No way, that they provide simple means of migrating away. No, better to hold the user captive in their BS software.

I have to agree with some other comment here, which mentioned, that the problem is, that the buyers are not the users. I would happily give my own wiki implementation to people at the job, but people, who make the decision on what is used think, that Confluence Wiki will be better than anything a single person can come up with.

I spent a few years doing indie game development, and one of this biggest lessons that stuck with me was how to think about a product as an experience. To me, game design was mostly that, but focused on making the experience fun, I guess. A huge amount of time and effort went into thinking about what the end user's experience would be like. By instinct, I tend to lean on thinking about the "player's" experience.