I got a bad case of vicarious cognitive dissonance from reading this article. At first I thought it was purely me, but the section on his reasoning behind limiting the option space for how long the clipboard history should be showed me, that he basically just doesn't like configuring stuff and will come up with any reason to avoid implementing user configurable options. To clarify: He clearly acknowledges that there are various good reasons for people wanting different behaviour, which generally contradicts his goal of arguing for not shipping preferences. This contradiction at the root of the article never gets resolved, which is what creates the cognitive dissonance, and which makes it almost comical when he arbitrarily limits the clipboard history length to be configurable as a value between 50 and 200. This limit cannot be explained through any of his arguments (like that preferences increase the space of possible states that need to be tested), and so I conclude that it is actually all about the vibes for him - he just doesn't like preferences. And that's ok.
Yeah I definitely noticed. I think it would be a good idea though to put a disclaimer for what already works and what doesn't. The website makes it appear like a finished product (props for it btw), and the goals of your app are quiet interesting. I will check back later once you got to it to a state that is more or less what the website promises
Keep at it, it looks really cool what you are trying here
I had copilot take a look through your repository, I will link its finding at the bottom. as a tl;dr for anyone who got interested by the sales pitch of the website:
"DevMem is a moderately clever idea — tying AI-generated documentation updates to git commits is a legitimate problem worth solving — expressed in a Go codebase that is not yet production-ready, does not compile as committed, contains a material security misrepresentation in its documentation, and ships approximately one-quarter of the features its website describes.
...
The marketing website is polished to a degree inversely proportional to the maturity of the code behind it. The claims about credential security are not merely aspirational — they describe a specific technical mechanism ("macOS Keychain", "system credential store") that simply does not exist in the implementation. Users who trust that claim may store secrets in a file that is one cat command away from exposure."
"The black box present in modern technologies such as mobile phones abstracts away what’s going on behind the interface and don’t require any process to interact with it"
Then they just added an artificial interaction process and somehow felt confident to having solved the problem.
I only get names starting with A, could not get it to do otherwise.
Also, is "popularity" of the name implicitly for the US..? I am not from the US and I doubt the results are as uncommon as my filter setting wants them to be.
Ideas:
- just add various optional filters that you can come up with, length, sillables, origin, first letter, must have letters etc etc
- allow the option to not select gender, since some people may look for names for children before pregnancy
- add some more info on the dataset, and maybe even let the users look at all the namesunfiltered. I was wondering how complete it is as well..?