Comment by brudgers | Hacker News Reader

Comment by brudgers | Hacker News Reader

Parent Full thread

Looking at the "Show HN" associated with your HN user name, one thing that pops out is that the post doesn't create much context for providing feedback. That is, there is nothing explaining what you are trying to achieve. Since I am clearly not in the target demographic, there's nothing for me really to evaluate the site against...I don't visit other street-fashion-rating websites or know why the woman on the beach is worth -422 for whatever it is that I'm picking.

It does something, but I don't know what let alone how well unless you tell me what it is supposed to do and why it is supposed to do it. I mean, I know there's a design vector under which I am simply not supposed to "get it," but I have no basis for evaluating it against a user story of someone who is supposed to "get it."

As others have said, the best feedback would be from people who are supposed to "get it". Absent that though, feedback requires an explanation...a blog or a comment or a link at the bottom of the page.

Good luck.

Thanks for digging deep. However, this question was not in relation to my previous Show HN post (damn internet :)).

Since you did bring it up, I like to think that I'm aware of the reason for it not getting any attention. Long version: I got fixated on engineering, while forgetting about the core UX.

Case in point:

1. the score you mention begins with a node service scraping a random page from lookbook, which contains a list of photos and their 'hypes' (likes);

2. then all the scores get normalized, where 1000 would be given to the highest hype score, all other being (hype*1000)/max_hype;

3. finally the sign of the score is based on whether you correctly guessed that the looks hype is more ('Fashn') or less ('No Fashn') than the average, non-normalized hype for a given page, and the nominal value of it is the difference between the normalized average and a given photo's score.

I guess I didn't bother explaining this to the user, since the initial idea was to compare your taste to that of the lookbook community (read 'objectively fashionable'). However in approach above, the score for a photo is determined relative to the other photos on the page, but not other pages. Since relativity is unavoidable with the set logic, the user must be presented with a design that allows relative comparison. This, in turn, brakes the current minimalist design, which was one of the main features.

This Buridan's ass state is where I stopped development, and decided to show it as is. HN wasn't the first place, and all the feedback resolved around a broken (not well explained?) core mechanic. Meanwhile, the restless mind queued up a number of ingenious ideas, and, not being socially pressed for time nor resources, I decided to move on.

Everything on the damn internet is forever...fortunately it helps with discoverability.

Anyway, where I was going is that everything that is in your comment here provides context and most importantly is interesting to the people who are likely to read it:

  1. It's technical
  2. It's on HN

Making a website that gets feedback on a website review website is...well it's like those big books of 1000 new house plans. The plans are optimized for selling from a big book of house plans, not for the buyer's lot or lifestyle or around their love of amphibians and British motorcycles. And everyone has their own amphibians-and-british-motorcycles crush and there's no way to know it without talking to them.

So what I'm saying is that the hard part is figuring out what someone cares about. I mean even a website about designing websites around set logic and relativism has at least one person who cares...so does british-motorcycles-and-amphibians because I went to Junior High School with her.

The challenge is that it is a lot easier to find people, like me, to comment on a british-motorcycles-and-amphibians website than it is to build one that my friend will care about.

Reply on news.ycombinator.com