Comment by moe | Hacker News Reader

Comment by moe | Hacker News Reader

Parent Full thread

Terrible.

There's a good idea buried somewhere in this, but they managed to completely miss the mark.

1. It needs an off-switch (globally and per repo). How did this feature get out the door without that switch? Does Zuckerberg manage product at github now?

2. The data is useless. What does it tell me that ednapiranha "contributed" to X during the last month, without any quantification? Nothing.

3. The data is useless. For most people the "popular repositories"-box shows 5 popular projects that they have forked...

4. The data is useless. When I want to see what a user (or myself) has contributed to then I want to see all of it, not the last 1 month.

Github appeals to the self-improvement, tomato-timer, fitbit crowd here. Those people will start blogging in no time about how "commit-streak tracking changed my life", from their barefoot standing desks. For everyone else these vanity-stats are worse than wasted space; they are visual noise that pretends to convey useful information but doesn't.

> Github appeals to the self-improvement, tomato-timer, fitbit crowd here.

What a bunch of bullshit. It's a tool. Thousands upon thousands of people use it daily because it helps them get stuff done.

If you don't happen to like the way it's presented, fine, but that's a personal preference, not a lifestyle choice.

I'm sure you also wanted to say something about the points I made but alas your tomato timer rang before you got to it, right?

No not really, I wanted to say something about the attitude.

I didn't think the points merited response, because it's self-evident that the blanket statement of 'the data is useless' simply isn't the case.

Anyway, let me be specific:

1. I can see why folks might want to be able to turn it on or off, but introducing visual inconsistency between different user profiles would feel pretty wonky to me.

2. Keeps the information dense. I don't need to know exactly what was committed in order for the information to be broadly useful.

3. Incorrect.

4. Recent = relevant.

That's not to say that I think it's perfect - for example, the 'current streak' will tend to 0 days for most users as noted by other comments, but overall I think it hangs together pretty well.

> Introducing visual inconsistency between different user profiles would feel pretty wonky to me

If GitHub made the default tab your repos instead (as it was before), all profiles would be consistent.

The default tab on projects is source—and projects can have optional wiki and issues tabs if they want. I think it should be the same with profiles' contributions tab.

You missed the word 'here'. The parent is talking about this one new feature, not GitHub as a whole.

GitHub is great, but I agree with moe. This seems like gamification, and GitHub's magic is in collaboration and communication. This turns me off a bit, as it smells like GitHub is trying to hook me rather than providing a tool I want to use.

For number 3, no they aren't for forked repos. See mine - https://github.com/deanperry - my popular repos aren't ones I've forked. I have quite a few forked repos which have been starred loads of times.

Hm, it seems I stand corrected, the list doesn't work like I thought it does.

However, I still don't understand what it's supposed to tell me. See for example (randomly chosen): https://github.com/peterb

According to Github his fork of "carrierwave" is his repo with "the most stars and watchers".

Well, it has 1 star. And I apparently can't see the number of watchers.

Neither can I see the list of commits/issues/pulls that peterb made to this repo, or which of his commits were merged (if any). Github shows me absolutely nothing that would let me gauge the quality or quantity of his contribution(s). Perhaps he's on the core-team of carrierwave, or he just corrected a typo in the README at some point - I'm none the wiser.

The key mistake here is to rank the contributions by the popularity of the project that was contributed to. Fix a typo in the README of Rails and that will likely remain your "top contribution" forever [unless I'm still misunderstanding how the list works]. It's the definition of vanity-stats; figures that can make you feel good but are ultimately misleading.

PS: If peterb is reading here: Sorry for using you as an example, I just clicked a random name on the list of rails-forks.

> The key mistake here is to rank the contributions by the popularity of the project that was contributed to.

That's not the case. Repositories Contributed To are ranked based on the impact an individual had on them (recency, # commits, # issues, etc). The popularity of the project has no effect on it.

Well, on just about every profile I tried there was Rails, JQuery or such front and center, on both sides ("Popular repos" and "Contributed to"). I still can't make out a pattern, it looks like a fairly random mix between the repos by the user and forked repos that they never contributed anything to.

Perhaps it'd become understandable if github displayed the actual metrics that lead to the ranking (commits#, issues, etc.).

Until then these new profiles remain sourceforgesque to me...

> Does Zuckerberg manage product at github now? I liked it.. No I loved it :D

I feel they're trying to address the general audience. It's good to configure the number of popular projects. but he popularity must be calculated automatically based on the user's contribution. But it's pretty good deal at least the for the sake of showing it at LinkedIn :)

Reply on news.ycombinator.com