Hacker Newsnew | past | comments | ask | show | jobs | submit | brebory's commentslogin

Next thing you know, they'll start intercepting Command-W's.


I suspect it's because the original app id and secret key are public on the github repo.[1]

I bet they're registering a new application and secret right now, hence the invalid API key error.

[1]: https://github.com/kylemcdonald/socialroulette.net/blob/ba32...


Either that or they're starting to exceed usage. I'm sure there's not some magic mass delete routes for facebook's api. I have to think they're going to be hitting some kind of limit here pretty soon.


I'm running into a bug on the second lesson. Unless there's something subtle I'm not noticing.

http://i.imgur.com/n2WcCzY.png


Oh that had a little trick, the last button was actually a "large alert radius button".

Sorry for that! Just pushed a fix to make it more fair, no longer a radius.


Funny how fast that was, I just did it and wondered what was wrong with his.


Ok, cool! Nice work on this!


It helps to remember that: 1. #RRGGBB 2. B + G = Cyan 3. R + B = Magenta 4. R + G = Yellow 5. Equal amounts of all three channels = Gray 6. #000000 = Black 7. #FFFFFF = White

Think of the rgb color space as a cube, with each of Black, White, Red, Green, Blue, Cyan, Magenta, Yellow on the vertices. Then you define any hex color as point somewhere in this cube and it's pretty easy to imagine what it'll look like.


I'm not defending layervault's overuse of the concept, but one of the (many) ways to achieve typographic hierarchy is to map the contrast of an element to the relative importance of that element. Look at it this way: Without even reading it, you know that the gray text is supplemental or auxiliary information, because it has this visual cue.

Of course, sacrificing readability to achieve this definitely isn't the best MO.


In this situation, it's probably best to differentiate between the two with color—but you're right, it's silly to lose the already-established visual cues.


Colour is a terrible visual cue because colour perception is heavily relative to the viewer. Strong visual signposting is more apt (hence why playing with underlining should not be abused).


They sum up their reasoning pretty succinctly here: https://yourkarma.com/help/75-why-do-i-have-to-login-with-fa...


Because it's so hard to make a fake Facebook account. If someone actually wants to cause trouble, requiring Facebook won't do anything except slow them down for about 20 seconds. I see this "reason" everywhere, but it's total garbage.


Seriously. That's asinine.


"Vomitous" is the word that came to my mind.


You actually can't name your hotspot currently. It automatically names itself [name]'s Karma, pulled from your facebook info.

Their reasoning - https://yourkarma.com/help/5-why-can-t-i-change-the-name-of-...


This makes me want to create an network of access point named "x Karma" that installs viruses and malware. "Viral marketing," I'll call it.


Exactly. The SSID proves precisely nothing. It is sad to see them to claim it "keeps people honest."

This "Karma FirstName" scheme is going to alienate some potential customers. Also, I suspect the added advertising boost from letting your users be clever would offset losses. But that is a simple empirical question. It wouldn't be hard at all. How many new signups per hotspot do you get from hotspots with custom SSIDs versus stock SSIDs?

Looking forward to seeing what security measures they do have in place to prevent someone from spoofing. At an abstract level, the device is going to need to contain some public-key (or something similar) that it sends somewhere on the Internet which is then verified with a private key. Then it's got to tell the user whether or not it's a real Karma hotspot or there's trouble.

There will always be a risk. But it could at least be managed to the point where it is one of the lesser Internet risks.

Remember this isn't really a new security hole for users of public wifi. Just setting up your own Starbuck's access point (dlink) and loading it with malware and snooping would do the same thing.

I don't really know how SSL works over a compromised access point. So maybe there are safeguards. But I tend to assume if the access point is compromised, all my base are belong to them. Or at least the data I send.


The difference between bootstrap and a CMS is that bootstrap doesn't mess with any wysiwyg interface bullshit. It's just a pre-fabricated way to architect your html, css, and javascript, and the default styling makes things look very professional.


I think you have that absolutely backwards. The difference between Bootstrap and a CMS is that a CMS is an application. Bootstrap is purely a sort of standard UI toolkit for the web.


A pre-requisite for being a CMS is managing content (Content Management System). Bootstrap has no content-managing facilities; It is purely a presentational (CSS, some JS) library with some well thought out defaults.

ie: you could still use Bootstrap to style a WordPress site.


Right, a CMS lets you implement an architecture for HTML and CSS, e.g. Bootstrap. I don't understand how what I said was "absolutely backwards".


I think "backwards" is the wrong word. I'd say "uninformed" would be closer to the point.

CMSs manage your content usually through a database, and most of them provide a way to display that content in a structured format. See Wordpress and Drupal. WYSIWYG is not a necessary part of a CMS.

Bootstrap is a (mostly) css library to help provide a better front end development experience for a web site/app. It can be used with the front-facing portion of a CMS or any other kind of site.


    Right, a CMS lets you implement an architecture for HTML and CSS, e.g. Bootstrap.
Huh?


Any chance of seeing a responsive build of the site down the line?


100% Chance.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: