Contest For a Better Open-WRT Wireless Router GUI 217
Reader RoundSparrow sends word of a contest, with big cash prizes, being mounted by a commercial vender of open source Open-WRT routers. You have 10 months to come up with "the most impressive User Interface/Firmware for Ubiquiti's newly released open-source embedded wireless platform, the RouterStation." Entries are required to have open source licensing and will all be released. First prize is $160,000, with four runners-up receiving $10,000. RoundSparrow adds: "Could be built on top of existing X-WRT or LuCI OpenWRT web interfaces. OpenWRT Kamikaze 8.09 was just released. Now is perfect timing for OpenWRT to get some kick-ass interface and usability ideas. I'm not affiliated with the contest vendor."
Slashdot Editor: OpenWRT not Open-WRT (Score:5, Interesting)
You guys altered the name to Open-WRT :) Anyway, thanks for spreading the world on this and Kamikaze 8.09 release. the OpenWRT devs work hard.
Re:X-WRT? (Score:3, Interesting)
Re:the best routers I have seen (Score:1, Interesting)
Re:X-WRT? (Score:1, Interesting)
It doesn't have to be though.
I have a flakey internet connection to a very small ISP. I find the status pages on Tomato's default install to be very useful.
Graphs of network usage based on QoS classification helps you tell what is using your network at a glance. Theres a lot of work that could be put into areas like that that make your router not somethin you set and forget.
It also doesn't have to be a self contained gui (i assume, didn't read the contest rules).
I certainly think theres a market for pretty desktop apps to both configure and monitor your router, letting you see network utilization and what all is connected to your router.
Re:I have a suggestion ... (Score:4, Interesting)
Sure Tomato is nice - as long as you can still get hardware for it.
Sure OpenWrt may lack documentation or tries to do too much stuff at once, but right now it's the only fucking router distro that makes actual progress besides tuning the ui.
DD-Wrt, Tomato, CoovaAP etc. are all stuck with binary drivers that require an ancient kernel to operate them and more and more devices that are supported by this software are already end of life or will be soon.
The point of the challange this article is to make a user friendly interface for OpenWrt (besides the three projects already working on it). If Tomato satisfies your need - fine, but if you rely an modern hardware you're out of luck. And at some point it makes more sense to spend 25$ for a cheap-ass Draft-N device with gigabit instead of being stuck with ancient Broadcom gear for 60+$ just to install Tomato.
Re:X-WRT? (Score:3, Interesting)
I think we should encourage other companies to join in the contest. Best idea I have is solicit router companies to do $25,000 donations - and allow them to independently judge and reward their own winner.
That way maybe someone who didn't make the top place could get a chance at another income boost. Would supplement the interest in people fearful of not making 1st place.
Also note that a single person can enter more than ONE entry - so if they come up with different design cocepts - they don't have to choose.
Re:I have a suggestion ... (Score:3, Interesting)
Personally, I don't have enough computers at home to need ipv6.
I take it you're not developing software that needs to support IPv6 either.
Tomato is great, but I need/want IPv6 autoconfig on my home net. A spare Cisco 1700 handles this nicely. One of these days when I have some spare time, I'll get a v6 tunnel set up on it.
Re:It is amazing how negative everyone is about th (Score:3, Interesting)
The thing is, GUI design is a complex art, one that takes a long time to learn to do well, so its hard to be good both at visual interfaces and the often very complex code that they control.
You hit the nail on the head. It's an art, and that means that, when all is said and done, it's the guy with the eye who polishes the job. Programmers can follow all the user-interface design rules laid down in the multitude of books on the subject, but if they don't have the touch, what they'll come up with may be functional, but will still look like crap. It's as inevitable as the tide. Good coding can be an art as well ... it is by no means always thus, but some developers do carry their work into the realm of true art.
Look at the history of video game development. Early products were as you describe: often designed by single coders, or maybe a team of coders. That worked, because the hardware was too crude to allow an artist to do much with it, although some companies did hire animators. Blocky programmer-drawn graphics were acceptable because that was about the limit of the equipment. However, as the resolution, color space, and processing power of graphics systems improved, you began to see specialization occur in game development. Much as happened in motion pictures decades before, the evolving complexity of the products demanded an expanding team effort. Designers, coders, artists, animators, level designers, writers, play testers, quality control testers, a whole host of wildly different disciplines are now required to produce a single game.
It's no different in the world of Web development. You need a team, with people capable of handling overall design responsibility, documentation (something coders are notoriously bad at doing), back-end requirements, the GUI, and many others. It's easy for developers who have no talent for user interface work to dismiss such as unnecessary glitz. Understandable, but entirely wrong when talking about software that is marketed or used by the general public (like a Web site.) Furthermore, in the real world people (especially people who write big checks) are impressed by a polished, well-written GUI.
Personally, I've spent almost thirty years in this business, and I started out doing largely embedded-type stuff for the game industry, and eventually got into doing higher-level graphics and artwork. Then I got into manufacturing and control systems, and discovered that users like software that is attractive, not overly-complicated for the problem it purports to solve, and above all does the job. The years I had spent learning graphic techniques paid off handsomely during that time, since my competitors generally couldn't draw their way out of a shoebox.
In any event, I found that while I had successfully worn a number of hats as an application developer over the years, it was getting harder and harder to be a solo act. So, nowadays I'm not, I work with lots of other people, and I've found that my skill set is complementary to many of them. It works out well, if you have a good team.
Re:needs an easy way to edit firewall rules (Score:4, Interesting)
1) If your router is also a switch/hub, it can analyze the internal network traffic and learn computer names (if windows file sharing is enabled anyway).
2) It can tap the internet to look up stuff like mac addresses and other statistical traffic patterns to identify things like your Tivo or XBox.
3) You can invent an internal protocol that enables your household computers and devices to communicate to the router what the fuck they are. Odds are good you can use fancy crypto to make sure that the computers and devices can't lie if they get compromised.
4) Make a training mode that lets everything go through and when you are done, the router uses the wealth of statistical bullshit it collected in steps 1->3 to give the user a report outlining the househouse hold traffic.
5) The user can then "lock" the router and not let anything but what was configured in #4
6) If something odd happens, or the router detects new computers (say a laptop, etc). The magic protocol in step 2 would send some kind of alert to a computer, your email, your phone... something... basically saying "hey man, something changed... you might have to retrain me".
My idea, obviously, is a very crude outline. But you get the idea. Everything can be simplified if you focus in on exactly what the task at hand is and leave the rest of the bullshit out. In fact, I bet you can design the firewall configuration in such a way that the user never needs to see IP addresses or port numbers. All they see is friendly computer names (deduced from #1->#3) and descriptions of the traffic.
Nerds, obviously, wouldn't like this--instead wanting some geeky bullshit. But they can piss up a rope as far as I'm concerned. This is a mass market device intended for people who just want to feel secure that nobody is hacking their shit.
Re:Naw... (Score:3, Interesting)
Re:Naw... (Score:3, Interesting)
I guess the hard part is the "working and secure" bit then. I think you are generalizing too much.
Networking used to be horrible. Different protocols, different hardware. Lots of fads, each as much as a pain to set up as the last. No point and click email server wizards. Site-to-site was by modem, not VPN. When men were men and the cable was coax, that sort of thing.
It has gotten to the point now where you have a router, modem, AP, and switch in one, for $50 or something. Network speed is automatically negotiated. Everything full-duplex. That's if you're actually running cable. You don't even have to anymore. The protocol is all TCP/IP, so there's no need to worry about having matching hardware as well as software that can speak those protocols.
If a user can't set up a 2009-era consumer router/switch/AP/toaster, there's nothing wrong with that, but they need to pay someone to do it for them. Because it's about as magically automatic as it's gonna get.
People that come on here and complain about networking not being easy enough must not remember a time when your OS didn't do the connection troubleshooting for you. So yeah, in general we can always strive to make things simpler, but specifically, how is it going to get simpler than "plug the blue wire into the port on the left and turn your laptop's radio on?"
Re:needs an easy way to edit firewall rules (Score:1, Interesting)
You've never actually configured a real firewall before, have you.
First, the automation you propose would be a huge gaping security hole no matter which way you look at it. It would never be acceptable from a security standpoint and you might as well just tell the end user "You're behind a port-address translating firewall so your network is 'safe'."
Second, UPnP already does 90% of the crap you just "outlined." The mass-consumers you speak of will and already DO just turn on UPnP and go about their business. Then the router can use adaptive QOS for on-the-fly traffic shaping (like they already do)...
It seems to me that you're one of the "people who just want to feel secure that nobody is hacking their shit" consumers who's pretending to know what a firewall is.
Odds are that if somebody has a need to configure their firewall then they know what they are doing and all that BS you spouted off will actually get in the way. In fact, if you need to be doing anything other than poking holes through the firewall, one of these dorky little home-router linux distros isn't going to cut it. You'll need something with RAM and then you'll need a real manly firewall like pf, so you'll throw your pathetic little 32MB embedded router and its busybox command line out the window and replace it with a blazingly fast *BSD setup.
*disclaimer: I'm a competent Linux user who earned an RHCE back in the day, so you whiny "linux >> BSD" geeks can roll your excuses into a neat little cylinder and shove it down your throat. When you want to do some real networking in a production environment, do yourself a favor and hide your Linux server behind a *BSD firewall.
some suggestions .... (Score:4, Interesting)