During the weekend I spent some time (almost the whole Sunday, actually) polishing the brag. The One Big Change Everybody Was Waiting For is obviously this – we now have RSS and Atom feeds! Loads of smaller or less visible ones got through, too.

For starters, now the whole layout actually behaves properly when on small screens. The min-height CSS properties are properly set and as I moved from position:fixed-based layout (which was much harder than expected), those now do their work.

We’re still bound to Firefox and Chrome/Chromium, unfortunately, but that’s the price I have to pay for the oh-so-cool horizontal layout (curtesy of CSS3 Multicol). And gradients. Let’s not forget gradients.

And speaking of gradients – have you noticed how there’s no official support for CSS3 Transition with gradients? Well, don’t move your mouse over the header area above, or you might get surprised…

Also added today: single-node layout. Yes, that really needed additional work, as the horizontal layout was not easily tamed for that. Well, it’s done.

You might also notice some CSS3 Transition effects (aside from the magical gradient transition on top) with links and images. CSS3 is fun!

Planned for inclusion sometime this year:

• more traditional layout for more traditional browsers;
• per-tag/per-category feeds (yup, I have tags/categories here, they’re simply not exposed yet);
• Identi.ca and maybe Twitter integration (the backend’s already there);
• comments.

To all the newcomers here – this is alpha version of my brag, and things will break. Please use Firefox or Chrome/Chromium to have the best experience; Opera will work too, but will not be as nice. Rekonq, Konqueror, other KHTML/Webkit-based browsers – newer version probably needed. To all the Internet Explorer and Safari users out there – please get a real browser, it’s free and it’s fun.

During the weekend I was responsible for the network and tech support at Startup Weekend Warsaw co-organized by my lab at Warsaw University of Technology. I cannot say I did a stellar job, but then again, I cannot say I didn’t try and do my best to keep the network alive and kicking. So…

The Scenery¶

100+ participants. Every single one of them with at least one laptop. Most - with some mobile, WiFi-enabled device in their pockets, too. Many with some third device - tablet, second mobile, etc. All in all, probably more than 200 devices trying to connect to the Internet. Quite a bunch.

The Setup¶

2 Access Points (Linksys WRT-54GL; one with vanilla Linksys software, the other with DD-RT), two 8-port gigabit switches, loads of cable, and the Faculty infrastructure. And it’s always about the infrastructure, ain’t it.

The Plan¶

The Plan was simple enough and looked quite well:

• set-up two different WiFi networks, on different, separated channels;
• get as many people on the wired network as possible.
• ???
• PROFIT!!

As always, the “???” turned out to be the crux of the whole thing.

The Unexpected¶

As per Faculty Network Policy, many outgoing ports were blocked. Obviously, in that NAT-ed network, all the incoming ports were filtered. I decided to set-up a tunnel (an SSH-based VPN) for those few of our users that would need some “exotic” ports (like, oh you know, 25/tcp if they would fancy sending an e-mail). That was supposed to be far from mission-critical and just a courtesy towards the technically-inclined guys and gals in the room – so, basically, 90% of them.

However, it became uber-critical as soon as it turned out the (important) live audio/video stream that was supposed to allow more people to participate on-line actually uses some of the blocked ports. Whoopsie! The quick-and-dirty solution became a very important piece of duct-tape.

Fun with Streams¶

And there were loads people watching this stream too! Problem was, many of them were in the very room the stream was transmitted. Now, sending a video stream was enough of a network hog to cause minor hiccups; when people started watching it within the same network, basically all hell broke loose…

Oh, and let’s not forget the great job Skype was doing to help our network tank even deeper. Yay for that.

The Fa(c)ulty Infrastructure¶

To be honest we had some real faith in the Faculty’s solid backbone. And with good reasons too. It is a solid backbone, so why shouldn’t we? Ah, faith, you are a funny thing. There comes a moment that reality catches up and, say, the Faculty’s DHCP server goes down. Good for us we had a nice Ubuntu box (yes, the one with the SSH tunnel/VPN running). 5mins with apt-get and dnsmasq.conf and we were back on-track to the next failure in the string.

WiFi Mavericks¶

Well, obviously, the wireless quickly started getting quirky. As in, not working properly. Or at all for that matter. When suddenly 100 devices try to connect to a single AP in a matter of minutes, the AP will go down in a matter of those same minutes. Vicious circle.

So people started using 3G connections, which would not be that bad, as it would lessen the traffic on the poor battered APs, right? If only those were used via Bluetooth or USB. But guess what? Setting up your own ad-hoc WiFi mininetwork is sooo coool, right? Hence, suddenly, we had about 15 different ad-hoc networks interfering with the two Startup Weekend official WiFi nets. Guess what, that was not helping.

How it all played out?¶

or “putting the fires out”

To be honest - not well. There were simply too many points of failure. Too many fires. Often times the APs got in some strange mode in which connectios already established work passably, but no new devices were able to connect. Should we reboot such an AP to get the new devices on-line, or just go with the flow and let the already connected use the network without interruptions? Damned if you don’t, damned if you do.

Lessons Learned¶

So, “mistakes have been made”, moving on with the knowledge. In particular:

• QoS! next time each and every single user will get a dedicated, albeit small, bandwidth channel.
• less security can buy quite a nice amount of reliability; seriously, we did not need WPA here, we could have gone with WEP – or no security at all.
• in-house as much as you can: get your own network segment, your own DHCP server, etc. – this way at least you are in control if something goes awry; and believe me, it will.

We got a few things right, a bit more on that further on down.

Blaming and Name-calling¶

We already know I was responsible for WiFi, but far from being able to do it all by myself. Special thanks for all the hard work go to Piorek and Karolina. Piorek was helping me all the time with tech stuff (and doing a great job); Karol was the bureau and chancellary, making everything go as smoothly as possible.

Many thanks to Kamila, Konrad and Krzysiek for the great atmosphere at the conference. You guys should have been pissed a few times, but weren’t, and that went very far in helping us deal with the Wireless Notworking rather than interpersonal stuff. Seriously, to all the conference organizers out there – take heed, as this may well save your WiFi!

Last but not least - [@zstanska](http://twitter.com/zstanska) and [@mpaluchowski](http://twitter.com/mpaluchowski), doing the social media and video streaming, were usually the first to nag inform us about any problems. And in style! Also made fun of us on twitter. Really, you guys could have used a better service.

Finally, heartfelt thanks to all the people at and around Startup Warsaw.

Icing on the cake¶

Microsoft reps doing a presentation on a Linux box (that was our presentation box, simple as that)… I am not even sure they knew that, to be honest. Ah well, fun anyway. If and when I get the video, I will drop it here.

Actually, quite a few people had Linux on their lappys. Interesting times.

Follow Up¶

Today Shot sent a great article about making conference WiFi work. Better late then never – but hey, turned out we actually did many, many things right! I’ll try writing a follow-up bragpost on that later, when I sleep a bit.

This is exactly I would like to tell millions of Skype users today, when Microsoft bought it. This is a perfect example why giving away control over entire communication in a given area to a single entity is not a good idea.

So, what could happen now? Let’s see:

• Skype client for Linux development might halt;
• same goes for the Mac client.

But, because of all the contacts that Skype users only have accessible via Skype, it’ll be hard for them to move to any other solution. In other words, damned if you do, damned if you don’t… Suddenly it turns out that a situation where all three elements – the protocol, client and server software – are in one hands is a very bad one.

How it could have been¶

A perfect example of a better world, or how it could and should have been, is e-mail. Because the protocol is an open standard, because anybody can set-up their own server – the situation is much, much better:

• there are multiple providers;
• there are multiple clients to choose from.

That means that regardless of the luck or ideas of a single provider or vendor of a single client, we can always go elsewhere. We can choose the client software, regardless of the service provider we use; we can change the service provider without changing the familiar client software we have grown accustomed to. We have a choice.

Had we all been using a standardized, open VoIP protocol, like SIP or Jingle, we would simply change the provider – without losing our contacts, having to learn new client software nor wondering whether our friends and family chooses the same new solution we will.

Who’s next?¶

Facebook, of course. Had I been using it, I would seriously start considering Diaspora. It’s alpha, a lot of to-do’s still down the road. But it works and is attracting more and more users. And most important of it all it works according to those few simple rules:

• standardized, open protocol;
• anybody can set-up a server;
• anybody can use any client they choose to.

…or release early release often

Well hello there! Yes, yes, I have a brag. Self made. As in – layout, engine, graphics (that would only be the logo for the moment), all made by myself. See, I’m bragging already!

Let’s treat that as alphabeta version. Loads of ToDos still, e.g.:

• a much better logo – current one is very preliminary, basically only for testing;
• graceful degradation for the so-called alternative browsers;
• RSS/Atom feeds;
• maybe a better slogan;
• interface translation;
• IoC config caching;
• nice urls;
• CSS cleanups and re-writing it in SCSS or LessCss;
• admin area (pasting entries directly into the database is cool and all, but not that convenient);
• comments;
• English version;
• paging;
• single entry template;
• font used – through @font-face;
• lots. and. lots. of. layout. enhancements (partially done).

Yup, the whole thing: the engine (used also on the website of a Lab I work at), layout, styles, graphics – everything will be released on free/open licenses. Working on it.

If you think I would need anything else here, you know the address – write. I’ll be listening.

Oh right. One last thing. Why “brag” not “blog”? Well, “blog” comes from “weblog”, an on-line log or diary. That doesn’t play well with me (a diary is a private thing, kept locked and published after the author’s passing, if at all). Here I’m going to, wait for it, brag and litter the net with comments on IT, philosophy, life, the universe and everything (42!). Hence – “brag”.