The sanoma Amsterdam branch moved to Hoofddorp last weekend. From now on we will work at the beautiful headquarters.
At home in The Netherlands, I still use an Apple Airport Express bought in South Korea. Last week, I was tinkering around in the Airport Utility settings and realized I could crank up the AE to 5 GHz and Wi-Fi link speed went from 54 Mbit/s to 300 Mbit/s. I noticed the difference immediately. However, one day later I couldn’t get back on my network and I saw this message in my Snow Leopard console:
802.11d country code set to 'NL'. Supported channels 1 2 3 4 5 6 7 8 9 10 11 12 13 36 40 44 48 52 56 60 64 100 104 108 112 116 120 124 128 132 136 140
Oh dear, the channel which I was using (161) was blocked due to 802.11d regulation, which I wasn’t able to override. I was back at the same old 54 Mbit/s but it felt even slower than before. The sad part is, it worked for a little while, so I got a taste of what was taken from me. I was able to repeat the behaviour: Setup with Airport Utility and get high speeds until reboot. It seemed like a software problem rather than hardware.
Starting with the 802.11d thing, I got tired of all limitations within OSX and tried to find a solution outside the Apple Operating System. Using Re-find I installed ubuntu 13.04 Raring Ringtail on my Macbook Pro 9,1, installing is a process that has become easier over the years. Next up was installing WiFi drivers and seeing if the 5 GHz network would be in my reach.
I’ll save you some time: I found a few howto’s online recommending a broadcom generic WiFi Linux driver called b43-fwcutter and firmware-b43-installer, these aren’t the optimal drivers as they don’t support 5ghz which is mentioned in the documentation. However, i’d recommend to install them anyway by means of building a same enviroment as I had. Hereis how I did it.
I had to reboot. Wi-Fi was working and I was able to see the following channels:
sudo iwlist wlan0 channel wlan0 14 channels in total; available frequencies : Channel 01 : 2.412 GHz Channel 02 : 2.417 GHz Channel 03 : 2.422 GHz Channel 04 : 2.427 GHz Channel 05 : 2.432 GHz Channel 06 : 2.437 GHz Channel 07 : 2.442 GHz Channel 08 : 2.447 GHz Channel 09 : 2.452 GHz Channel 10 : 2.457 GHz Channel 11 : 2.462 GHz Channel 12 : 2.467 GHz Channel 13 : 2.472 GHz Channel 14 : 2.484 GHz Current Frequency:2.412 GHz (Channel 1)
But like I said, we need higher channels to access 5Ghz. Instead:
sudo apt-get install bcmwl-kernel-source
So after all the work, I was still looking at a limited channel range. Wicher pointed me in the right direction. There are files that are limiting us, let’s find them!
sudo find / -name regulatory.bin
Mine was in /lib/crda. First we copy the file to a safe place to have a backup
cd /lib/crda/ cp regulatory.bin ~/applications/db2bin/regulatory.old
now we need to extract the data
regdbdump regulatory.old > regulatory.redb
Now we edit the file. I copied the Korean settings over the dutch one in an editor.
Band [MHz] Max BW [MHz] Flags Max antenna gain [dBi] Max EIRP [dBm (mW)] 2402.000 - 2482.000 20.000 N/A 20.00 (100.00) 5170.000 - 5250.000 20.000 3.00 20.00 (100.00) 5250.000 - 5330.000 20.000 DFS 3.00 20.00 (100.00) 5490.000 - 5630.000 20.000 DFS 3.00 30.00 (1000.00) 5735.000 - 5815.000 20.000 3.00 30.00 (1000.00)
Band [MHz] Max BW [MHz] Flags Max antenna gain [dBi] Max EIRP [dBm (mW)] 2402.000 - 2482.000 40.000 N/A 20.00 (100.00) 5170.000 - 5250.000 40.000 NO-OUTDOOR N/A 20.00 (100.00) 5250.000 - 5330.000 40.000 NO-OUTDOOR, DFS N/A 20.00 (100.00) 5490.000 - 5710.000 40.000 DFS N/A 27.00 (501.19) 57240.000 - 65880.000 2160 NO-OUTDOOR N/A 40.00 (10000.00)
Then we go back to bin, I used a python script I pulled from github.com/zioproto/.
python ./db2bin.py regulatory.output regulatory.redb
and now we overwrite:
sudo cp /home/<user>/applications/db2bin/regulatory.redb /lib/crda/regulatory.bin
And after the reboot:
joop@joop:~$ sudo iwlist eth1 channel eth1 26 channels in total; available frequencies : Channel 01 : 2.412 GHz Channel 02 : 2.417 GHz Channel 03 : 2.422 GHz Channel 04 : 2.427 GHz Channel 05 : 2.432 GHz Channel 06 : 2.437 GHz Channel 07 : 2.442 GHz Channel 08 : 2.447 GHz Channel 09 : 2.452 GHz Channel 10 : 2.457 GHz Channel 11 : 2.462 GHz Channel 12 : 2.467 GHz Channel 13 : 2.472 GHz Channel 14 : 2.484 GHz Channel 36 : 5.18 GHz Channel 38 : 5.19 GHz Channel 40 : 5.2 GHz Channel 42 : 5.21 GHz Channel 44 : 5.22 GHz Channel 46 : 5.23 GHz Channel 48 : 5.24 GHz Channel 149 : 5.745 GHz Channel 153 : 5.765 GHz Channel 157 : 5.785 GHz Channel 161 : 5.805 GHz Channel 165 : 5.825 GHz Current Frequency:5.745 GHz (Channel 149)
As it turns out, I am able to access the faster WiFi channels at my home. This brought enormous joy to me, a sense of liberty from the software regulations. But it took my entire Friday evening to achieve it. So next I’m in a pickle: high speed WiFi Linux or back to OSX. Let’s see…
Exactly two months ago, Google announced that they were ending Google Reader on July first. Google reader is a service which aggregates content from various websites served by web feeds. For me, it’s my news feed to stay up to date with the people back in Asia and the IT industry. Could they be closing Reader because the free service is still driving more traffic than Google+?
Anyway, since that sudden decision of Google, people seem to be taking stock of the company and started to be reserved about trusting their services, like Jeff Hunsberger:
When Google announced that they were shuttering Reader it made me take stock of how I felt about the company and how I interacted with them. I looked around and saw how heavily invested I had become. Google’s interests and mine were diverging. When they were innovating they always seemed to be pushing the boundaries of what could be done on the web and focused on making it better.
But somewhere during the rise of Facebook, things began to change. Google’s focus was on ad revenue and how to monetize these great base technologies they had helped create and foster. Their focus shifted subtly at first and I was forced to ask the question more and more I am willing to give up access to my personal information for this product? Is it really that good? In most cases, the answer was “yes”.
Long story short, Jeff has been moving away from Google. I read this post a month after he wrote it but I have been trying exactly the same, stopped to use Google services here and there. My thought on the matter is: you get exactly what you pay for. In the end Google is a company that’s in the business to make profit. So then I started to wonder: Oh no! what if Google quits this and that service? So, without further ado I proudly present: A top 15 list of Google services in the priority I need them. Google: Please don’t close anything in my Top 5 anytime soon ok? OK TNX Bye.
My grand ranking of Google services:
I tried Yahoo for a week. Didn’t even try Bing. Seriously. Google is a mindreader knows what I’m trying to find. Nothing to change there. However, for some specific searches I started a move to duckduckgo, wolframalpha and nerdquery.
Absolutely required for my life.
I tried Apple, Bing, Yahoo and Naver maps but they all fail to get me on my way. For now, Google remains.
Keepers weepers! I need to get home some times…
3. Google Analytics:
From a professional standpoint, I can’t practice my job without google analytics and webmaster tools. However, it was fun giving piwik a go on this private blog for a week. It did show a lot more then GA, like IP addresses but I abandoned this trial because I saw that pages loaded 20% slower compared to the GA embed code. Instead of Piwik, I’m thinking of Logstash + Redis + Elasticsearch + Kibana 3 for a future project. For now GA remains.
Keeper! until the world moved on…
In our house, we watch a lot of video from bloomberg, tudou and dailymotion. But for silly cat movies, there is no place like the youtubes. A cool thing of youtube is it’s HTML5 player (no Flash), one annoying thing is that youtube has been repeatedly suggesting/forcing a ‘real username’ down on all it’s users recently. But most content is here…. I guess I won’t be blocking youtube any time soon.
Tough one. I’d say close it and see what the rest will do, talking about that level playing field but I can’t do without cat video’s. Keep it!
5. Scholar & Public data
Google scholar is underappreciated. It’s free and it’s informative. However, my university still grants me access to a range of libraries. So I’m not dependent on Scholar anymore for research. However, I’d still like to make use of it’s vast contents and old library books. Also, Public data shows OECD information work hour information better then OECD could do.
No, please keep Scholar & Public data alive.
- Adwords and Adsense: Have you tried a CPC campaign on Facebook? For now, Google is the standard. Also for banner income.
A tough one… We have two Android phones in our house and they are old and painfully slow. My current device is an iPhone. Perhaps ubuntu is an option by the time I want to replace that? Seriously, for now I’m burying my head in the sand and want to say get rid of it. But in the sake of a balanced world i’d say:
Ok fine, keep it… there needs to be more than iOS out there… For the sake of choice.
8. Google Docs
I have a dozen of shared documents on Google documents but use the service once a month at most. I noticed more people are using other cloud services and personally I have been using Naver nDrive.
Gone! No tears would be shed here.
9. Google+ and hangouts
I quit google+ on 13 March 2013 and it felt good. I wasn’t waiting for another Social network. However, I’m still active on Twitter and Facebook but less then before.
Google employees swear by Google+ Hangouts, which I closed down. I’m still using Skype since everyone is still there at the moment.
Gone! I wouldn’t even realize if they closed it tomorrow.
On the desktop, I’ve always been a Chrome evangelist, converting many IE, FF and OP users to Google’s browser. So I wondered how it was on the ‘other’ side. At first I gave Maxton browser a try, then Sleipnir. After a week I had enough and moved to Firefox and love it. It syncs, has addons and is fast. However, I am still running Chromium on my laptop with a logged in Google account for work related matters, on the other side, I have been logged out of Google on my Firefox for a few weeks now. More on that later. On my phone I am a Mercury user.
Gone! As I do believe in choice but webkit YAY! And forking webkit wasn’t a nice thing to do Google.
11. Calendar & Contacts:
At the office we use Microsoft exchange, at home I share a calendar with my wife on Google. Last month, I setup an owncloud server and setup ical to make use of those instead. Owncloud supports CalDAV and CardDAV, syncing all my devices to each other.
Hah, don’t need those anymore.
I started using Gmail in 2005. I have moved all my family members there as well. Now, eight years later I have noticed that other mail services have evolved as well. There are many out there like freemail and foreign services. I’m using a Korean one called Naver. The point is, I left my @gmail account completely for a month and seem to have no problems sending private mails using my private @joop.in domain. However, Google spam filtering is better. For the rest, a painless switch.
Gone! I could live without Gmail. So can you. Believe in yourself!
I do a bunch of translation from and to Korean every day. Bing and Naver are far superior to Google translate.
In the trash! I could live without Translate.
I use news to search for real time events. There are other services for this. So, I guess I’d rather had seen news go then my beloved reader.
Yes, I could live without Google News.
Google bought Feedburner in 2007 but sadly, the product has hardly been developed after the transition. I was using feedburner about insights in RSS and for email subscribers, I left feedburner for mailchimp back in 2012 and haven’t looked back since. I have feeling Feedburner might not live for too long without Reader anyway so I’d recommend moving as well!
Yes, I could live without Feedburner, in fact, close it right away and see if I care. Ta ta…
My colleagues seeing me use yahoo jokingly said that Google wouldn’t notice my abandonment of their services. But that wasn’t the point of this odd hobby I picked up in the last two months.
I wanted to know how dependent I was and tinkering away from the big Google seemed like a fun way of demonstrating this. I have friends and collegues who work at/with Google all the time. It was also fun teasing them as well. A month in, I know what I definitely need (Search, Analytics, Maps and Youtube) but wouldn’t be sad to see other things go. In the end, I keep on search and might perhaps go back. Alternatives are always a good thing.
The whole reason why I started to write this silly blogpost was about Google closing down their Reader. Up to today I haven’t decided on the replacement service. Luckily, I still have until July for that choice. For now, all feeds are still maintained by Google.
A female minister in the Netherlands got a lot of press yesterday after saying women in The Netherlands are too dependend on their husbands. One in five Dutch women work more than 35 hours a week.
A foreigner observed the following:
Dutch markets and town squares are busy at midday because Dutch stores still close at 6 or 8 pm, often by law. This is one reason why women are less likely to work full-time: who would do the shopping, and when?
Let’s just calm down and do groceries together at 7pm, makes it more fun anyways.
The state of silicon valley today in a nutshell:
The general public doesn’t understand start-ups at all, Paul Buchheit said. They’re mystified how a company with no revenue can be worth a billion dollars. It’s because of this power law: If a company has a 1 percent chance of being a hundred-billion-dollar company, then it’s worth about a billion dollars. Imagine an assembly line where Facebooks and Googles come along every few years. You can either pick that cookie off the assembly line or not. If you pick it off, it’s market price, which varies. But if you don’t pick it off, you’re out of the game. We’re in the early days of the Internet, Buchheit said. Every other industry will be eaten by tech.
Is there a punk scene in China? Places like Yuyintang and Logo in Shanghai hosts hardcore and punk shows quite often. The Beijing punk scene is even bigger. When visiting shows, I was handed out magazine’s featuring biographies, upcoming concerts and yes, stickers. The music is punk but it’s very Chinese, the crowd is very diverse. Mostly teenage boys and girls looking for fun and something different. Mohawks, beer, tattoo’s, pits and Chinese buttons. Yes, punk is very much alive in China.
☆ Watch the video at China Calling! - Crane.tv
☆ Picture from Shanghai247.net
Chris Sauve has a great presentation over at TED about managing habits. Write down everything. Reduce your life to the essentials and keep questioning them.
He placed habits in four categories:
- Don’t have to do and don’t like
- have to do and don’t like
- don’t have to do and love
- have to do and love
If it’s a habit you don’t love and don’t like to do, simply eliminate the habit. He continues, if it’s something you don’t love but have to do, simply automate it. (Just do it, don’t make tough decisions about it), make time for the things you love!
A colleague visited a baseball game in Seoul, Korea in 1995. During match break time, a legendary LG Twins player signed a baseball and hit it into the crowd. What were the odds of my Dutch colleague to catch that very ball? He took it back to Holland and kept it proudly at his house, not knowing who that player was.
Until one day 18 years later, the collegue mentions this very story to me and we both became curious. The next day, he brought the ball to our office and we were staring deeply at the autograph to decipher the origin, but we gave up after some time. Then we decided to take a picture and put it on Facebook. A few hours later, people (Mostly Korean men) found out that the player was 김재현! In fact, the discovering commenter was bright enough to notice that we had actually been holding the ball upside down! So far for our own search but again solved thanks to the power of social media…
The mystery is solved for now but now the next question is, what is the ball worth?
This blog is now powered by Jekyll. What is it and should any wordpress user consider the same?
If you are familiar with wordpress, you will have certain expectations from a blog: A backend to login, edit posts, manage comments and oops! roll back mistakes through revisions. However, a database (MySQL) driven blog can be quite heavy for webservers, resulting in performance problems and higher costs to run the blog.
Furthermore, storing blogposts in a database doesn’t make a lot of sense. On a typical wordpress setup, the server calculates and draws the page for each page view. Both from your database and your PHP content. This doesn’t make sense on static content like a weblog post, because the calculation stays the same as long as the blogpost doesn’t change. At first, I put caching techniques like Varnish on my blog to reduce the amount of (re)calculations. But I was still left with a heavy page with a very long Time To First Byte (TTFB), relying on frequent wordpress security updates and bulky plugins. I wanted something better.
With Jekyll, a blogger writes in so called “markdown” (.md) files, which are an easy-to-use formatting syntax created by John Gruber of Daring Fireball. You simply write your text on your computer (there are editors for this like Texts) and then save your .md file to the specified directory and type ‘generate’ in your terminal. Now, each blogpost is generated to it’s own HTML file, which can be uploaded to the web server. No more database, no more PHP.
Here is the interesting part: Without PHP and database requirement you don’t need a fancy server. I’m hosting the blog on Amazon S3 now. Others even put it on Github. Just as long as you use a service like Disquss to take care of the only non static content: comments, you don’t need any PHP/MySQL server. S3 is way faster then my own server and more reliable.
The process to move
Last weekend I’ve put some time into moving to Jekyll. I got the job done with a few guides. First I exported my wordpress blogpost to .md files, I used a plugin called Exitwp and wpXml2Jekyll but finally used Jekyll + MySQL server connection to maintain UTF-8 and blog attributes as much as possible. Then I made the layout work, for which I used Octopress . Octopress comes with a nice preset of HTML, CSS and JS templates. I decided to design my own template and build it to look like my Wordpress blog so blog looks the same but it’s faster and easier to manage.
After it was setup, I noticed the import didn’t know how to process [caption] tags, which were generated by wordpress to put captions underneath images. I found a nice plugin by Robert Anderson and then wrote some regex to batch convert wordpress captions to fix this problem:
and change that to:
Captions had to be scanned because this regex didn’t solve everything but it did most of my work.
Then I rendered the blog and uploaded it to Amazon S3. I also use cloudfront CDN to serve pages quicker around the world. I used a few guides for this: ’Octopress and Amazon CloudFront’ by Jaco Belder and ’Jekyll blog on Amazon S3 and CloudFront’ by Tobias Sjösten.
And that was it in terms of getting the blog up and running.
In terms of webpage performance, I still need to wait for data but I can see there is a big improvement compared to the load time of my (Micro EC2) Wordpress blog at 4.7 seconds last Friday, with an average website speed of 10+ seconds over the last month. I just ran the same test and the same looking homepage came in at a whopping 1.2 seconds. I’m expecting a 400% performance bump in Google this week!
In terms of how this system works compared to wordpress…. I’m going to evaluate how it works in the coming weeks. I can always go back if I wanted to. I raised a question in the introduction of this post: What is Jekyll and should any wordpress user consider the same? As it seems for now is that generating posts seems to difficult for a typical user. On my system I render and deploy with a single commands. Perhaps these steps can be improved some way.
A big take away for me: You don’t lose central management without a database. This is a huge eye opener to me. Now I have a folder full of .md files on my hard drive instead of the database, I feel like I am MORE in control of the content. As said, I ran into a problem related to [caption] tags across my imported wordpress blog posts and I was able to solve it as easy as it would have been within a database.
For the next step? I just have one more thing to do… I need push it into Git or Mecurial. Then I would be able to perform revisions. Besides that, I’m now thinking that with some aggressive optimization, I should be able to hit the 1 second threshold? Further study should dive into how static hosting help page speed of the e-commerce websites which are only slowing down lately.