CarmaBlog

 Version française disponible

Google is part of our daily lives

Countless news reports are chronicling the hegemony of the US firm. Google’s success fascinates, irritates and sometimes scares. Like most Internet users, it is difficult to spend a day without using one of the services provided by Google. At some point, you ever :

• Browsed the web with Google Chrome
• Searched the Internet with Google Web Search
• Read your emails with Gmail
• Organized your appointments with Google Calendar
• Managed your photo albums with Picasa
• Chatted with your friends with Google+ or Google Hangouts
• Watched videos with YouTube
• Read books with Google Books
• Managed your documents with Google Docs
• Translated an article with Google Translate
• Planed your moves with Google Maps
• Managed your phone, contacts or alarm clock (and more) with Android
• And even more…

The giant has diversified and is now competing on all fronts (automation, robotics, automotive…). Its strategy is still focusing on individuals and how to improve everyday life. You can read this list of mergers and acquisitions by Google to convince yourself.

One search may hide another

Did you know that Google is gathering and storing all your searches? Google is able to time travel in your search history, to the nearest second. The fact that you have searched for information on your future trip to Thailand on August 10th, 2005 at 10:05 may not be very useful to you, but this information among other represent a goldmine for Google, in order to establish your very detailed profile. It is however possible to disable this feature if you go to your account’s advanced settings. This is not something that an average user does. This is not a coincidence if this setting is enabled by default. A Google search may be trivial at first, but imagine that you’re looking for:

• Cars and brand comparison, Google will determine that you will change your car soon
• Fashion and clothing throughout the year, Google will determine that you are working in this sector
• A disease and related remedies, Google will determine that you have this disease
• Tips to save money, Google will determine that you are barely making ends meet at the end of the month

Jobs, places, favourite musics, movies, cooking, hobbies or political views. These examples are fairly simple, but it can goes very far when overlapping the information. It becomes possible to tell very accurately the events of a lifetime. If you think that Facebook is the biggest player, you’re probably wrong.

You’re the product

Google is continuously collecting huge amount of data from and about its users. It builds and consolidates complete profiles of our lives, and probably knows you better than your own mother or even yourself. Google’s services are free in exchange of your personal data. It is an implicit contract that you agreed (not necessarily read) when creating your account. The data collected can be sold and used for targeted marketing campaigns, statistics, etc. After all, it is fair game!

Smile, you’re on filter

The search engine displays a customized list of results, showing only the results that are relevant to your profile. You can do this simple experiment at home: run several searches with the same set of keywords but using different Google accounts. The list of results will probably be slightly different, especially if people are different (e.g. different gender or generation). The fact that the results are relevant to the connected user is a positive aspect, but it can also be a disadvantage. The primary purpose of the Internet is to make the universal knowledge accessible with a censorship as limited as possible. If you add an intelligent filtering system on your searches, there will be articles or even entire sections of the internet that you’ll never be aware of, simply because Google has decided so. Somehow, the knowledge and access to the information become limited or even incorrect.

I’m not anti-Google

Do not be mistaken, this article is not a pamphlet against Google. I am a very satisfied user for years and it will be probably painful to navigate the web without their services. This is more of a wakeup call. It is worth to spend some time to set up your account rather than letting Google decide for you. Unfortunately, I don’t know any magic tricks to continue using Google while preserving your privacy. However, I can advise you an alternative search engine DuckDuckGo. This search engine has gained great popularity by focusing on the respect of your privacy without neglecting the relevance of the results.

 Version française disponible

Quite recently, I was looking for a way to generate tweets from a RSS feed.

This is not something very complicated to develop, it is a simple service that will pull your RSS feed regularly and create a tweet (through the Twitter API) for each new item.

Rather than programming a tool with all the maintenance involved (e.g. change of the Twitter API), it is better to see what is already there. So I decided to use Twibble.io, a quite popular and promising solution. Unfortunately, Twibble has huge limitations. If the user interface is user-friendly, the site is often unavailable, RSS entries are sometimes skipped for no reasons. The free version comes with some advertising, as all generated tweets are ending with a Twibble reference. Honestly, if the service was meeting my expectations, this second point would not have bother me at all.

So I looked for another service and I discovered Twitter Feed. I must admit that the interface does not look like much, but it has the advantage to work very well. I am now using it for the past few months without having any issues. It can tweet on Twitter and also post on other social networks like Facebook or LinkedIn, without advertising. Sometimes there is no need to force a user, I was not forced to write this, I’m just an happy user.

Of course, there are other services available such as dlvr.it and probably many others. But for now, I haven’t felt the need to try them.

 Version française disponible

A WordPress blogger has a tendency to install many plugins on his/her blog in order to add functionality and some interactivity. Read WordPress plugins of CarmaBlog if you are not convinced. Over time, the blog becomes slow and longer to load. I think I got to the point I just had to do something…

The site GTmetrix is able to calculate the performance of any website by assigning an overall grade and provides a detailed and useful report to help you optimize and improve your rating. The goal is to get a decent score, I would say between 75 and 85%, which corresponds to a rating between B and C. For CarmaBlog, the result was not very glorious with a rating of 60% and an overall grade D, the loading time of the home page (i.e. 3 articles) was about 6-7 seconds, it was just ok.

A cache management plugin: W3 Total Cache

At the end of the year, I decided to tackle this problem by installing a WordPress plugin to add some caching features and speed up the blog. For those who don’t know, a cache is a place (usually some disk space directly on the server) where pre-computed pages are stored. The server displays these pages directly instead of recalculating the original page every time. This process significantly improves the performance. But note that you have to clear the cache if you modify a page, otherwise you won’t see your changes. To be honest, at the beginning I was a bit afraid of breaking my site, but the poor grade of GTmetrix gives me some strength to do it.

I finally choose the W3 Total Cache plugin which is very popular. My choice was somewhat arbitrary, depending on the number of users and also the ratings pretty good. For this kind of plugin, the configuration is the most important and probably the trickiest part. You have to go step by step and be patient… The best way to break everything is to activate all optimizations simultaneously and see the final result (that’s also a good option if you like taking risks…).

When I was going through the configuration, I realized that W3 Total Cache was supporting plugins, including plugins for Feedburner and WordPress SEO. A good thing as I use both of them.

Some manual image optimizations

Once the cache management was set up and properly configured, I noticed a slightly better performance, but that was not transcendent. GTmetrix also provides many indicators to help you improve the performance, here are the two manual things I realized:

• Optimize images. GTmetrix provides optimized images you can download and use to replace your old ones. You can easily save 50% space while preserving the quality… Impressive! Imagine the gain when you are handling a couple of pictures of 200-300 kb each. Also, I suggest that you bookmark this site: TinyPNG, definitely a valuable help for your next images. This optimization is particularly important for the common images including in the header, footer and sidebars… as they will be downloaded for sure by your visitors.
• Define dimensions on images. Attributes width and height on img tags. We usually ignore them, but they prevent the browser to do the calculation for you and therefore it will increase the performance. In addition, the space occupied by the images is already set so you’ll avoid some annoying page or scrollbar moves (if you didn’t understand, trust me that’s something positive).

These are the two major optimizations that I had to do manually. W3 Total Cache handles the rest automatically.

A free CDN: CloudFlare

Second step, I see the possibility of coupling CloudFlare with W3 Total Cache through an internal plugin. I take a look at CloudFlare. In fact, I had already heard of it, but that was all.

Good news, they had a free plan and there are no ads or specific limitations (at least for a personal blog like mine). But what is it actually? CloudFlare acts as an interface between the visitor and your host (1and1 in my case) and serves as a CDN (Content Delivery Network). It’s all Greek to me, but it seems to have a great impact on the perfs and GTMetrix advised me to use a CDN so… Unfortunately, there is some DNS reconfiguration to do, brrr that is scary. I’m not specialized in networking, but the DNS (Domain Name System) is like a directory, from the domain name, it matches the IP address of the server to reach. Changing of DNS is like changing of address, and if the move goes wrong, you are no longer accessible. After some thinking, I decided to continue, it is better to break everything during the festive season when the traffic on my site is quite low.

So I create my CloudFlare account, I add the site CarmaBlog, I leave the default settings for most of the options. I choose a fairly low security profile (anyway, it could hardly be worse than before). For the performance profile, I choose “CDN only” as W3 Total Cache will handle most of the optimizations, and conflict is the last thing I want to see (be reasonable). The setup wizard is very nice, especially when you are a beginner. Then comes up the DNS configuration, CloudFlare asks me to update my DNS settings from my 1and1 account… Oops, let’s move on to something more serious!

This is the touchy part as your website can become completely unavailable in case of some misconfiguration. I feverishly logged in to my 1and1 account, select the domain config, there are a lot of warning messages like “Are you aware of what you do?”. I am not convinced myself, but I click “yes” then I fill in the fields to point to the two new DNS provided by CloudFlare. This DNS change can take between 1h and 48h. It is all about network configuration, far beyond my skills, so I confirm my choice. At the same time, I enable the CloudFlare plugin in W3 Total Cache, and establish the link with my CloudFlare account by filling in a couple of fields… It is already late, so I go to bed.

The next morning, one of the first things I do was to go on the blog. Phew, it was still available. I already noticed some speed (perhaps psychological, we’ll see in the conclusion…). I logged in to my CloudFlare account and… Victory! CarmaBlog is well configured, the DNS reconfiguration went well, now the traffic goes through CloudFlare. CloudFlare provides a dashboard showing visits, threats (yes!), the bandwidth saved, etc. Then I ask myself if going through a third party provider was really a good choice? What will happen if the CloudFlare DNS is down? But at the end, a DNS server is rarely down and if it falls, there is always the backup to take over. And the problem might have happened with 1and1 DNS, so there should be no worries! In the worst case, I can always switch back to the previous settings and use the 1and1 DNS again within 24 hours.

Conclusion

I go through the blog to make sure everything works, I turn on private browsing to be sure that the browser is not using cached data. So far so good, it is fast and fluid, really good news. I submit the site again on GTmetrix, and there you go… The performance is much better. I have more than 80% and an overall grade B. A is available from 90%, it is difficult to achieve such a score if your blog is using a lot of plugins and a quite complex theme.

After all these optimizations, the page size go down from 1.2 mb to 800 kb (this obviously depends on the pictures, but I made the comparison on the same page to give me a good idea), I got 33% improvement. Meanwhile, the number of requests falls from over 100 to around 65 (thanks to the CSS and JS combination among other things). I had already noticed a faster loading time with W3 Total Cache, but having CloudFlare on top, I went from 6-7 seconds to below 3 seconds (over 50% gain).

The combo W3 Total Cache + CloudFlare really works great! This article is more like a feedback than a tutorial, but you will find all the resources needed on the Internet if you decide to do this change.

Not only my host should be happy, because most of the bandwidth is now rerouted to the CloudFlare servers (much more powerful architecture), but my visitors too, as the page is displaying much quickly than before. Today, the only question I ask to myself is: why did I not do it earlier?