Cela fait à présent un peu plus d’un an que Google a arrêté son service Google Reader. Ce dernier était un agrégateur RSS qui permettait notamment de centraliser tous ses fils rss, de pouvoir y accéder de n’importe où du moment qu’il y avait une connexion internet pour les récupérer et bien sûr de lire le contenu.
Pour rappel, RSS est un format de données qui est utilisé pour syndiquer certaines données. Concrètement, la plupart des blogs et sites de news proposent un lien “fil RSS” auquel l’utilisateur peut souscrire. Grâce à ce lien, il est possible de savoir si du contenu nouveau a été publié sur le site en question, et si c’est le cas, de récupérer le contenu sous forme de flux RSS qui est converti en HTML pour être lu grâce à l’agrégateur.
Google Reader
L’apport de Google Reader était pour moi la synchronisation. Au lieu de souscrire aux différents fils RSS via un client sur son ordinateur, il était possible d’utiliser le service de Google afin de les stocker tous en ligne et de les lire via l’interface Google Reader. Ainsi, lorsqu’on est en déplacement sans son ordinateur, il était possible d’y accéder en se connectant au service. Cela est devenu encore plus important avec l’avènement des plateformes mobiles.
Personnellement, je n’utilisais que très peu l’interface de Google Reader. J’utilisais plutôt le service comme backend pour pouvoir centraliser et synchroniser mes flux. Cela était couplé avec une solution front-end qui se connectait à mon compte Google Reader pour récupérer mes différents articles. Ainsi, je pouvais utiliser n’importe quelle application du moment quelle supportait Google Reader.
Un nouveau choix s’impose
Aujourd’hui Google Reader n’existe plus. Il a fallu trouver une alternative. Plusieurs contraintes se sont alors posées à moi.
La première était que je voulais absolument dissocier l’agrégateur de la solution de lecture. Cela me permet principalement de ne pas dépendre à 100% d’une seule solution et de diversifier d’une certaines façon le risque. Cela permet aussi de se construire la solution idéale, à savoir choisir l’engin qui va agréger et synchroniser le contenu mais aussi l’application qui va le lire, que ce soit sur mobile, tablette ou ordinateur. Cela sous-entend donc une API pour que les développeur d’applications puisse intégrer le service, et un semblant de popularité.
La deuxième était de choisir un service qui ait tout de même une interface web au cas où je souhaite lire mes articles alors que je n’ai pas mes terminaux habituels.
La troisième était de prendre un service payant. Le but ici est de ne pas être assommé de publicités ou de voir ses données venues ou utilisées d’une quelconque façon. Toutefois, mon objectif principal était plutôt de choisir une solution de stabilité. Cela est rendu possible en ayant derrière des gens avec un business rentable. Ma peur était de changer pour un service gratuit qui n’arriverait pas à tenir le coup faute de revenus suffisants pour l’entretenir.
A few weeks ago, I was looking for a new desktop wallpaper but couldn't find a great source of wallpapers. Some websites offer giant selections from which I could choose but the quality wasn't that good. I decided to buy the domain wallpapermonthly.com at that moment with the hope of building and curating a website with a monthly selection that favors quality over quantity. I don't expect to select more than 5 items a months. This happened during the time I was supposed to revise for my exams so I couldn't act on it but I was still very excited.
After I took my exams, I allowed myself a lot of time to rest and one day — actually on my birth day — I thought it would be fun to build a website in a day. The general idea and the domain were already here, but I didn't know how the website would look like or work. Even if I only had a few hours, I didn't want to rush it and took some time to think about it and put my ideas on paper.
The Product
The idea is to provide a monthly selection of wallpapers. I don't want it to be complicated. Just display a few images in one post per month. I want it to be very simple, with no unnecessary information. It must be easy to navigate through the months with newer and older links. There also should be a simple archive page with only links to each month. I also don't want to have any problems with copyright information, so I will only host 800px large preview files and offer a download link to the author's page unless they ask me to host the hi-res wallpaper. People should also be able to submit their work or work that they find worthy of sharing with our readers.
Which CMS ?
It was out of question to build a static website of course. Wordpress is too heavy with and I wouldn't use half of its features. In the last year, I've become very familiar with Kirby CMS, a file-based CMS that requires no database. The good thing with Kirby is that it is so flexible that you can do almost everything with it.
Sketching it
That part wasn't that hard because it was clear in my head how the website would work. I only used my pen and paper, sketched something in 5 minutes and designed it directly in the browser.
I am aware that the interface design is far from perfect, I plan on improving it without affecting its simplicity.
Building it
Once again, everything was clear in my head so that wasn't very hard. I'm going to explain the general structure of the kirby-powered website.
Structure
In the content folder, I created 5 folders :
error : 404 "not found" page.
entries : front-page where the magic happens; all the subpages consist in monthly entries. Envision it as a blog with monthly blog posts.
archives : generated page with all the links to previous entries.
about : display the purpose of the website, a colophon.
contact : to submit an entry.
An entry
The entries directory acts like a homepage. In this folder, there is a folder for each month named with this pattern : yyyymm - name (y: for year, m: for month) and that contains an entry.txt file redacted in markdown with a title and a date only. It also contains the previews for each wallpaper named with this pattern : nn - name.jpg (nn being the number that represents in which order I want to display the images). For each image, there is a text file named nn - name.jpg.txt with the metadata for each image : title, author, link.
The magic happens in the template
With kirby, you can create templates for specific pages. For this you just need to use the same name for the .txt page and the template file.
The entry template is easy to understand and to build. For each image contained in the entry folder, kirby should create a list element that contains the image followed by the title with the link to the wallpaper and the author's name. If there is an older or newer entry, a link to this/theses entry/entries should be created.
The entries template fetch the latest entry's information and displays a link to the previous entry.
Publishing
I don't even use an administration area, it all happens with a text editor and an FTP connection. For each month, I just have to create a new folder yyytmm - month year, with a text file entry.txt containing the name and date. I then download the wallpapers, resize them and rename them. For each wallpaper I create a .txt file with the same name containing the title, name and link to the wallpaper. I upload it. I'm done.
Pour finir
I encourage you to visit wallpapermonthly, submit your work or someone else's and give me feedback.
I judge a book by it's cover. I judge an app by its icon. Yes, I might be missing something really good and worth my time and money, but you also might miss many consumers like me who do the same. I know it's not fair because your product could be excellent if we look beyond the interface or app icon for instance. But the thing is I sometimes don't have the time to take into consideration all my options and compare all the similar products, so I select. Which ones do I take into consideration ? Those that catch my eye. Take the time to think about your packaging, user interface or icon. It'll be worth it. I might even notice your app, realize how great it is and end up buying it.
