The idea is pretty simple but could come in handy: when your OS’s windows are moved off-screen, their essentials UI elements (think: buttons) stay at reach, sticking to the screen’s limit. Go ahead and play with the demo, you’ll get it!

Thoughts welcome, of course.

ps 1: A hot thread about it on Hacker News.
ps 2: @raphaelbastide’s take on the idea: greasy!

Then I started working on a freelance project and Crunch came out; I was sold. This app made me choose LESS over Sass, which may not be the best of reasons, but it seems like I’m not the only one to pick this language. As you may know the second iteration of Twitter Bootstrap also relies on LESS, even encouraging you to use Crunch.

Saturday @ilovetypography tweeted this:

Dreaming of something that takes my CSS & LESSifies it — automagically.

To which I replied that:

This seems possible! Will try and see if I can write such a tool, interesting idea…

Sunday morning I started writing JavaScript, and a few hours later I had a good-looking script. Took some time to create a simple yet efficient UI, and that was it. Least.

The tool takes basic CSS as an input and tries its best to output the LESS or Sass source it would correspond to (including nesting, classes and pseudo-classes). I also added an option to output something for the not-so-popular-but-interesting Stylus language.

As you’ll notice in the footer of the tool’s page it’s a “five-hour project”. In order to have a decent UI as fast as possible I used Twitter Bootstrap, which has its downsides but fits perfectly this kind of projects. A “five-hour project” also means it may still be a bit buggy. Don’t hesitate to ping me on Twitter or to drop comments here.

ps: Least both references LESS and “To Say the Very Least”, a track by Buck 65 on Vertex. Ah!

It is a mix of Flash (most part of it), JS (for the add form and its communication with the main UI) and PHP (CRUD logic). Read the blog post for more details on it.

I recently started looking at JavaScript frameworks and tools, probably like most Flash developers. Not because Flash is dead, but because the more tools you master, the better you are. And because yes, everyone is asking for JS and CSS3 these days. After some investigation I decided to give Backbone.js a try. And re-writing Kwot from scratch with it seemed like a great exercising opportunity.

So here its: Kwot JS!

It should work really well on Safari, OK on Chrome, not-that-perfect on Firefox and poorly on Internet Explorer. You were looking for a reason to continue using Flash for building “experience sites”? There you have it: compatibility. Kwot JS relies on CSS3 transforms to allow this perspective view on quotes, and on property transitions to animate the quotes’ colors. Of course this doesn’t work in all browsers!

Now, about Backbone.js. The framework’s requirements and logic are good, but coming from an AS3 world mostly everything looks odd, for better or worse. Note this is my first project with it and I might have missed some good practices or handy tips.

The good thing is its persistence layer. In Kwot, quotes are stored in a database, I simply wrote a simple RESTful API with Slim (try this PHP framework: win) and connected it to my Model classes. That’s it, it just works. Underscore.js’ templating system is handsome, too. Never used that before, cool stuff.

The bad things are, as far as I’m concerned:

• No strong typing. Of course this isn’t Backbone’s problem, but rather JS’ one. On this particular point your IDE can help, but you won’t get runtime errors.
• The “this” keyword. You always end up with plenty of lines all starting with a “this.something”. It frustrates me.
• Scope issues: when listening to events you pretty much always have to use Underscore.js’ bind method in order not to lose your infamous “this” scope. I don’t like that.
• The MVC pattern. If you ask me: it sucks. I’m no integrist but I wouldn’t have called what Backbone provides MVC. I was always wondering “who should do that?” and never really came up with a satisfying answer.

Overall the learning curve is pretty steep, I think it took me less time to write this version than the Flash one, but I’m a 3 year more mature developer. Also, when I wrote the Flash version I didn’t really know where I was going and I kept experiencing with concepts and ideas to see what would look best.

On a side note the Flash version uses Neutra, whereas this one uses Questrial (hosted and served by Google Webfonts).

Tell me what you think, both about the app and the framework!

I then contacted the fine folks at Adobe Developer Connection and offered to translate it to English. They accepted, reviewed it, and boom: published it.

Wether you’re a French or English reader, I got you covered. Tell me what you think.

It quickly became successful. It now has been downloaded 168,000+ times, a best-seller of sorts. Except it’s free.

But with success comes feedback, and with feedback comes feature suggestions. Most of them were included through updates, some of them didn’t make the cut. Probably because of me being lazy or because of AIR’s limitations.

Then AIR 2 came out, then I learnt Robotlegs… So I re-wrote it from scratch! Introducing Shrink O’Matic 2, the same quick and simple app but with more features and a nicer theme.

Here’s what’s fresh off the oven:

• Drop folders onto the app: every image in it (or in its sub-folders) will be shrinked.
• New “Rotation” settings pane: either use a specific angle or let the app read your images’ EXIF data and decide what to do.
• Custom name option: choose exactly what the output name will be using your own pattern and injecting the original file’s name (using $name) and/or its position in the queue (using $num).
• PNG files now keep their transparency when shrinked.
• Watermark: watermark your images, even choose where to place the overlay.
• Drop files onto the app while it’s processing, no problemo!
• No more dimension limits.
• Shiny new theme!

But! I decided some features had to go. I removed the “name preview” that used to be in the status bar. I also removed the ability to drop images from web pages. If you need these features and want them back, make sure to drop a comment and let me know!

That’s it, go get it!

I recently published their brand new website, designed by Christelle Bonnet and developed by me. It’s been a pretty long process (they are very busy guys) but it’s here and I like it!

I used mostly WordPress and MooTools to develop it. Everything was designed to be light and subtle, I think I can safely say that it is.

Pretty technical stuff in it, not sure everyone will want to install it but hey, there may be some doctors out there!

The hospital is French but the app is both in French and English. Built in Flash Builder with Flex Mobile, Robotlegs and AMFPHP, UI design by Jumo.

Now go show off browsing images of “Scaphotrapezial Synostosis” or “Rectus Femoris intramuscular Haematoma”, whatever it may be!

The CSS community is really active and fast-moving. When I wrote Boks, Blueprint CSS was one of the most watched and forked project on GitHub which is mostly why I chose to base my UI and logic on it (it still is at the top, by the way). But with today’s CSS3 hype and because of the ever-growing list of CSS frameworks (not going to list them here) I started receiving lots of feature requests.

At first I thought I could wait and handle them later, but I quickly realized it would need a lot of time, and I definitively didn’t have it in my hands (or at least not for this project). The idea to OpenSource it seemed obvious and I’ve been slow doing so, but here it is. If anyone’s willing to take a look at what I wrote and fix or improve it, do it! Don’t forget that this has been written a while back and I wouldn’t re-write it this way (think Robotlegs); I know the code will look crappy to some but hey, we all learn and evolve, right?

I’ve licensed Boks’ source under GNU GPL v3 in order for it to remain OpenSource, but if you have other suggestions, just tell me.

Well, it’s still here but I don’t use it much anymore, since I’m a big fan of Robert Penner’s Signals! So I wrote Signl. It’s basically the same thing, but for Signals.

I know writing Signal classes is an easy task; but, why not make it easier? I hope this helps! Oh, and by the way, it’s OpenSource, just like Evnt was.

Si comme moi vous ne pouvez pas vous empêcher de dire “flim”, “ouiche” ou encore “un pour l’argent, deux pour le spectacle et trois pour le caillou”, cette application est pour vous.

Après une semaine de vacances avec une personne touchée par cette maladie et sans avoir accès à Internet j’ai vite réalisé qu’il me fallait une application contenant tout le script de ce magnifique flim : La Classe Américaine. Je savais qu’un fou avait déjà fait tout le sale boulot et qu’il ne me restait plus qu’à extraire ces données et créer l’interface pour les parcourir…

Grâce à Google Chrome et ses Outils de développement j’ai pu injecter MooTools dans la page. Petite astuce très simple et très pratique (script à copier/coller dans la console) :

var scriptNode = document.createElement('SCRIPT');
scriptNode.type = 'text/javascript';
scriptNode.src = 'https://ajax.googleapis.com/ajax/libs/mootools/1.3.2/mootools-yui-compressed.js';
var headNode = document.getElementsByTagName('HEAD');
if (headNode[0] != null) headNode[0].appendChild(scriptNode);

Après ça, quelques lignes pour extraire les données, les nettoyer et les stocker directement dans le presse-papier au format JSON :

var data=[];
var images=$$('table.script img');
for (var i=0; i<images.length; i++) {
    var tr=images[i].getParent().getParent();
    var o={};
    var scriptTag=tr.getElements('td')[2];
    scriptTag.getElements('a').dispose();
    data.push({
        ts:tr.getElement('small').get('text'),
        script:scriptTag.get('html')
			.split('’').join("'")
			.replace(/\n/, '')
			.replace(/<br>\n$/, '')
    });
}
console.log(data.length);
copy(JSON.encode(data));

Un petit coup de Flash Builder, saupoudré de Robotlegs et hop, une appli Android ! Pas de version iPhone pour l’instant, mais si quelqu’on m’offre de quoi me payer un certificat de développeur, je ne dis pas non !

Au revoir, Messieurs-Dames. C’est ça, la puissance intellectuelle. Bac + 2, les enfants.

But I wanted something more powerful/thorough that would remember my frequently used characters. So I wrote Unipasta!

Here’s what you should know about it:

1. Every input under the selected character (char, code and hex) can be edited and will update each other. Easily jump to any character!
2. The font metrics (baseline, x-height and cap-height) are auto-calculated and will help you know where the char lives;
3. Click on the “More Info” link to jump to fileformat.info and access a lot of details about the selected character;
4. Every character your click will be automatically copied to your clipboard, handy!
5. Use the “Recent characters” list to quickly access your favorite ones (latest used will always be listed first).

If you think some missing Unicode blocks are important to you or if you’d like to add a new character listing, just ask for it!

Remember Shrink O’Matic, the “oh, so easy to use” image shrinker for Windows, Mac and Linux? Introducing Shrink O’Mobile, the “oh, so easy to use” image shrinker for Android!

Because cameras on phones take big pictures and because you might want to send smaller/lighter versions, Shrink O’Mobile is here to help out. Just launch the app, choose the way you want your image to be shrunk, pick your image and BOOM! Your fresh, smaller, new version is instantly stored in your camera roll. Easy as pie.

And did I mention the app is free? It is.

Le bureau baroque is an architecture agency in Bordeaux; playing with art, design and — of course — architecture. They’re the ones that set up the Pecha Kucha event in Bordeaux and who invited me to show off a bit! Of course they work on lots of other great projects so they needed a site to let the world know what’s up.

They wanted a highly visual site that would be easy to update, we considered a few options but quickly came to the conclusion that Indexhibit was exactly what we were looking for. But one thing that was really bugging us is how most (if not all) Indexhibit sites look the same. Indeed they’re a great way to show big images and the navigation is damn simple, but hey, why not keep these great qualities but in a more original layout?

Since we’re really keen on horizontal layouts we went for this. But something bothers me with this: you get a horizontal scrollbar at the bottom of the page; which is hard to see, because it sits at the bottom of your screen and we — dumb humans — are not really used to it. So I thought: why not have a horizontal layout scrolled by a vertical scrollbar? That seemed a bit tricky at first, for two reasons:

1. Is that technically feasible?
2. Isn’t that too weird, for the end-user?

Both questions could not be answered without trying, so I tried. And it looked cool!

When the prototype was ready I started taking a look at Indexhibit’s guts; and although it looked a bit ugly to me, I realized pimping it was no big deal… A few e-mails and burger-meetings after we were happy with the newly created theme. Christelle Bonnet helped out with her great typographic/balance eye and we were ready to go!

Go see the site live, enjoy those nice projects, play with that side-scroll concept and tell me what you think!

My “anniversary” intro

It’s been five years (this monday) since Google added the ability to simply chat inside Gmail and to store your chat history, just like your regular e-mail discussions. This poster is a celebration of that, plus a big high-five to my “chat pal” (who hopefully received my package on time), plus a technical and aesthetical look at what we wrote during these years.

Let’s make history

Back to the chat history thing… I remember being pretty happy when Google announced it, mainly because I knew I’d use it for later reference, archiving links and thoughts had become much easier.

Here’s a copy of the announcement they made:

Chat with your friends from right inside Gmail. There’s no need to load a separate program or look up new addresses. It’s just one click to chat with the people you already email, as well as anyone on the Google Talk network. And now you can even save and search for chats in your Gmail account.

So it’s been five years. And I’ve chatted quite a lot; mainly with one guy, my buddy Renaud. We chatted around 2,800 different discussions so I thought there might be some interesting data to dig in these archives… So I dug.

But digging thousands of discussions is not an easy task, so I had to take a look on the tech side of things.

Join the tech side of the force

Before digging, I had to retrieve all the discussions we had, in an easy-to-analyse format. I used Gmail’s offline feature: applying a new label to our conversations and locally syncing this label. For some unknown reason it would crash on Google Chrome so I had to use Mozilla Firefox. When syncing was done I got a pretty big file in my “Google Gears for Firefox” directory.

Cool thing is, Google Gears stores data as SQLite databases, so I fired up Lita in order to understand what the structure was like… Things looked a bit messy but I eventually found everything that would interest me; and it was in the “MessagesFT_content” table. Here’s the query I ran:

SELECT c1Body FROM MessagesFT_content WHERE c0Subject LIKE '%Chat%'

Almost cool. The query still returned a bunch of HTML code, our names, and other useless crap. So I fired up Flash Builder, imported the SQLite file and wrote a few AS3 lines, in order to grab the results and filter them with regular expressions. Bang: plain text! Oh, this useless AIR app is OpenSource, by the way.

Now that the data was clean and ready to be analyzed I had to find a cheap or free way to do it. I chose Primitive Word Counter, not because it’s perfect but rather because it’s very simple and could handle the large amount of data I was going to feed it (some other apps simply crashed)…

Running it gave me the most used words and phrases, I only picked the most interesting (at least to me) and launched InDesign.

A celebration poster

I decided to go for an A1 poster, mostly focused on those words and phrases but with a tech twist to it. I kept it all secret, got it printed, and sent it to my pal… Happy fifth Google-talk-history-enabled anniversary to him; and to all of you out there that use it on a daily basis!

Six month ago I’ve been contacted by an architectural firm to provide some consulting on a project of theirs (I’m not going to name names, you’ll understand why). The goal was to find ideas to make a building’s front more interesting. The building being a place to help and promote Hip-Hop culture.

So I started working on it and came up with ideas and concepts. The architect I was in contact with seemed pretty happy with it and everything was looking good.

Until I no longer received any answer to my e-mails… Our last interaction is now 5 month old and I think time has come to mourn. What I came up with can be interesting and since it involves an OpenSource project, here are a few bits about it.

At that time I was discovering  GML (Graffiti Markup Language) and Evan Roth’s work. Bordeaux hosted Les Grandes Traversées and all of this really inspired me. So I thought of a mash-up between GML’s #000000book (black book, open archive of GML tags), a player of my own (Okr), the building itself and Twitter. Here’s the document I presented to explain what I had in mind.

The steps are:

1. Creating and sending a graffiti;
2. Receiving data;
3. Converting it to an image;
4. Projecting it on the building’s front;
5. Photo-shooting of the front;
6. Sending to Twitter;
7. Online consultation.

After a few e-mails with Jamie Wilkinson (heads up!) I started working on the core classes writing GMLPlayer and GMLCreator. The goal was to provide both a way to display tags and to create/upload them. I then built a UI around all that (a Flex one, after noticing Minimal Comps didn’t work the way I expected).

iframe: <a href="http://toki-woki.net/p/Okr/">http://toki-woki.net/p/Okr/</a>

Note: you’ll also find the app on its dedicated page. Try searching for “dasp” or “hello world” for example and play with the settings (the 3 top sliders).

Unfortunately it is only after creating all this that I realized the project would never become real… So I simply stopped working on it. I am well aware that some parts of the code is a bit raw and could be optimized and I haven’t built the creation/upload feature into the UI yet. Don’t know if I will, but the project is OpenSource so feel free to give it a spin! I also share my initial attempt and a pixel version in case you’re interested.

Pretty happy that — even if not feature complete — Okr made it to the GML project gallery, yay!

And just because a project will never see the light of day doesn’t mean it doesn’t need a proper logo, right?

To avoid breaking things you have to create an intermediary app version that will smoothly switch from 2.0 to 2.5, here’s what you can read on the release notes page:

In order to be able to update from version 1 to version 2, an intermediary update step must be added as follows: application version 1, packaged with AIR 2 and using the 2.0 namespace gets updated to: application version 1.5, packaged with AIR 2.5 and using the 2.0 namespace. This version of the application must include the version of the Application Updater SWC/SWF included with the AIR 2.5 SDK. This gets updated to: application version 2.0, packaged with AIR 2.5 and using the 2.5 namespace.

Where “application 1.5″ is the intermediary step.

All of this looks quite simple but really it isn’t; or at least it wasn’t for me. To be really explicit here are my 3 application descriptors and the update descriptor (version numbers changed to match Adobe’s example).

Application descriptor – Version 1 (packaged with AIR 2.0):

<?xml version="1.0" encoding="utf-8" standalone="no" ?>
<application xmlns="http://ns.adobe.com/air/application/2.0">
	(...)<version>1</version>(...)
</application>

Application descriptor – Version 1.5 (packaged with AIR 2.5):

<?xml version="1.0" encoding="utf-8" standalone="no" ?>
<application xmlns="http://ns.adobe.com/air/application/2.0">
	(...)<version>1.5</version>(...)
</application>

Application descriptor – Version 2 (packaged with AIR 2.5):

<?xml version="1.0" encoding="utf-8" standalone="no" ?>
<update xmlns="http://ns.adobe.com/air/framework/update/description/2.5">
	(...)<versionNumber>2</versionNumber>(...)
</update>

Update descriptor (PHP script receiving the caller’s current version as “currentVersion” GET variable):

<?php
header("Content-Type: text/xml; charset=utf-8");
echo '<?xml version="1.0" encoding="utf-8"?>';
$currentVersion=array_key_exists('currentVersion', $_GET) ? (float)$_GET['currentVersion'] : 1;
$isNewNamespace=$currentVersion>=2;
$ns='http://ns.adobe.com/air/framework/update/description/'.($isNewNamespace ? '2.5' : '1.0');
$version=$currentVersion>=1.5 ? 2 : 1.5;
$versionTag=$isNewNamespace ? 'versionNumber' : 'version';
?>
<update xmlns="<?php echo $ns ?>">
	<<?php echo $versionTag ?>><?php echo $version ?></<?php echo $versionTag ?>>
	<url>http://www.your-site.com/update/your-app-<?php echo $version ?>.air</url>
</update>

Not that simple, right? And this is not only annoying to the developer, but also to the end user. He will be notified of an update from version 1 to 1.5 and when he’s done he’ll get prompted about the new-new version (2): bang, another update process.

If you’re curious of how I send the app’s version to the update descriptor, here it is:

_appUpdater.updateURL='http://www.your-site.com/update/version.php?currentVersion='+App.getVersion();

The App class is available in my as3bits repository.

Some helpful links on the subject:

• Update AIR 2 or earlier to AIR 2.5 Application Updater framework
• Updating AIR applications
• Release notes for Adobe AIR developers 2.5

I hope this helps!

I wrote a quick and simple golden ratio tool (dedicated page). Basically it helps you find “golden ratio neighbors” for a given number: every number in the list divided/multiplied by its neighbor = φ.

Pretty straightforward but could come in handy. I’m aware it could be improved; if you have suggestions…

When I moved in I bought an IKEA Ramvik table and while travelling this summer I had an idea (don’t ask why): decorate its top with Lego bricks used as pixels. Here are the steps I went through. If you don’t care about those steps and want to see a nice time-lapse video, scroll to the end of the article!

Lego bricks

First things first. What are the Lego brick sizes and colors available? Oddly enough this question is not that easily answered. Probably because Lego’s site is crappy, or because nobody really cares… I eventually found Brickipedia which happens to be a much richer resource than the official ones. Everything I was looking for was there: the Lego “unit” is 8 millimeters and the color palette is pretty simple.

Table specs

Knowing my table size I had various options, depending on the “pixel size” I’d choose. Of course the number of bricks (and the price) would also vary. So I created a dynamic spreadsheet on Google Docs that’d do the calculations for me… Here it is, with all the options possible (French, sorry).

iframe: <a href="https://spreadsheets.google.com/pub?key=0AmyvU-n2aOBpdG5oNUN6UW0xcUJ4a2E3eXFrTTJIVnc&amp;hl=en&amp;output=html&amp;widget=true">https://spreadsheets.google.com/pub?key=0AmyvU-n2aOBpdG5oNUN6UW0xcUJ4a2E3eXFrTTJIVnc&amp;hl=en&amp;output=html&amp;widget=true</a>

I chose the 4x4 option, quite cheap and still offering a cool number of pixels.

Design

36x16 pixels of freedom, that’s it. I tried lots of different designs, from lo-fi photos to pixel-art drawings. I decided to go for a Heavy Oblique Futura.

Actual size:

Looking good.

The Lego palette

I set my type to white, on a black background. The anti-aliasing process creates gray-scale pixels to smooth the curves, which is great, but Lego bricks aren’t available in all colors! To have a realistic preview of what it would look like I had to create a Photoshop Color Table matching Lego’s gray-scales (if you’re interested, just ping me [UPDATE: here they are]). Here’s a comparison between Photoshop’s default gray-scales (left) and Lego’s palette (right):

You may notice that Lego’s black is a little bit light and the grays are yellowish.

Time to order bricks!

Already? Nope, not that fast. Before ordering I had to know exactly what to order, that means counting the pixels. Well, I’m not this kind of guy. I’m a developer; I hate repetitive chores, you know.

So I fired up Flash Builder and came up with PaletteCounter a simple, OpenSource, app to count pixels of each color. I also added some kind of “assembly instructions generator” to help us build it. Handy.

Time to order!

Really? Yup. I placed an order on lego.com’s Pick A Brick and received it a couple of weeks later. Yay!

Let’s do this

I’m not going to describe the process (that happened this saturday), just have a look at this time-lapse vid. 1020 pics shot in about an hour, yummy. Thanks to Céline and Julie for helping out!

I remembered hearing — a few months earlier — about Raphaël (a JS lib to render vector graphics). At that time I found the concept really interesting but couldn’t see where to use it.

OK, so I know what I want to do, I know the tools I’ll use, but how do we do that? Raphaël is cool, but its syntax looks a bit cryptic at first. It says it handles SVG, and — guess what — Illustrator can export to SVG. We’re getting closer…

SVG is XML, but Raphaël doesn’t eat XML, it eats a specific syntax meant to be compatible with SVG and VML. Hum… This is when I fire Flash Builder and start writing code. The goal: accept an SVG XML file as input and spit some “Raphaël-compatible script” out.

Boom (plus a little demo of an output).

I then simply had to export my Ai logo to SVG, open it in SVG2RaphaelJS, copy paste the output, and wire all that to the scroll/color tweening logic. Pretty easy.

Everything worked and looked nice on all browsers and OS’s but I realized the sole idea of having a logo changing colors didn’t work that well. For purely visual reasons. So I dropped it from my theme and rolled back to the black PNG version you can see right now.

The good thing is, the tool works pretty well (at least with Illustrator SVGs) and is OpenSource. Either use it or modify it, and tell me what you think. I did this in a hasty fashion and I probably won’t use it a lot, so don’t expect updates or bug fixes… But if you feel like improving it, please do so!

An intro

We create and distribute pedagogical simulations and videos, and our school users can download them to prepare and illustrate their lessons. Until now we offered server-side generated ZIP files with an HTML, some CSS and the SWF, but users (who often are not computer geeks) wanted something simpler and more powerful… So we designed the eBox!

The eBox

Basically the eBox is an empty media library waiting to be filled. The first time you launch it, it installs itself (and the AIR runtime, if not already there) and then fills itself with the medias you chose. This happens thanks to AIR’s BrowserInvoke logic and removes the “Save As” and “Uncompress” steps.

You can also fill it with “local medias” (of any type) but dragging and dropping files on it, or by browsing and selecting them. If these medias are SWFs they’ll be opened within the app, if not the OS will open them with their default app. Handy.

You also have the ability to create directories to organize your library, and to reorder medias (via drag and drop). Pretty common, but cool.

Let’s talk tech

Flex 4.1, localized with Lupo (now free and OpenSource!) and designed with Illustrator. Server-side communication is made through AMFPHP. I used some tiny techniques that helped deploy two different versions with only one project (we have a school and an individual version), I might detail that in a next post.

This is my first real-life project with Robotlegs, it helped me learn how to use it and how to write really clean code; with view, mediators and all. I loved it and think I will continue building big projects with RLs, makes you feel pretty.

So?

I’m pretty happy with the final result (not that final, expect updates!), both on visual and technical points of view. Go grab a free media and tell me what you think! We also set up a special page with a nice presentation of how it works, in case you’re lost.