Qt Quick by Alexis Menard at Cybercom Developer Day 2010 part 1
Transcription Whisper (large-v3), avec identification des locuteurs. À recouper avec la source d'origine.
Analyse automatique
Générée par IA — peut contenir des erreurs. Méthodologie
Chapitres
Répartition du ton (temps de parole politique)
Propositif 100 %
Affirmations vérifiables (citations exactes)
- 2:29PrésentateurFait
« Qt is quite old. It's an old toolkit, right? Back in time, when you think about Qt, that's what you think. »
- 3:34PrésentateurFait
« it's a component-based architecture. And even though we said Qt is cross-platform, it's not really cross-platform in the sense that you can take your program, let's say, on the computer and put it on the device. »
- 4:11PrésentateurFait
« I don't know any designer that actually speaks C++. »
- 6:09PrésentateurFait
« it's still, you know, based on top of the foundation of Qt. »
- 8:53PrésentateurFait
« QML is a language. It's called Qt meta language, a Qt meta object language. »
- 9:23PrésentateurFait
« It's very inspired from JavaScript. It's actually really similar to JSON syntax in JavaScript. »
- 11:23PrésentateurFait
« let's provide basic primitives, like I don't know, basic geometry primitives, like rectangles, pixel map, rectangle pixel map and small geometry form factor. »
- 12:14PrésentateurFait
« We support states and animated transitions. »
Temps de parole
- Présentateur89 %
- ?11 %
≈ 79 interruptions / 10 min (chevauchements et interjections détectés).
Modèle mistral-small-latest · prompt v1 · les citations sont vérifiées mot pour mot dans le transcript ; le taux de réponse directe est calculé à partir des verdicts question par question. Aucun label idéologique n'est produit.
So,
hi, welcome everybody. First, I'd like to thank Cybercom for inviting me here. I'd like also to thank Marco and Tanya for the organisation. Everything went smoothly, so perfect.
So
today I'm going to talk about Qt Quick and a bit showing what it is and going into details and how you can use it.
So,
the agenda for today will be a bit of an introduction. The opportunities with Qt Quick, then we'll meet QML, it's a language.
I'll
talk about the tooling support for QML
and
a bit of the language basics, how it works, and how you can integrate already existing code into QML
or
native components.
And
I'll
give
an example as well, which is KDE Plasma. So, about me. I'm a Qt developer since 2008. I'm working in Oslo.
During
two years, I was mainly working on QGraphicsView, which is the canvas in Qt.
For
people that follow a bit the news of Migo, it's the framework that is used by the Migo Touch UI library.
I
was involved in QML development. I mainly work on the integration with GraphicView, so low-level stuff in QML, not necessarily high-level.
I
recently moved to Qt WebKit and WebKit development. So, right now, as we speak, I work on the HTML5 video and audio element support in Qt.
My
KDE background, so I'm a KDE developer as well. I started in 2006 on K-Plateau. It's a project management tool. It was a student project, and since then,
I'm
stuck in KDE, and I work mostly on Plasma. I'm not as active as I used to be, but
I
still stay around.
So,
moving to Norway was fun, because I left the French fast food for the Norwegian gastronomy,
which
I heard is not far away from the Finnish gastronomy, too. So, no, it's a good job. So, I like my job. I like Norway. It's just joking. So, let's talk about a bit of the past of Qt. So, Qt is quite old. It's an old toolkit, right? Back in time,
when
you think about Qt, that's what you think. This is KDE, right? So, regular UI on a desktop, buttons, menus, and everything.
The
problem is the market is moving, and we have those kind of things. And you don't have a mouse. You have a finger, and a finger is less accurate than a mouse.
So, if
you compare this and this, you can tell that this user interface will not work on a small form factor like this one.
Not
because it can't run. It's because it's not well suited for it.
So,
as I said, Qt comes from the desktop world.
It
was initially designed to build a UI on top of Motif and Windows
15
years ago.
So, at
that time, you know, having something on a seven inches or less screen was not in the mind of people. So,
also,
it's a component-based architecture. And even though we said Qt is cross-platform, it's not really cross-platform in the sense that
you
can take your program, let's say, on the computer and put it on the device. It will run. That's for sure. If it's a supported platform by Qt, it will. But, as I said,
it
will not be optimal.
So,
you still need to tweak the UI.
So,
it's not a really cross-platform toolkit in the sense that you always have to tweak a bit. Also, if you look Qt, Qt is C++, right?
I
don't know any designer that actually speaks C++.
I
mean, like 100% designers that actually speak or write C++ code or understand C++.
So,
what we see is, like,
the
designer actually designed the UI of the application, but the developer actually implements the user interface, and it's always different than what the designer asked for. Or,
they
implemented less because they couldn't. Performance reason, time problems, time issues, and stuff like that.
So,
the fact that the designers speak another language, and the fact that they use other tools, let's say Flash, for instance, or even PowerPoint. I've seen PowerPoint used for designing user interface.
It
makes the collaboration difficult between both worlds. And when you create a user interface for a phone, now, the user interface matters a lot.
Like,
it has to be fancy, it has to be nice, otherwise, you won't sell it.
You
won't sell your device. Also, if we look again, the imperative languages like C++, it's a bit counterproductive, in a sense that it requires many steps to reach what you try to achieve. Like,
if
you take a C++ framework, you need to inherit from the base class, or to actually customize them, and so on, takes a lot of time.
And
you always struggle with the C++ or the language itself, than actually describing what you really want to have on the screen, the look and the feel. That's something that is not easy to do today in Qt. Also, we see that canvas-based architectures become prevalent. Qwidget,
the
way the widget work in Qt,
is
not based on the canvas.
It's
based on the old approach. We still have a canvas, where, as I said, because I'm working on it, it's called Graphic View,
but
it's still, you know, based on top of the foundation of Qt.
So
it's not the best canvas that we could have in Qt.
Like,
let's say, compared to a canvas in the iPhone, for instance, or, I don't know, Android, probably.
So with
the canvas, you can do animations, fluid transformations, like nice effects, and so on. You can talk to the graphic card and, you know, add some fanciness in your UI.
And
this is important because
for
us, for Qt, for the people that use Qt, it's important they can actually create UIs that are different than the others. So they can actually sell the products and say, hey, look, mine is different, right? A bit like they do on top of Android. Oh, here is my UI layer on top of Android. It's better than this one. And
so
we still think that you should be able to customize your UI quickly and do some fancy stuff on it. Also, if you look at the market today,
the
hardware capabilities of, let's say, a small smartphone and a computer, they're slowly merging.
I
mean, who thought about, like now, for a recent phone, around the one gigahertz processor? Who thought that it would happen two years, three years ago? So, and if you look back in 1999, probably something like that, or 2000, like desktop computers had the one gigahertz processor, right? So the hardware are merging. Graphic cards are on mobile phones getting more features and features and features.
And so
the hardware capabilities of both a computer and a mobile smartphone, they are slowly merging.
But
the UI and the four factors are completely different. My laptop is a keyboard.
The
phone is a touchscreen. Who knows what will come next, 3D user interface, or I don't know. So, but the hardware capability is converging. So what we could run on the desktop now, we can actually run quite easily, on quite easily, but we can, we could run it on the phone quite easily.
But
still, the UI
will
be different, just because the screen, the form factor will be different. The way you interact with it will be different.
So
that, so that's, that's actually the background of Qt. So now how we can solve that? Well, we think that what we add in Qt does not fit well, mobile development.
We
are in a transition phase.
We
slowly move to a canvas with Graphic View.
That's
what Nokia is using for MIGO.
But
we still think it's not the optimal approach. So, and that's why we're working on something that we call QML. And QML is a language. It's called Qt meta language, a Qt meta object language. And the marketing name of all that project is actually Qt Quick. It's a number in a term for the declarative module in Qt, which is QML. And creator, which is our tool for it, our IDE.
For
us, we, it's a new way to create UIs in Qt. It's nothing to do with the old designer. It has nothing to do with C++. It's a declarative language.
It's
very inspired from JavaScript. It's actually really similar to JSON syntax in JavaScript.
We
also support JavaScript. You can do JavaScript inside.
You
can import libraries from the web.
It's
intuitive. It's really fun to use. I'll show you afterwards.
And
it has a similar structure to HTML, but it's not the same.
You
create parents and then children, and it's an object tree that you create, actually, a bit like HTML. For the fun story, back in time when we were starting to work on QML, QML, the language itself was not JavaScript or JavaScript JSON. It was actually HTML
or
XML, to be honest. And
after
hacking demos and everything, like when the file started to be really big, we said, wow, it's insanely hard to read. XML is not human readable language.
So that's
why we changed the syntax and now we use JavaScript JSON, because it's easier to read and to understand what's going on in the code.
The
approach we have also is completely different than what we had in QWidget. QWidget is, for instance, you get a checkbox
with
everything into it, all the features you can imagine of a checkbox. And
what
happened is like it creates
huge
fatty components. And sometimes people say, yeah, but I don't need all those features. What I want is just a checkbox with a cross and I can click on it. I don't need auto layout. I don't need other features like that.
I
don't need to be able to style the buttons and so on. So
the
feedback from customers of QWidget or users of QWidget was pretty much that it's actually hard to extend the existing components
to change
them the way they want.
So
that's why we said, okay, let's do the other way around is, let's provide basic primitives, like I don't know, basic geometry primitives, like rectangles, pixel map, rectangle pixel map and small geometry form factor,
I
would say. And then you build your button.
So
that means that you can actually build it like you really want
by
adding primitive one afterwards, like a rectangle is a pixel map with the text on top of it. So
we
create components from simple other elements and then each item is independent. You can access to each sub-item independently.
For
now, the canvas in QML is QGraphicsView.
So
you can do a fancy animation. You can also reuse the code that you have on GraphicView and then you can bring it to QML.
We
support states and animated transitions. So you can define states on the UI, like, for instance, portrait landscape. And you can decide how the transition will happen from the portrait to landscape.
You
can configure, as I said, the interpolation between the properties and everything.
So why
Qt Quick? Recode. It's to ease the question of animated user interface, something that we didn't have in Qt. It's to bridge the gap in Qt for mobile development. We want to be able to create quickly applications and put them on, let's say, OV store.
We
focus on performance. It's designed since day one for performing well on devices.
It's
not necessarily for device, for mobile or embedded development. We think that it will have a use case on the desktop and we hope that at some point we'll bring features so that you can actually use it as well on the desktop and get rid of Qt slowly.
It
increases the productivity a lot. You can create custom items really quickly and it's completely different than C++ where you need to subclass and everything. And what works really well is since it's an easy language to read, designers are able to understand. And we have a lot of success stories of actually designers doing working prototypes.
They
create the design and they actually hack it into QML and then it's actually a working prototype they can give to a developer and say, okay, now you have to implement the logic. The UI is made but implement the logic.
That
happened in KDE.
That's
what, so that's the last one. Empowers designer to actually create real application instead of a flash that's actually is useless for a developer which can't do anything with a in
KDE. That's all fresh. You You