How Qt could bring better third-party software to Ubuntu (ars technica) [LWN.net]

Ars technica looks at the advantages of the Qt toolkit. "A point that I think often gets overlooked in the toolkit debate is that adopting Qt doesn't necessarily imply ditching GNOME or switching to KDE. As we discussed in our review of Qt 4.5 last year, Qt has relatively robust support for Gtk+ theming, including conformity with the GNOME HIG and support for native GNOME dialogs. When everything is properly configured, Qt applications look entirely at home in GNOME environments. Adding a standard Qt library stack to a fresh Ubuntu installation requires only 16.5MB of packages, which expands to approximately 50MB on disk."

to post comments

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 20:34 UTC (Fri) by JohnLenz (guest, #42089) [Link] (32 responses)

I think a stronger backing behind Vala has a better chance of drawing 3rd party development to Linux and Ubuntu than QT with C++. Perhaps we should highlighting two options to new desktop linux programmers: use the scripting language python or the compiled language Vala.

Lets face it, most 3rd party deskop developers are going to be coming from Windows C#/VB.NET programming. Vala is a much better fit for them than trying to get these programmers to learn C++. Vala has a comparison on their web page showing how the features of Vala match C# and Java. With Vala, they get 90% of what they are used to in C#.

Vala vs. C++

Posted Oct 22, 2010 21:00 UTC (Fri) by ncm (guest, #165) [Link] (1 responses)

... and Vala experience looks so much better on your résumé.

Vala vs. C++

Posted Oct 22, 2010 22:42 UTC (Fri) by salimma (subscriber, #34460) [Link]

I tend to list frameworks (Java, .NET, GNOME GObject) separately from languages. And pair relatively obscure languages with the better-known language they are associated with (e.g. Java/Scala, Lisp/Scheme/Clojure). Since Vala and C# are very similar languages, it would not be misleading to lis them together in the same way.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 21:28 UTC (Fri) by flammon (guest, #807) [Link] (28 responses)

How about JavaScript?

http://live.gnome.org/Seed

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 22:18 UTC (Fri) by BenderBendingRodriguez (guest, #60391) [Link] (27 responses)

Doesn't Qt also have this? I mean QML. What's the difference? All in all i'd prefer that there would be one toolkit, less fragmentation among developers, less confusion among users ("you want GNOME or KDE?" "What? I'm staying on Windows...")

Also in the light of recent changes in the way qt is being developed (more openly) it would be best for everyone to embrace Qt considering it is backed (for now) by Nokia and definitely won't dissappear in a year or two.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 5:14 UTC (Sat) by drag (guest, #31333) [Link] (25 responses)

If QT becomes embraced by developers and people start to produce applications that users desire then QT may become vogue. Who knows. The only way it's going to happen is through a natural selection by programmers and users.

The way things are now Gnome has made large improvements and I am quite happy and comfortable with my Gnome system and applications. There is a tremendous amount of value in all of that and I don't see why anybody would want to give that up now just because somebody says that one toolkit is funner/easier to program with then the other (regardless on how true that is).

As far as the mobile space goes it seems very likely that Nokia has missed it's boat. They had a window of opportunity there and I am afraid they blew it. They don't even have one product out yet using Maemo and there does not seem to be anything coming out in the near term either. By the time they get into gear they are going to be so far behind the curve they may have almost no significant impact at all.

It's hard to express in words just how much momentum Android now has in the market place. It may not seem like it right now if you look at what is out there right now on the store shelves, but after looking around and seeing reports coming out of various conventions and whatnot; Android is going to very quickly become the de facto standard in phone operating systems and that corporations are looking ahead to the day when the mobile devices with web-based applications become dominate. The only people that have a chance of stopping that is going to be Microsoft.

It looks more and more likely that Java and Javascript is going to be the wave of the future on mobile devices; despite how ugly that sounds. It's either that or .NET/silverlight.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 9:05 UTC (Sat) by MKesper (subscriber, #38539) [Link] (4 responses)

Hmm, what about the n900? Or did you mean meego (N9 delayed...)?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 13:19 UTC (Sat) by drag (guest, #31333) [Link] (3 responses)

> Hmm, what about the n900? Or did you mean meego (N9 delayed...)?

That's one of the issues. They started shipping a working OS and basically stuffed it all in order to go with QT. Moving away from what they had that worked may have been a big mistake and lowered their chances considerably.

Maybe QT is so nice that it was worth the sacrifice and will make up for it in the end. But fundamentally users and customers (the people that make Nokia money) don't care about toolkits. Maybe it will pay off. It certainly is a big risk.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 10:05 UTC (Mon) by buchanmilne (guest, #42315) [Link]

That's one of the issues. They started shipping a working OS and basically stuffed it all in order to go with QT.

It's not all-or-nothing, Qt is available and updated on Maemo (pr1.3 will probably ship Qt 4.7 with Qt mobility APIs, QML etc.).

Moving away from what they had that worked may have been a big mistake and lowered their chances considerably.

Moving to Qt is only part of the difference between Maemo and Meego. One important aspect is the entire open-sourcing effort, and ensuring that all components required for the device to work are either open-source, or distributable (firmwares for some components), and upstreamed first. This is important if other vendors are going to ship Meego devices, which is important to increase the target market size.

Maybe QT is so nice that it was worth the sacrifice and will make up for it in the end. But fundamentally users and customers (the people that make Nokia money) don't care about toolkits. Maybe it will pay off. It certainly is a big risk.

Meego means developers will have a bigger market than just Maemo (which isn't very big at present), but devices by other manufacturers in a variety of applications.

Qt provides a platform for developers to target all of Nokia's platforms (except Series40, which most developers probably wouldn't target anyway) and Symbian 3 provides a migration path from old Symbian toolkits to Qt.

Users don't just flock to a platform with no applications, Nokia is taking care of their existing developers, and building a bigger market, and reducing the effort developers need to spend to reach a larger audience. And they're doing this with an open-source project that provides benefits to users of other open-source platforms.

Do you really think GTK+clutter on Maemo and legacy Symbian-specific toolkits on Symbian 3 would be a better situation for anyone?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 10:25 UTC (Mon) by jamesh (guest, #1159) [Link] (1 responses)

I don't think you can really comment on Nokia's Qt on Meego strategy without also looking at their Qt on Symbian strategy.

Looking at only Meego, the decision looks crazy. While the external developer community was small, replacing the graphical toolkit essentially rips the rug out from under their feet. It has also likely set the internal developers back as they learn the new toolkit and rewrite applications.

If you bring Symbian into the picture though, the move means that there is a single source-compatible application development platform that spans both their Symbian phones and Linux phones. This has the potential to increase the developer base for both platforms, and make it easier to migrate users and developers off of Symbian in the future.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 9:43 UTC (Tue) by buchanmilne (guest, #42315) [Link]

replacing the graphical toolkit essentially rips the rug out from under their feet.

The graphical toolkit hasn't been replaced, the default has been changed. Both GTK+ and Qt have been available in Maemo5 for some time. GTK+ will be present in Meego. Even hildon will probably be available in Meego (AFAIK gpodder works fine on Meego, so I think it is present in current Meego), but not an officially supported API. The recent Maemo5 update (PR1.3) provides compatibility with Meego APIs.

Other vendors (Apple, Microsoft) are allowed to deprecate APIs (seemingly for little benefit to developers), why are people complaining that Nokia is doing it (to provide a single API across more platforms)?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 9:22 UTC (Sat) by oak (guest, #2786) [Link] (9 responses)

Why do you think Microsoft to have so much better chance than Nokia?

Nokia is shipping much more devices (than anybody else), has the operator relations (at least elsewhere than in US) and the just released new (Symbian) phones (with estimated sales in tens of millions) include Qt. Qt Creator has gotten pretty positive press + they just announced the support for Qt Quick thing.

What they'll need is the service side getting better (Microsoft probably has upper hand there), but at least OVI has been stated to have over hundred million users so I assume it (and its phone & desktop integration) to eventually get on par with its competitors... Its current advantage over other services is support for carrier billing.

Although Apple & Android have head start, I think the race is just beginning, not over.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 13:14 UTC (Sat) by drag (guest, #31333) [Link] (8 responses)

Symbian made a lot of headway in the past because of third parties licensing it for their phones. But most of that is gone now, except for the low-end smartphones because those phones are too slow and too limited to run Linux well. However in the next year or two even the very low-end phones will be able to run Android and the handset makers are anticipating it.

As low-end hardware gets capable enough to run a Linux system then your going to see Symbian evaporate. It'll stick around for a long time, but it's going to be at a evolutionary dead end.

If Nokia is able to get something reasonably quickly AND works very hard to aggressively market it to other handset carriers then it might have a chance still. However I feel that Microsoft is the one that is being much more aggressive so they will have a much higher chance of actually being competitive.

That's it. It's a race; aggression and a impressive user experience is key to getting a leg up on Android and iOS. I think that Nokia had a chance, but right now Microsoft is in a better position and is willing to do what it takes.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 0:01 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (7 responses)

No, Nokia is playing smart.

Their Symbian OS is, frankly, a piece of crap. It's getting better with each new release, but it's still stuck in 90-s. So they brought in QT and it now allows developers to write applications for Symbian _and_ full-scale Linux simultaneously. And Nokia produces a _very_ large number of devices for this to be tempting.

Android is still nowhere close to being cheap. The low-end communicators with Android are still quite expensive and it'll take them at least 2-3 more years for prices to fall enough to be competitive with Symbian phones.

So we have two strategies: Android begins at high level and grows down, while Nokia starts at the bottom level and grows upward. I think in the end they both will be successful.

And yes, Nokia _had_ to drop GTK and rewrite everything in QT. Because GTK is crap.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 0:35 UTC (Sun) by drag (guest, #31333) [Link] (6 responses)

> No, Nokia is playing smart.

No, I think that their organization is not terribly well thought out with different departments purposely pitted against one another in a corporate management style similar to Microsoft, which is just horrid. It makes it difficult because instead of having people work together they are competing.

But that's just a opinion. If you disagree with me I won't take it personally.

> Their Symbian OS is, frankly, a piece of crap. It's getting better with each new release, but it's still stuck in 90-s. So they brought in QT and it now allows developers to write applications for Symbian _and_ full-scale Linux simultaneously. And Nokia produces a _very_ large number of devices for this to be tempting.

Why on earth would you care about programming on Linux and Symbian if Symbian is such crap? Why not just make everything Linux? That way you don't have to give a crap about having to maintain 2 entirely separate operating systems with layers of abstractions so they can only-sorta have the ability to run the same software?

> Android is still nowhere close to being cheap. The low-end communicators with Android are still quite expensive and it'll take them at least 2-3 more years for prices to fall enough to be competitive with Symbian phones.

Right now there are several phones on the market with 300-400 dollar price tag, retail.

Some cheaper, for example:

http://www.huawei.com/news/view.do?id=11162&cid=42

Sprint is offering a 150 dollar Android phone, very soon... contract free.

http://phandroid.com/2010/10/06/huawei-ascend-150-android...

That's _cheap_ for a smartphone.

Here is the LG GT540 Optimus, which should be around 240 dollars for a unlocked verison according to the article:
http://news.softpedia.com/news/LG-GT540-Review-162490.shtml

For lots of other examples Google around for 'Entry-level Android' in Google news. ;)

One of the big boons for Android in Asia is going to be China's clamp down on so-called KIRF phones, which are look-a-likes of more expensive phones. This has left many companies with the choice of trying to figure out to make their own real smartphones and Android is the key for them. Sure most of it is shit now, but successful companies are going to rapidly improve things for themselves. Companies like Qualcomm are starting to work with them so they can crank out Android platforms for as little as $75 (probably 140-180 by the time it reaches customers, I expect) a pop. And Qualcomm seems to only have eyes for Linux anymore. (just a impression I get)

Sure Symbian will drop in price, also, but one thing to keep in mind is that these smartphones are specifically targeting Linux in order to target Android. Symbian will have to be ported to these platforms were Android is going to support them out the gate, more or less. At least by their chipset manufacturers.

The big thing I am hoping for is that these unlocked Android phones for 200-500 dollars will be widely available so that people can afford to purchase them outright without having to deal with carrier's restrictions. I think then we might see some significant activity with not only third party Android roms like Cyanogenmod, but also Maemo and a resurgence of other attempts at making Linux phone-friendly in the past.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 15:46 UTC (Sun) by nix (subscriber, #2304) [Link] (3 responses)

No, I think that their organization is not terribly well thought out with different departments purposely pitted against one another in a corporate management style similar to Microsoft, which is just horrid. It makes it difficult because instead of having people work together they are competing.

OK, so a few posts ago you were all libertarian, and now you come out with this, which a Soviet central planner would have agreed with... I am confused.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 16:15 UTC (Sun) by mpr22 (subscriber, #60784) [Link]

Competition between companies is qualitatively different to competition within a company, particularly of the "battling business units" variety.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 16:18 UTC (Sun) by drag (guest, #31333) [Link]

> I am confused.

If you understood what I was about then you wouldn't be. ;)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 4:23 UTC (Mon) by eru (subscriber, #2753) [Link]

Have a look at this: http://www.hs.fi/english/article/Knock+Knock+Nokias+Heavy+Fall/1135260596609 (an English version of a well-researched article in the latest monthly supplement of the largest Finnish newspaper).

One gets the impression that Nokia now seems to be a case study of how to screw up a succesful company by bad reorganization.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 16:16 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link] (1 responses)

"Why on earth would you care about programming on Linux and Symbian if Symbian is such crap? Why not just make everything Linux?"

Symbian is orders of magnitude simpler and requires far less resources. It is geared towards low-RAM environments and works really well there.

"That way you don't have to give a crap about having to maintain 2 entirely separate operating systems with layers of abstractions so they can only-sorta have the ability to run the same software?"

Right now, a Symbian phone can be bought for like $30. Android is not going to be that cheap any time soon. Also, don't forget non-touch phones (Android sucks there). So Symbian is going to be slowly scaled down, until it disappears. But it'll take some time.

And on the high end, QT allows to write programs for tablets, smartphones and PCs.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 16:28 UTC (Sun) by drag (guest, #31333) [Link]

> Right now, a Symbian phone can be bought for like $30.

I don't think you'll find a S60 Symbian for that, retail. Maybe you can find one for 30 dollars if you combine it with a rebate and a 2 year contract, but I was talking about retail prices.

If you have links I am all eyes.

If your talking about Series 40 operating system, then it's possible you can find a phone like that if you find a sale or a really good deal.

But S40 is a different beast entirely. It's a couple generations old and really is in the feature-phone category. S40 and Android are not competitors, really,

> And on the high end, QT allows to write programs for tablets, smartphones and PCs.

(So does Java.) That's all fine in theory, but I doubt your going to see to any programs of any significance that are running on Symbian, Maemo, and a PC without a significant amount of effort going into maintaining/porting all 3 different versions. Not unless your happy using a mouse on your PC to interact with a application designed to work on a 2.5 inch 320x240 touch screen.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 17:04 UTC (Sat) by anandrajan (guest, #146) [Link] (9 responses)

The GENIVI alliance has decided to go with Meego for in vehicle information and entertainment. This is a win for Nokia/Intel and they could build on this first win. While Android has huge momentum in the smartphone space, it remains to be seen if Android/Chrome will win on tablets/netbooks and other wireless devices.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 18:36 UTC (Sat) by drag (guest, #31333) [Link] (8 responses)

Yes. Android is at a disadvantage with Tablets. It won't be until Android 3.0 before Google has a OS that it intends to compete with iOS4 on the ipad.

This is certainly a opportunity.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 21:07 UTC (Sat) by jspaleta (subscriber, #50639) [Link] (7 responses)

Which makes it that much more amazing that OEMs putting Android tablets in the retail market already. That's actually pretty astonishing considering that Google really hasn't made a direct push there. With OEM support like that, the window of opportunity to turn OEM-heads away from Google is pretty small.

-jef

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 21:50 UTC (Sat) by anandrajan (guest, #146) [Link]

With OEM support like that, the window of opportunity to turn OEM-heads away from Google is pretty small.

Yes, this is really weird. It reminds me of what happened circa 1995-1997 with Windows 95. The Android smartphone momentum has clearly spilled over into tablets with OEMs tracking Apple's moves but over in the more free Android playground. However, the tablet space (non IOS) is still wide open at the moment due to Android 2.2's inadequacies. Furthermore, with Nokia's strong presence in Europe/Asia, it is possible that we'll see Meego wins first in that sphere and not in the US.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 8:07 UTC (Sun) by cmccabe (guest, #60281) [Link] (5 responses)

> Which makes it that much more amazing that OEMs putting Android tablets in
> the retail market already. That's actually pretty astonishing considering
> that Google really hasn't made a direct push there. With OEM support like
> that, the window of opportunity to turn OEM-heads away from Google is
> pretty small.

Well, if you look at it from the OEMs' point of view, the design considerations for tablets and smartphones aren't that much different. A tablet is almost like a phone with a bigger screen. It makes sense to run Android apps on a tablet; it doesn't make sense to run Windows XP applications there. Then if you add to that the need to support ARM, the case for Android becomes even stronger.

On netbooks, on the other hand, every OEM does the same thing: they just try to stick the skinniest version of Windows on there that Microsoft will sell them. It's gotten very difficult to find Linux-based netbooks. And it is impossible to find ARM-based laptops. Microsoft's dominance on the desktop prevented that whole technological shift from ever happening. Instead, we have to make do with laptops that run too hot to ever be used on our laps. It will literally burn your crotch if you put a "laptop computer" on your lap. It's a strange world...

To try to get vaguely back on topic... I don't really understand who the "third party developers" are that Ubuntu is going after. Mobile developers? I honestly can't think of any mobile apps I would want to run on my desktop. Maybe other people can think of some?

Also, C++, which QT is based on, has some serious security issues. It's not a managed environment. The designers of Android spent a long time thinking about how to sandbox third party applications effectively. Using Java helped, and there's also a lot of work they did under the hood at the system level to make true isolation possible. On the other hand, I wouldn't want to run "third-party", non-open source code on an Ubuntu installation. It's just so, so easy to get root. Even when there's no an actual vulnerabilities to exploit, you have clueless users running sudo all the time and then executing unrelated stuff in the same shell, and stuff like that.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 13:19 UTC (Sun) by drag (guest, #31333) [Link] (4 responses)

To be honest; the versions of Linux that shipped on Netbooks tended to be absolutely retched.

Linux on the netbook had a very strong chance, but lack of vendor sophistication blew it completely and none of the Linux distributions were anywhere close to aggressive enough to get OEM acceptance on those netbooks. Linux netbooks outsold Microsoft-based netbooks by a wide margin for a while, but once people started using Linux they started to realize how horrible it was and how it would not run any of the applications they needed.

The hardware OEMs tend to have very poor skills when it comes to software. They are very good at taking reference platforms produced by chipset and processor manufacturers, redesigning them enough to be efficiently/profitably produced. But they are not good terribly good at software and drivers and cannot be trusted to provide a good user experience.

They really really really need a standard platform they can provide to customers that will be consistent and have wide ranging applications suppose.

This is something the Linux distribution model did not and possibly cannot provide and is something that Android is doing. Sure the OEMs slap a trashy skin on them and theme on Android phones, but you are not going to run into the application compatibility and usability nightmare that lay people face-planted into with Linux system on a netbook.

*(Ubuntu is may be close, though, and 10.10 is pretty slick. But the better Ubuntu gets the more resentful people seem to get, bizarrely. So I don't know if it ever will be possible to get a traditional Linux distro desktop that will be acceptable at this point to a wider audience. Maybe Google Chrome...)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 13:52 UTC (Sun) by anselm (subscriber, #2796) [Link] (2 responses)

The problem with putting a good Linux distribution on netbooks is that the sort of well-customised distribution (say, Ubuntu) one would want to see there does not come free (as in beer). Somebody will have to spend engineering time to see that all the hardware is supported, all the dialog boxes fit the screen, etc. etc., and all of this must be ready by the time the device hits the market (so no hoping the Debian project will get things right a year or so after the initial launch, by which time the manufacturer will be on the second next model after the one in question). A serious Linux netbook distribution would also require after-market support in the shape of updates and customer support, which is probably more expensive than its Windows equivalent.

On the other hand Microsoft can afford to basically give Windows away for netbooks in order to keep Linux from making inroads in that market. Since everybody except a few geeks would really rather have Windows on their netbooks, anyway, why should a netbook manufacturer risk alienating a majority of their potential customers?

The only way a manufacturer can apparently save money using Linux is by slapping on a Mickey Mouse distribution that barely works and then forgetting all about it once the device is done, which as we have seen does nothing for the credibility of Linux in general. True, a netbook that came factory-preloaded with a well-adapted and well-supported mainstream Linux (such as Debian or Ubuntu) would be very nice to have and might even convince a few more people to give it a try, but on the whole, Linux on a netbook does not seem to be an important selling point right now, which is why nobody seems to be pushing the concept.

Android on iPad-like devices is acceptable now because people want the hardware, because it is reasonably cheap for the manufacturer, and because there is nothing out of Microsoft right now that would be viable competition – not because Android as such is such a great thing. If people had the choice between Android and some Microsoft operating system that worked halfway well on an iPad workalike device, it is a fairly safe bet that most of them would go for the Microsoft option.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 0:47 UTC (Tue) by cmccabe (guest, #60281) [Link] (1 responses)

> Android on iPad-like devices is acceptable now because people want the
> hardware, because it is reasonably cheap for the manufacturer, and because
> there is nothing out of Microsoft right now that would be viable
> competition not because Android as such is such a great thing.

Personally, I believe that Android is a great thing. In a lot of ways, it's ahead of Linux on the desktop. It has a better security model, gets better power usage, isn't difficult for novices to set up. Almost all of the software is written in a managed language (Java) so buffer overflows are a thing of the past. Most of the code is open source. I have developed for Android in the past and it's a good platform.

> If people had the choice between Android and some Microsoft operating system
> that worked halfway well on an iPad workalike device, it is a fairly safe
> bet that most of them would go for the Microsoft option.

People do have the choice between Windows Mobile 7 and Android. And they're mostly picking Android. Anyway, users don't care about operating systems. They care about applications and total cost. The Chinese and Korean OEMs who actually build most of the phones care about not paying royalties.

Nobody ever picked Windows because of its features. You picked it because it was the operating system that ran all of your applications. How is it that people still don't get this, 30 years later... ? They won by being there "firstest with the mostest." (Cue long discussion of DOS and its silly features.)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 1:35 UTC (Tue) by anselm (subscriber, #2796) [Link]

People do have the choice between Windows Mobile 7 and Android. And they're mostly picking Android.

Which isn't all that surprising given that the first devices based on Windows Mobile 7 (or Windows Phone 7, as it's called these days) are only now becoming available, and that Windows Phone 7 isn't specified for iPad-like devices at all – Microsoft markets it strictly as a smartphone/PDA operating system. We shall have to see what Microsoft will be doing as far as iPad-like devices are concerned.

Anyway, according to Steve Ballmer a Microsoft »iPad killer« would be based on Windows 7 rather than Windows Phone 7 (where, as usual, the one has not much in common with the other except for the similar-sounding name), and I'd suggest that this might appeal to people who would prefer their iPad-like devices to support an operating system that they know and that is the same as the one on their desktop. Whether that will work out in reality is, of course, anybody's guess; traditionally Microsoft hasn't been all that successful in anything not to do with PCs (with the possible exception of the Xbox) but they might get lucky once in a while.

Finally, just for the record, I'm all in favour of Android – I have an Android phone and as far as I am concerned it is very nice indeed.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 2:39 UTC (Tue) by cmccabe (guest, #60281) [Link]

> Linux on the netbook had a very strong chance, but lack of vendor
> sophistication blew it completely and none of the Linux distributions were
> anywhere close to aggressive enough to get OEM acceptance on those
> netbooks. Linux netbooks outsold Microsoft-based netbooks by a wide margin
> for a while, but once people started using Linux they started to realize
> how horrible it was and how it would not run any of the applications they
> needed.

Yeah, the OEM Linux installations were always bad. Then again, their Windows installations aren't so great either. You really would be better off just wiping the drive and doing a clean Windows install without all the spyware and adware. Too bad they don't give out Windows install media with the computer any more, bwahaha.

I think the biggest reason why Windows has the netbook market cornered is compatibility. People just really want to run their applications. There were probably some backroom deals made too. Every OEM pays a different price for Windows depending on how much Microsoft likes them at the moment. It's not fair, but that's life.

> They really really really need a standard platform they can provide to
> customers that will be consistent and have wide ranging applications
> suppose.
>
> This is something the Linux distribution model did not and possibly cannot
> provide and is something that Android is doing.

I don't think that Linux is that fragmented. You can run almost every application you can run on Red Hat on Debian, for example, without going outside the package manager.

Unfortunately, the reason why Linux-on-the-desktop is such a bad OS for grandma is that it's such a great OS for developers and enthusiasts. Code gets accepted into projects because it's cool. Testing isn't as much fun so not enough of it gets done. Package management makes it easy to install anything, but unfortunately "anything" includes some very beta-quality stuff sometimes, and Grandma doesn't know what a FileNotFound error is.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 6:50 UTC (Sat) by wstephenson (guest, #14795) [Link]

Qt has the QtScript engine. This is used in Qt Quick to provide logic between the markup language's object.

QtScript can also code complete Qt traditional Qt apps.

Here's an example: http://labs.qt.nokia.com/2009/01/06/simple-qtscript-based...

The code has moved with Qt's move to gitorious, you can now find it here (alongside a PyQt version of the same program):
http://qt.gitorious.org/qt-labs/graphics-dojo/trees/maste...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 18:55 UTC (Mon) by aleXXX (subscriber, #2742) [Link]

I haven't met any professional software developers who develops with vala yet.
I also haven't met any professional software developer who is developing GUI software in C.

In my 8 years of experience as professional software engineer 100% of all colleagues who were writing software for a desktop PC were using C++, be it on Linux, OSX or Windows (in 4 different companies).

Alex

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 21:07 UTC (Fri) by tuna (guest, #44480) [Link] (29 responses)

One of the advantages of the GObject-family over QT seems to be that there are more bindings for different languages such as Vala, C++, Python, Javascript, C# and others. Are there any QT developments adressing this?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 22:21 UTC (Fri) by foom (subscriber, #14868) [Link] (23 responses)

Are you kidding? Qt has awesome bindings, for python at least. I haven't tried any of the others, but wikipedia says they all exist.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 22, 2010 22:59 UTC (Fri) by kripkenstein (guest, #43281) [Link] (22 responses)

He's not kidding.

The state of Python/Qt bindings is confusing (there's a GPL option, which is good, but Nokia are pushing forward with an LGPL effort. That's not done yet, but when it is, the business model of the GPL version will vanish, so not clear how that will continue). Qt/Java bindings have been abandoned by Nokia, and not clear what their status is. Other languages are not doing well either.

This is a significant advantage of the GTK+ side of the family.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 1:21 UTC (Sat) by foom (subscriber, #14868) [Link] (2 responses)

So, because Nokia is funding development on LGPL bindings for Python so that non-GPL'd software can use them, that means things are not going well?

The Qt Java bindings appear to still be under active development:
http://qtjambi.sourceforge.net/blog/

Now, things don't look that great on the Gnome side to me:
http://www.gtk.org/language-bindings.html

And the latest addition to the Gtk Java bindings is support for "accelerators" (aka menu shortcuts)? They didn't have that yet? Sounds pretty incomplete...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 7:50 UTC (Sat) by drag (guest, #31333) [Link] (1 responses)

> Now, things don't look that great on the Gnome side to me:
> http://www.gtk.org/language-bindings.html

That page does not make any sense. It has to be screwed up. How does it make sense that GTK+ does not have official C# support when Gnome is shipping at least one C# program in the latest version of the Gnome Desktop Environment?

How can Vala be officially supported for GTK+ 2.8 when GTK+ 2.8 was released in 2005, but Vala development did not even start until 2006?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 14:59 UTC (Sat) by Ed_L. (guest, #24287) [Link]

Just a nit, but Vala is being developed specifically to compile against Gtk+ objects without change of ABI. I realise there's little challenge shooting at well-defined static target, but such appears to have been their choice.

:-)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 9:59 UTC (Sat) by rdale (subscriber, #70788) [Link] (18 responses)

"..The state of Python/Qt bindings is confusing.."

Yes, it might be confusing but it is hardly a problem that there are two actively developed Python bindings projects.

"..Other languages are not doing well either.."

The PerlQt/KDE bindings are being very actively developed. Ruby and C# are waiting for the move from KDE's svn to git before so they can be reegineered/refactored - the projects are far from dead. The PHPQt bindings haven't been maintained recently, but someone has recently voluteered to start working on them, and we'll have to see how that works out. There are two QtScript based bindings projects, one in Qt Labs and another one called JSmoke in the KDE git repo. There are Common Lisp bindings, Lua bindings and others. There is a lot going on. The more people that want to use Qt bindings the more effort will be put into developing them. There are no technical advantages in GTK over Qt with respect to language bindings.

There are language independent bindings libraries called 'Smoke' for Qt and KDE which is equivalent to Gnome's GObject Introspection libraries, and several bindings projects are based on Smoke (Ruby, C#, Perl, PHP, QtScript and others). The is much less manual effort involved in producing a Smoke library than there is for a G-I library. Once a Smoke library has been created, it is relatively trivial to add language bindings for it - mainly a matter of writing language specific marshallers for any of the more unusual types.

In addition to all this activity, the very heart of the UIs of Qt apps in the future will be based on a form factor independent visualization layer written in QtScript as QML Qt Components.

It is certainly true that there are many more mainstream Gnome apps written using language bindings than in C. But in my opinion that is because writing apps in Gnome C is a lot less viable than writing Qt apps in C++ is. There is simple a greater need to use language bindings for mainstream apps in Gnome than there is in the Qt or KDE worlds, as the GTK C api is so use hostile. You are welcome to chalk that up as an advantage if you like - it depends on how you look at it.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 13:21 UTC (Sat) by danieldk (guest, #27876) [Link]

Ruby and C# are waiting for the move from KDE's svn to git before so they can be reegineered/refactored - the projects are far from dead.

Yes, I have been using the QtRuby bindings for a project, and they worked just fine. They recently updated OS X support, so the application now works fine there too.

There is simple a greater need to use language bindings for mainstream apps in Gnome than there is in the Qt or KDE worlds, as the GTK C api is so use hostile.

I would like to emphasize this a bit more. It is really easy to expand on the C++ Qt API. For instance, we wanted to do custom tree-drawing (rather than converting the tree to SVG, and displaying it on a QGraphicsView), deriving QGraphicsScene and QGraphicsItem to do such a thing is intuitive, and doesn't cost a lot of time. And this it true for many parts of Qt - it is clearly written to allow customers to expand functionality.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 13:36 UTC (Sat) by drag (guest, #31333) [Link] (4 responses)

> Yes, it might be confusing but it is hardly a problem that there are two actively developed Python bindings projects.

It's been a big problem for Python in the web scripting sphere that there are so many different ways to do the same thing. People try to look forward to a simple and straightforward platform they can base their programs off of. With redundant choices then it becomes difficult to know what is the correct choice to make and what they can expect that user's devices to support without making them jump through hoops to install the software.

I don't see how it would be much different with mobile applications.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 14:03 UTC (Sat) by rdale (subscriber, #70788) [Link]

"..It's been a big problem for Python in the web scripting sphere that there are so many different ways to do the same thing.."

Yes, one principle of Python is that there should only be one clear way of doing the same thing. In terms of providing this sort of clarity with respect to which Python bindings implementation to use, Canonical are quite good at understanding the problem. To solve it, they can either provide more funding to the LGPL PySide project to sort out any technical issues, or they can negotiate with Phil Thompson, the GPL PyQt bindings developer more successfully than Nokia did, to sort out any licensing issues.

The implicit reason for Nokia starting the LGPL PySide Python bindings project is that there is a large constituency of commercial developers who want to use Python, but wouldn't be prepared to pay a commercial license fee to use PyQt. I'm not sure if this is true - the cost of a PyQt license is not much. Anyhow, in my experience, C++ is easily the best language to use for writing commercial apps. Certainly Python, Ruby and other languages, have a place for writing custom apps developed 'in house', or for protyping new ideas and so on.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 14:33 UTC (Sat) by foom (subscriber, #14868) [Link] (1 responses)

> I don't see how it would be much different with mobile applications.

One way it's different is that they both have the same (or at least quite similar) APIs -- PySide is being developed with the explicit goal of being API compatible, just license-incompatible.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 19:32 UTC (Sat) by drag (guest, #31333) [Link]

Well as long as it's ABI compatibly without forcing either the developer or the user to take into account which particular bindings they are using then I suppose it's not bad. It's going to be hard to maintain 100% compatibility though.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 23:51 UTC (Sat) by pboddie (guest, #50784) [Link]

It's been a big problem for Python in the web scripting sphere that there are so many different ways to do the same thing.

I don't disagree that people felt intimidated by the choice, and it's somewhat counter-intuitive that when people see a vibrant community actively producing lots of things and another producing very few things (or even a single thing), they often pick the one with fewer choices (or a single choice) because they (or it) must be "better supported", but in many cases people could just pick one thing from the very long list of solutions, or at least one of the most actively developed things, and it would do the job without much cause for complaint.

That said, I always wanted to see more collaboration instead of unnecessary competition and duplication in the Python Web scene, but then Django came along and filled the role of "obvious choice" that everybody seemed to be craving.

People try to look forward to a simple and straightforward platform they can base their programs off of. With redundant choices then it becomes difficult to know what is the correct choice to make and what they can expect that user's devices to support without making them jump through hoops to install the software.

Well, on the Qt bindings front the tried and trusted choice is obviously PyQt since it has been around for years, has been used in lots of stuff, and is well supported. Given that the other bindings try to emulate the PyQt API, I don't see how you can go wrong writing applications against that API: ultimately, whatever runs on the device will be supporting it.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 16:29 UTC (Sat) by RobSeace (subscriber, #4435) [Link] (11 responses)

> as the GTK C api is so use hostile

As a life-long C programmer, I disagree completely... I always hated GUI programming, until I encountered GTK+, which was the first semi-reasonable GUI API I've ever encountered... It seems to be a nearly perfect API from a C programmer's prospective... Of course, the issue is more that less and less modern programmers want to (or know how to) write C code... So, that's probably more the reason that more people use non-C bindings for GTK...

But, for SOME of us, having a pure C API available is still important, since we do write in pure C and have to interface with other pure C code, and using C++ just isn't a palatable option... So, as good as QT may be, it'll just never be a valid option for people like me...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 21:59 UTC (Sat) by Ed_L. (guest, #24287) [Link] (3 responses)

Facinating. I mean, I share your appreciation for Gtk+'s clean C ABI, as the host of (allegedly) higher language wrappers would not otherwise be possible. Without people like you we wouldn't have that. Still, while reasonably small functions are one thing, the thought of writing a large program in pure C just isn't a palatable option... ;)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 14:23 UTC (Sun) by RobSeace (subscriber, #4435) [Link] (2 responses)

> the thought of writing a large program in pure C just isn't a palatable
> option... ;)

And, I'm sure many modern programmers share your view... But, personally, I
love writing C code, and am happy using it for all manner of projects, from
tiny to massive... I've written in many a language over the years, but there's
none which I enjoy more than plain old C... Yes, it doesn't hold your hand
for you, and yes, it lets you shoot yourself in the foot if you're not careful;
but, I find those to be GOOD qualities, personally... I don't like languages
that think they're smarter than me...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 23:47 UTC (Sun) by HelloWorld (guest, #56129) [Link]

You clearly didn't understand what high level languages are about. Hint: they're not about outsmarting the programmer.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 16:08 UTC (Mon) by jabjoe (guest, #70720) [Link]

I'm with you. C is a really underrated language. It really keeps it simple.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 12:56 UTC (Mon) by ringerc (subscriber, #3071) [Link] (6 responses)

I tried to like GTK+, I really did. In fact, I have nothing much against the toolkit its self - though its documentation and tools really suck compared to Qt's.

For me, the problem is GObject. Having used much more C++ than C, I found myself really frustrated by GObject's attempt to provide an object model in a non-OO language. It works, but IMO compared to using C++'s native objects it's painful and clumsy. The horrific syntax for C function pointers doesn't help any. Writing an object system with macros, function pointers and struct aliasing is ... unpleasant. They've done some great work to polish over the ugliness and produce something quite usable from it, but I can't help wishing the ugliness wasn't there. Like any abstraction layer, it's leaky, and the leaks smell BAD. They also make debugging ... exciting. It's a lot better than some other C-object APIs I've seen, like Python's C API, but it's still not nice. Using it, I always wish I was using Qt instead.

I've used Qt a fair bit, and came to really like it. It has a clean, simple API that's very, very nice to use, and never feels like a hack on or around the language. It tends to do the vast majority of what you want right out of the box and be easily extended to do the rest. I think it's a great shame its licensing (and the dire state of C++ compilers in the late 90s) kept it from becoming the de-facto standard toolkit for Linux systems, as it's a dream to program with. Whenever I work with GTK+ or with Java/Swing I wish I was using Qt.

As a C++ based tool you have to be aware that you're using a foot-cannon language, and know what not to use as well as what to use. Then again, the same is true of C, with horror macros, abuse of longjmp, using goto outside error handling, etc.

The biggest complaint I have about Qt isn't really about Qt at all, but about the tools on Linux systems. gcc produces awful error messages from hell when dealing with C++ code. gdb has somewhat clumsy support for C++, and worse support for the STL. I actually found myself coding on Linux and doing my memory debugging (valgrind) on Linux when that became necessary, but moving to Windows with Visual Studio (Express) when I had something tricky I needed a debugger for, because gdb was so annoying. It's ok for C, but for C++ ... argh.

For the C++ naysayers: C++ can be an awful language when abused, and it's easily abused. Personally, I think parts of C++ are just nasty, but I still prefer it to C any better for high-level / GUI programming. My dream would be the Qt APIs on C# or Java, but integrated into the core of the language runtime so you didn't have to fight with native libraries like Qt Jambi. I love C for implementing libraries and low level systems, but not user apps and GUIs. That's where Qt and C++ shine.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 14:10 UTC (Mon) by jwakely (guest, #60262) [Link]

> gdb has somewhat clumsy support for C++, and worse support for the STL.

GDB 7 is a big improvement for debugging C++

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 2:56 UTC (Tue) by cmccabe (guest, #60281) [Link] (4 responses)

> gdb has somewhat clumsy support for C++, and worse support for the STL.

For what it's worth, gdb 7.0 is supposed to allow you to inspect STL containers with the "print" command. My distribution hasn't bundled it yet, though.

> gcc produces awful error messages from hell when dealing with C++ code...

This is another thing that's gotten better over time. Part of the problem is the complexity of C++ itself. Its grammar isn't even LR. Also, the way templates work makes verbose error messages a part of life.

Despite that, I think templates are the best feature in C++ -- the only one I really miss when I use C.

> Personally, I think parts of C++ are just nasty, but I still prefer it to
> C any better for high-level / GUI programming.

Yeah, but in the end, we ought to write GUIs in a truly high-level language like Scheme or Ruby. The best thing about Web development and Android development is that it's finally opened people's eyes to the silliness of developing GUIs in C/C++.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 3:33 UTC (Tue) by ringerc (subscriber, #3071) [Link] (3 responses)

I miss basic templates when coding in C. I'm kind of glad the fancier template features aren't there to be abused, though. Just because you can do things like template metaprogramming doesn't mean you should.

I really, really hate the lack of template functions in C. I can live without template classes, but the lack of template functions is horrible. Whenever I want to write something that takes a variety of arguments and outputs either a fixed result or a result of the same type, I have to hand-code all the variants or use horrid macro hacks, I can't just write

template<T> T somefunction(T a, T b) {
return a [someoperator] b;
}

I also tend to miss operator overloading. It's one of the few features I like to see used in C++ libraries, not just final C++ executables, because cautious and appropriate use of operator overloading makes working with things like complex types, binary-coded decimal types, decimal floats, etc infinitely less painful.

I sometimes wish for a pure C binding to Qt, though, just so the C-only folks who won't learn enough about C++ to see that they can adopt a restricted subset without fear can learn how nice working with Qt can be.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 10:38 UTC (Tue) by RobSeace (subscriber, #4435) [Link] (2 responses)

> horrid macro hacks

I really don't see how this:

#define somefunc(a,b) ({ typeof(a) _ret = (a) [operator] (b); _ret; })

is any more horrid than that template ugliness you posted...

> just so the C-only folks who won't learn enough about C++ to see that they
> can adopt a restricted subset without fear can learn how nice working with
> Qt can be.

Just FYI, that's not the actual reason for some of us... I, for one, simply
can't use C++, because all the code I write needs to be able to interface with
tons of other pure C code we have lying around from dozens of years of
development... Sure, C++ can call C lib code no problem; but, the reverse
is not true, and probably most of what I do is write lib code... And, even
when I write user-facing apps, they're generally of the generic variety,
where the interface and behavior are configurable at someone else's whim
(because, as a geek, I know my idea of a good interface isn't necessarily
everyone else's), which usually requires (among other things) supporting
plug-in shared libraries to implement arbitrary behaviors (including possibly
calling some of the app's own functions and messing with things)... So,
writing it in C++ would basically require C++ plug-in libs, which wouldn't
make me very popular among whoever got stuck writing them later (since we're
all used to dealing with pure C code all the time)... Yes, it's sort of a
circular "We write C because we write C" argument; but, you know, momentum
is a hell of a thing... It's not just anyone stubbornly refusing to use C++
out of spite or something; there's a very real cost involved in switching...
Even for something totally stand-alone that needed to interact with no other
code at all, you have the issue of maintaining it, which now becomes harder
for everyone else who isn't intimately familiar with C++ like they are with
plain C...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 11:27 UTC (Tue) by ringerc (subscriber, #3071) [Link] (1 responses)

The horrid macro hack you posted is worse than the horrid template function I posted only because the template is portable and standard. The macro isn't; it relies on gcc extensions that'll fail hopelessly on pretty much any other compiler. If C had a typeof() feature like the gcc extension, even if it was c99-only, there'd be much less need for template functions. Perhaps such a beast HAS been standardized in C recently, but if so it doesn't seem to have made its way into compilers like VC++.

"Sure, C++ can call C lib code no problem; but, the reverse is not true"

C code can call some C++ libraries fine - those where the C++ library author makes some effort to make it possible. You *do* have to limit your public interfaces to "extern C" functions without overloading or templates. This is often quite acceptable, and lets you do your core implementation in C++. The trickiest bit is probably ensuring that no exceptions thrown internally escape into the C code - and that's not overly hard either.

That said, if most of your consumers are C-only, I doubt it'd be worth the hassle without a strong need to use C++ features.

(BTW, even if you have a library with a purely C++ interface you can write a wrapper library that presents only "extern C" functions - but it's not a lovely job. I've done it where I need to use a C++ library from plain C code, and would choose to do so again in those situations, but I doubt I'd choose to write a new C++ library if I only expected to have C consumers.)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 26, 2010 12:46 UTC (Tue) by RobSeace (subscriber, #4435) [Link]

Even losing the typeof(), just:

#define somefunc(a,b) ((a) [operator] (b))

gives you most of the same functionality, really... The expression will
naturally cast to a's and b's type... But, if you pass in differing types
for each arg, then you may wind up with a result of either's type, depending
on which way things get promoted, or a compile time error/warning if the
types are incompatible...

Of course, I know that was a very trivial example, too... For more complex
functions/macros, you probably do need to define temp variables that take on
the appropriate type... So, yeah, if you care about compiling in anything
other than GCC, you're boned... But, thankfully, I don't... ;-)

> That said, if most of your consumers are C-only, I doubt it'd be worth
> the hassle without a strong need to use C++ features.

Exactly... And, I know everyone says QT is the greatest thing since sliced
bread, but frankly we've been plenty happy with GTK+ for our needs, so there's
been no real compelling reason to even look into QT at all, let alone tempt
us to start coding in C++ just so we can use it... If we found GTK to be
horrible and unusable, it'd be one thing; but, honestly, I find it kind of
a joy, really... And, this from a low-level system coder, who HATES GUI
coding at all!

> BTW, even if you have a library with a purely C++ interface you can
> write a wrapper library that presents only "extern C" functions

Someone ought to do that for QT then, just to bring the wonders of this API
to us heathen pure C people, then... ;-)

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 6:56 UTC (Sat) by wstephenson (guest, #14795) [Link] (4 responses)

There are Qt (and KDE) bindings for Python, Javascript (from Nokia), C#, Ruby, Perl and PHP (from the KDE project).

See here: http://websvn.kde.org/trunk/KDE/kdebindings/

And although they come from the KDE project, the pure Qt bindings don't depend on KDE, so don't start with that.

And it's Qt as in Cute, not Que - Tee. Capisce?

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 11:50 UTC (Sat) by jengelh (subscriber, #33263) [Link]

>And it's Qt as in Cute, not Que - Tee. Capisce?

But people will come up with their own language pattern for a given name anyway (cf. /ˈlɪnʊks/ vs /ˈlɪnəks/), so don't count on Qt being cute everywhere. Especially since Qt used to be QT (esp the logo when under Trolltech reign), and people retain their patterns.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 23, 2010 21:51 UTC (Sat) by mpr22 (subscriber, #60784) [Link] (2 responses)

They can pronounce it /kjut/ all they like. I'm going to pronounce it /kju ti/ in the certainty that I will get misunderstood less often than they will.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 0:24 UTC (Sun) by Jonno (guest, #49613) [Link] (1 responses)

They can pronounce it /kjut/ all they like. I'm going to pronounce it /kju ti/ in the certainty that I will get misunderstood less often than they will.

Following English language praxis, QT (pronounced /kju ti/) would be an abbreviation, and most (computer interested) people I know would read it as Quick Time...

All in all, I think spelling and pronouncing it the intended way (Qt, /kjut/) is less confusing...

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 24, 2010 1:28 UTC (Sun) by TRS-80 (guest, #1804) [Link]

I bet you pronounce SQL as sequel too.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 25, 2010 22:43 UTC (Mon) by BeS (guest, #43108) [Link] (1 responses)

I did some GNOME/Gtk+ programming a few years ago.

Lately I tried to reactivate my activities regarding Gtk+ programming. But what I saw was quite confusing. I was used that I use a GUI/Widget-Toolkit if I want to create a graphical user interface. But with the GNOME stack we have Gtk+ (the classic widget-toolkit), Cairo (if I understand it correctly this is what I use if i want to do same drawing (within Gtk+?!)) and Clutter a new "3D accelerated toolkit" which is used for the GNOME3 desktop.

I think this is highly cluttered ( ;-) ) for people comming new to GNOME/Gtk+ land.What should be used in which situation? Should I use simple Gtk+ for my GUI, should I use Cairo and draw some nice widgets by my own or should I do everything in fancy Clutter-3D-fashion. Do I really have to learn all three APIs to participate in GNOME/Gtk+ land? Will it be possible to keep an unique look-and-feel with that much different options?

It seems like the situation for KDE/Qt is clearer and more tidy compared to GNOME/Gtk+. But maybe I'm just wrong and you can help me to understand the situation at GNOME/Gtk+.

How Qt could bring better third-party software to Ubuntu (ars technica)

Posted Oct 29, 2010 2:55 UTC (Fri) by njs (subscriber, #40338) [Link]

Cairo's a library for turning "hey I want to draw some lines" into actual pixels representing those lines. GTK+ is a library for making windows and buttons and text entry forms and that sort of thing. They're at different levels on the stack; GTK+ uses Cairo to draw those buttons, and if you're using GTK+ for your widgets and then discover you want to draw something that's weird and unique to your app (like, I dunno, the board inside your chess program or a graph inside your spreadsheet program), then you can use Cairo yourself to do that. Or if GTK+ already has all the stuff you need, then you can ignore it. (Qt has this division too, just their drawing library is a particular set of functions inside Qt instead of having its own name.)

Clutter is... it's own thing.