I've no problem with multiple toolkits on Linux, but I really don't think there is any point in innovating on File Dialogs, or Button Orders. I don't care about whether the Gnome dialogs are better than the KDE ones. That stuff was done 20+ years ago, and anyone who thinks that designing a better File Dialog in 2004 is 'innovative' has lost the plot. So what gets up my nose somewhat is when a Gnome blogger just completely fails to acknowledge that KDE exists.
Luis Villa says:
I installed evo and galeon back in 2001, with apps in gtk, motif, and even raw X. Frankly, when I install stock firefox and stock open office, that's what I feel like I have again- the dialogs don't match, the icons don't match, the behavior doesn't match.
He then goes on to discuss various things:
Almost definitely we've made it technically too hard to follow them.
Umm.. In what way? Lets just co-operate via freedesktop.org and ensure that a Gnome app looks just the same as a KDE app. Cut this crap about how Gnome has 'innovated in the GUI field' because it hasn't. And neither has KDE. If you want to investigate something that really is the future check out Croquet. There's just no comparison, we need to consolidate those 20 year old ideas right now, so we can move onto the next thing. Pervasive use of 3D graphics with OpenGl, peer to peer networking all the time, dynamic scripting languages etc and so on..
Syndicate Blogs
context
I admit I am a gnome user generally, but lately doing KDE development at work, I have started to appreciate and understand parts of KDE.
In your post you admitedly link to Luis's blog entry in the spot where you dropped the one line that provides the context to the line you seem to take issue with. I am refering to, Now, maybe GNOME's standards are not the right ones to follow. Almost definitely we've made it technically too hard to follow them. With that context it is clear that Luis is refering to the GNOME standards being too hard to follow. I generally believe that interoperation between KDE and GNOME is a good thing. However, there is also nothing wrong with enabling firefox and other 3rd party applications to have a more consistent look with the rest of the desktop environment. It is one thing that Luis is working on on behalf of GNOME. Bu this in no way precludes the same sort of integration with the KDE environment. Frankly, firefox doesn't fit nicely into either environment. Whether the rest of GNOME succeeds with the file dialog changes etc shall be seen in the future and they should be able to stand on their own merit.
What Do You Expect?
The Linux desktop is Gnome to these people.
This is basically a rant to turn just about everything into a Gnome application, including Firefox and Open Office, conveniently painting over the fact that they are cross-platform applications. He's living on another planet if he thinks everything, unpatched, is going to be a Gnome application. Although people can look at KDE integration and provide access to KDE extensions and technology you're never going to turn applications like Firefox and Open Office completely into KDE applications, nor would you want to. As a desktop gets more widely used, and we've seen this with Windows, you're just not going to get total integration everywhere as people branch out and write different kinds of applications. How many toolkits does he think there is on Windows, not to mention third-party ones? Windows is write whatever you want, with a common bit of technology underneath
. I think he wants to look at what people actually write for Windows and pick up the cluestick.
Bizarrely, he also seems to see Gnome as providing a set of standards that somehow other projects like Firefox and Open Office should just follow, and he calls this writing apps for Linux.
I just saw this for the political bollocks that it is.
ha
Yeah, Luis' post really annoyed me and I was going to reply... but now you've stolen my thunder.
Re: ha
You’ve stolen my thunder
No problem. I like the idea of Free Software Hackers co-operating. But to what extent is Luis Villa a corporate drone? He gets paid fulltime to work on Mono by Novell I believe. And Mono is the 'One True Way' according to these people. I read Miguel doesn't like Parrot:
"Dan Sugalski was a man of his word, and brought a couple of pies for the contest. I do not like Parrot, but the man and his team stood by their word. Guido was a gentlemen and refused to throw the pie. So no pics."
Why write "I don't like Parrot" with no technical qualifications?
So as an unpaid volunteer should I attack our fellow Free Software developers? Yes, I think I should if they obviously don't have a clue, and clearly have a hostile attitude to our KDE project by avoiding mentioning it for no apparent reason.
Parrot.
Hello Richard,
I have described elsewhere my problems with Parrot, a google
search finds:
http://primates.ximian.com/~miguel/archive/2003/Jul-22.html
Enjoy,
Miguel.
Not entirely on-topic really.
Not entirely on-topic really.
That's not a description of Parrot - it's a reason of why .Net is better in a fairly narrow set of circumstances.
.Net is basically designed to produce large-scale systems that are designed to be built in a batch fashion and left. Yes, even Visual Basic is like this and VB.NET is definitely not a full scripting language - and its definitely got further away from it as the years have gone by. VBScript/JScript/Python etc. are, and they are used in a different way. Parrot, although it has shortcomings like everything, is designed to be in a dynamic environment. Just plain logic tells you that the two worlds are different and their designs are going to be different.
I just don't get the comparison to be honest.
Re: Not entirely on-topic really.
Well my original point was that Croquet is a lot more interesting than (incompatible) file selectors, before I got sidetracked with my Parrot rant.
Anyway I'm going to download the various Croquet white papers and study them. I wonder if KDE/Ruby/Croquet integration is possible? It could be called Kroquet of course..
-- Richard
Re: Parrot
Hi Miguel
Thanks for an interesting link there. And I'm sorry if I sounded so stroppy.
I like the idea of layering a tagged architecture onto the CLR. I'm personally very interested in whether it is possible get Ruby running on Mono. Alex Kellet with his Rubydium project is using libjit to directly interpret an AST, without going though the byte code stage, and I think that's a very interesting approach.
From what I've read about the CLR, it sounds as though anonymous delegates will map well onto ruby blocks. I think that the main thing it needs are proper class methods with dynamic dispatch. That design error seems to be hard coded into the class model. You can always map class methods onto Java or C# static methods, but not vice versa. As it stands I think you would need two CLR classes for every ruby class, one for the metaclass to act as a factory, and one for the actual class.
-- Richard