Slashdot Mirror
Interview with Taylor & Pennington from Red Hat
RH-Gimp writes "OSNews has put together a long and informative interview with Havoc Pennington and Owen Taylor from Red Hat. They discuss about the KDE issues, the UI on Red Hat 8.0, the future of the Linux desktop and XFree and other interesting stuff."
RH 8 is slick. Kudos to RH for raising the bard for other dists that claim to be for the desktop. RH has shown consistent improvement from release to release, that's all we can ask.
Give us back the fucking virtual desktops! (aka viewports)
What the hell were you thinking when you said that "Multiple Desktops" had all the functionality needed, so viewport people were out of luck?
I mean, come on. With viewports, all you needed to do was turn off edgeflipping and you were done. Instead, you rightly say that changing your current code to allow edgeflipping would be a pain in the ass.
Man, Penn and Taylor are my favourite comedians of all time, I had no idea they worked for RedHat... Guys I totally loved that time you threw the glass of water in the talk-show host's face, and you rocked on Fear Factor! ;-)
At first I criticized Redhat's blending of KDE and Gnome, but now I am beginning to appreciate it. It is adding yet another dimension to Linux on the desktop, and seems to be doing so in the same spirit of creative development that has driven Linux as far as it has come. Maybe having only two choices wasn't enough?
The article states that after many years of "It's not a bug, it's a feature!", real resolution changing has just been added to XFree86's CVS.
About bloody time.
Oops!
;-)
Sorry - that would be Penn and Teller. Silly me.
Soko
"Depression is merely anger without enthusiasm." - Anonymous
It was really a no brainer that some distribution would finally decide to try to piece together the different UI's, and at least try to make them look similar to one another. Imagine a bathroom with a gold shower head, a chrome drain, and marble knobs. Things look much better when they are in sync. Kudos to RH for finally getting the ball rolling in this area.
(+1 Funny) only if I laugh out loud.
It's good to see young developers getting involved is OSS. I mean, what are these guys, 11, maybe 12 at most?
:P
why run from Vincenzo?
It is a shame that OSNews did not question the UI designers about the removal of the taiwanese flag from redhat. It is even more shameful that no one on slashdot cares about it.
The average reader appears to take the stance that "Redhat exists to make money, and if this is what they have to do, then so be it." I find this insulting coming from a movement which is supposedly all about freedom.
It appears that free software is merely all about not paying for software and the downfall of microsoft.
I'm glad they're busy tweaking out a decent interface and making it more readily-usable -- making sure everything's aligned right, implementing an xFree86 that can actually do dynamic resolution changes, sticking to a GUI standard, and so forth.
But if people spent more time working on or with Redhat and less time talking about/flaming it, it would have a lot more commercial success and would serve as a better bulwark against you-know-who.
RedHat is one of the most important companies involved in getting linux accepted and used outside of its traditional audience, along with IBM and now Sun. I personally like RedHat 8 and wish them coninued good fortune.
Listen, you have to run out and get an exta rpm for playing mp3s. As always you have to download Nvidia drivers and if you have an ATI card I suggest going to the gatos.sourceforge.net and using those drivers. Fonts are an install away from the corefonts sourceforge project and dvd playback requires an ogle download.
I can understand every single bit of this. However, apt-rpm needs to come with the distro.
Also what is the deal with the extras submenu? I understand simplifying the menu structure. The SuSE distro menu is a huge mess with a hundreds of programs organized fairly well but still hard to find and half with no icons in the menu! Still, when a new program is installed the user should have a choice of whether they want it merged into the main or the extras menu (can't they come up with some better frickin' title for the thing?) not very easy for an end user.
Finally they need to be hit by a clue-by-four from of all places with the dipsticks at Lindows. Every desktop OS has at one time or another a compatibility layer to ease users over to its use. Mac OS X has one for old OS 9.2 apps. Windows had one for dos and Win 3.11 apps. We need a compatibility layer that runs Windows apps and it is called Wine. It is time that the distros come together and I mean everyone including the OpenLinux distros, Redhat and Mandrake and figure out how to make Wine as good as it can be without it being completely taken over by codeweavers and transgaming.
A good compatibility layer that works as well as CrossOver Office does right now out of the box with no messing around. Install Redhat, and then install Office 2000 and it just works. This is needed not by me but the newbie easing into Linux use.
It is still going to take a shift in thinking to get Linux to the desktop in any numbers even within IT departments.
Currently the Distro is still seen by too many as simply being the OS layer -- kernel, GNU shell and the GNU utilities.
The Distros need to think of the Linux OS as being made up of three parts as most OSes do --
OS layer -- kernel, GNU shell and GNU utilities.
Compatibility layer -- Wine
GUI layer -- kernel frame buffer support to Xfree86 to finally the desktop
Redhat is almost there and considering how quick the shift in focus came from Redhat they did a pretty good job.
ACK
Now all the other desktop distributions that wants to attract users should also create a unified desktop. The problem is if all the desktop distribs create their own standards, it will create more confusion. What the distribs needs is the committee that actually makes worthwhile decisions like a desktop standard.
She did not bring this up because when the questions were written & sent RedHat 8.0 was in beta and many issues hadent came to light yet. Here is a quote from her :
"I sent these questions to the Red Hat guys, BEFORE I actually installed Red Hat 8, get pissed off with the nvidia drivers, and wrote that review. If I had sent it later, my question would have been different, more direct, and maybe even a bit rude (eg. "you are a big company now and nvidia has the gfx market. Why don't you PARTNER with nvidia to make sure the damned driver works with your OS?"). Count on it."
-- "of course thats just my opinion, I could be wrong." --Dennis Miller
I never thought about:
"OWT: A lot of it was misunderstanding, but there are certainly real issues as well. Red Hat is interested in a desktop that is well integrated into the OS. The KDE project is interested in a desktop that is well integrated with itself. These goals don't always completely coincide. "
Now that I think about it this is so true KDE seems to try and do everything itself, gnome apps seem to add onto things while KDE makes its own program for something already there.
"of course thats just my opnion, I could be wrong" --Dennis Miller
-- "of course thats just my opinion, I could be wrong." --Dennis Miller
Redhat would never put apt-rpm in by default because then why would anyone pay $60/year to get their system updates via up2date? I would gladly pay for up2date if it were $60/user but there is no way I'm paying a $120/year for free software updates on my 2 redhat boxes at home. So I installed apt-rpm on my redhat machines, and gave my $60 to the mandrake club.
From page 3 of the interview:
"9. How do you feel about XFree's inability to fully function as a modern graphics subsystem? (e.g. you can't change real resolutions on the fly, no support for OSX's and BeOS6's smooth window moving etc)..."
BeOS 6?
Did I miss something? Yes, the R5.1 'Dano' developer release leaked, but I hadn't heard anything about R6.
First of all, there is a committee for free desktop standards[freedesktop.org]. Secondly, all other distributions have no need for creating a unified desktop. Red Hat did it because it felt that it was its responsibility, as market leader, to provide something that offers new users that extra cushion. Distributions like Slack, Gentoo etc, on the other hand, don't command such a high market share, and therefore, do not feel the need to spend thousands of $$ developng a unified desktop. The user base of such distros(Slack etc) is already well accustomed to the idea of two separate DEs and a bunch of WMs. If these users want a unified desktop, they should go for Redhat. If RH users want the full freedom of choice between DEs and not something modified by RH, then they should go ahead and try out other distros like Mandrake, Suse etc. Linux is all about choice and every distro has something unique to offer. If distros start copying each other, then linux bubble will burst, just like the unix bubble did a few decades ago. Kudos to RH for trying to stand out from the pack.
-- Reality is just an extended dream.
For those who browse without sigs
Redhat kowtows to China
And to the moderators--saying something bad about a Linux distributor doesn't automatically make a comment flamebait. Sometimes it's true.
They've stopped sucking for a long time now. I don't know what cookie-cutter distro you use, but my Gentoo installed automatically included a TT-bytecode interpeter enabled FreeType and MS Windows fonts. When I got my UXGA laptop, there was initially some problem because the LCD looked best when all fonts were anti-aliased (it has enough resolution that AA text looks sharp rather than blurry) but all I had to do was download FreeType2 CVS, get some nice Postscript fonts, and I ended up with font rendering that whips OS X all over the place and easily matches ClearType.
A deep unwavering belief is a sure sign you're missing something...
we do not need a compatibility layer, we need native apps!
"Every desktop OS has at one time or another a compatibility layer to ease users over to its use. Mac OS X has one for old OS 9.2 apps"
Yea so what's your point? These compatibility layers were for running apps from the previous OS, NOT apps from an entirely different OS! OS 9.2 couldn't run windows apps, and Win 3.1 couldn't run Mac apps. We already have an OS that can run all of the windows apps, it called Windows.
Making wine work perfectly only serves to enforce the Windows monopoly. Do you now want the MS Office monopoly, proprietary file formats and all, to dominate the linux platform? Because that is what your idea leads to.
How is a "compatiblity layer" even a marketing tool?
You: Hey switch to linux you can run all of your old apps.
Customer: But I can already do that now.
You: Yes but you also get to enjoy zero tech support because your running in an unsupported configuration
Customer: runs away
Wine is a crutch that keeps people stuck in the windows world. It's not like I don't understand why you or anyone else wouldn't want to stick with some old app you've been using for years, but the fact remains native apps are better in every way imaginable way.
I look at Open Office when I'm in Redhat 8.0 and think God, I remember using Netscape composer for word processing because there were no gui word processors for linux. It apps like that and Evolution that will over time surpass the very same MS versions you want to bring over. Don't you think that end result is better?
Linux gaining the ability to run all windows apps natively leads to a windows clone, and I didn't switch to linux so that I could use IE, Office and Photoshop.
If you wanna get rich, you know that payback is a bitch
I really do hate displaying my ignorance like this, but what does this mean? Admittedly, I've never really played with resolution, but as I understood it, if you really wanted to, you could set it up so ctrl-alt-[KP+|KP-] rolled you up or down one resolution level... and at any rate I know that certain apps (eg tuxracer) run at an apparently lower resolution than I usually have... so what exactly is "real resolution changing"? I don't doubt that it's something extraordinarily exciting, I just haven't the foggiest idea what you (or they) mean...
"The best argument against democracy is a five minute chat with the average voter."
--Winston Churchill
Hmm. "attention to detail" was a phrase they used a lot. Ok, sure, these systems need polish.
But -- fuzzy icons -- unified shortcuts -- moving text around by a few pixels. Not _bad_ things to do, by any means. Still.
What about -- "attention to architecture", so that all of these "details" don't turn into infinitely long task lists, so that apps are far better and more consistent at being self-documenting, so that it doesn't take a ton of new code for every little app, so that interactive extensibility is built-in to the core, so that process are managed less horribly....
- The control center contains an incredibly useful "Information" section that unifies a lot of the info that can only be found on the console with a bunch of different utilities or cat'ing parts of
/proc.
- The incredible integration of konqueror with manpages and infopages (if you haven't tried this, you really ought to, the manpages are much easier to read in a konq window.)
- I haven't tried the gnome-terminal from gnome2, but the kde2 terminal allows multiple terminals from one window, which allows very easy access to those console utils.
KDE can import gtk+ themes quite handily if you like.
- There's a wealth of basic KDE utilities to do OS specific functions like changing password, managing users, and runlevels.
- The power of Konqueror as a file browser. Easy Samba browsing. Automated CD ripping and ogg encoding with drag and drop. Multiple window configurations. Embedded terminals. FTP and web browsing.
- The KPackage program for use with RPM or dpkg/apt.
These are just some of the basic things that KDE2 does to integrate in to the OS (I haven't even tried KDE3 yet). The KDE project just focuses on integrating with itself because they want to have a very well integrated environment. And they've obviously succeeded in that (just look at the programming model for evidence). Sure, they don't go out of their way to integrate with the Gnome folks, but then Gnome isn't doing any better. I think Redhat has a bias towards Gnome (there's a lot of historical evidence of this), and it's on display here. KDE does a good job of integrating with itself, but that doesn't mean that it doesn't integrate with the underlying OS as well."I may not have morals, but I have standards."
What the hell were you thinking when you said that "Multiple Desktops" had all the functionality needed, so viewport people were out of luck?
They do have all the functionality needed.
If you need more pixels on your screen, go buy a bigger monitor. For everything else, there is Mast^H^H^H^H the multiple desktops functionality. Duplicate features must be occamed out. And boy, did I ever cringe over the clipboard implementations galore that coexist in Emacs.
To be true, I too have issues with over-simplification. Right now, I cannot drag windows between desktops with the deskguide applet, but I don't think it's infeasible to implement without re-inventing yet another way to provide more than single desktop.
My exception safety is -fno-exceptions.
A lot of goodies which 1.4 had missing from GNOME2 IMHO.
The ability to dynamically bind keychords to menu items, lot of customisability options (panel behaviour etc.) etc. are all missing. I tried to customise Metacity and I get a small menu from the gnome-control-center with two or three options which is definitely less than what I can do with sawfish.
You have to love the fonts though :)
CheerioA lot of people seem to misunderstand that most of the debate surrounding Red Hat's changes has to do with unified look and feel.
It's not that that I, and many other KDE-philes, have a problem with (in fact, I and many others support the idea), it's the seeming favouritism for GNOME applications for the default shortcuts. In fact, I don't know if any of the default shortcuts link to a KDE-based application. For example, why not setup a default shortcut to Quanta? It's a highly respected web-development editor.
// file: mice.h
#include "frickin_lasers.h"
I cannot agree more with their stance on the closed-source nVIDIA drivers. Not only their closedness hamper the development of their open counterpart, it slows down their very adoption and resolution of real problems, as opposed to imaginary IP threats.
My exception safety is -fno-exceptions.
Timing is everything.
http://shotgun.shacknet.nu:81/leet.png
"If he thinks he can hide and run from the United States and our allies, he's sorely mistaken." Bush on bin Laden
Now it is easy to install TrueType Fonts in Linux /usr/share/fonts/ttf and within 30 secornds the Xft2 font engine finds the fonts.
Copy your truetype fonts in
Is the decision to label the Flag of the Republic of China as a bug. Seems wourthy of mention. Read more about it here.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
Red Hat 8.0's new consistency is provided using Xft2/Freetype with 'heavy modifications' and the Bluecurve theme. Underneath remember, it's still Gnome 2 and/or KDE.
The world is a big free and happy place. Which means you too can have the Red Hat desktop goodness on your distribution of choice and not have to complain about Taiwanese flags, RPM, additional packages for MP3 playback etc.
I haven't got a hard disc spare to install Red Hat 8.0 on (I'd really like to see it based on all the screenshots), but I do have a couple of Debian systems. Someone could make me a very happy man (and earn some serious karma) by taking the bits that are good about Red Hat 8.0 and making them available in other distributions.
That's how Linux works. Take the bits you like, ignore the bits you don't. Is Bluecurve on Freshmeat's themes section yet?
The only thing i think needs attention is installing/uninstalling applications. Today that is a pain in linux and should be addressed. To get adoption on the desktop that have to be as easy as in Mac or Windows to install and uninstall software.
Besides that i really like RH 8.0 and it works just fine for me and my wife.
HTTP/1.1 400
Texstar has some "Freecurve" RPMs for Mandrake 9.0 up as well, for Mandrake users who want a taste of Red Hat's new theme.
Thanks... with one simple comment on /., you've managed to make me feel old and ruin my day, and I haven't even had my coffee yet!
It's notable that the interviewees admitted that RH is still not ready as a consumer desktop.
Rather, RH is meant to be used by non-technical users on a carefully controlled system, installed by a knowledgeable systems administrator.
A lot of the PR I've read on 8.0 are breathless in proclaiming 8.0 as a Windows replacement, but as RH's own developpers point out, this is not the case.
evanchik.net
Sigh... it always comes back to this. "If Linux doesn't have it, do it yourself!"
Then why bother trying to make normal desktop users, companies, anybody that doesn't program, use Linux? And when they come back and tell you "Umm, I'd like to, but it doesn't have [missing feature], so I can't" you reply with "gee, that's too bad, you fix it".
They didn't miss the boat or anything. Slashdot asks time and time again why Linux isn't being adopted. Then you have to accept to be answered by people that have no skill, time or interest to actually fix it (or money, if there was a commercial alternative running under Linux). It's not like they all see GPL programmers as a bunch of people they can leech free (as in beer) software of, they just gave you the facts. Deal with it.
Kjella
Live today, because you never know what tomorrow brings
One of the interviewees said, "But in general: say you woke up one day and everyone on earth simultaneously agreed to switch from windows/icons/mouse to some new paradigm. It would still take 20 years, trillions of dollars, and be mind-blowingly difficult."
This is more than an understatement. We've been trying to make the metric system switch for more than 20 years and we're still only inches off the starting line.
System/application developers frequently forget this point and underestimate the importance of backward compatibility. Evolution will always win in a war with revolution; even if revolution wins a few of the early battles.
This is all fine and good, but how about giving to the rest of the OSS world something to use and build upon, without pulling the entire project with it?
Like, a powerful C portability/utility library, a standalone signal/slot framework, a lightweight ORB (this means CORBA the standard, not another crippled-by-design object framework?), a configuration management server, or a fully internationalized font rendering system?
My exception safety is -fno-exceptions.
Congratts! This qualifies as the most bizarre post on slashdot in recent history.
I just emailed this link to everyone in the office.
I can only think of a single KDE/Qt app that blows the GNOME/GTK equivalent out of the water, and that's licq, which Red Hat *does* ship.
* Konqueror? Not as widely used as Moz.
* Anything in the KDE office suite? Not even close to Open Office, even for the most die-hard KDE fans.
* Kmail? Not as comfortable as Evolution for people used to Outlook, and doesn't have PIM/schedule management a la Outlook (frankly, I think everyone should be using mutt, which has the best PGP support of all time, but...:-) ).
Actually, I take it back. There is no GNOME/GTK equivalent at all to kcheat, a program designed to let you cheat in video games by editing memory. Kind of silly that there's nothing else like this for Linux, esp. since kcheat is KDE 1...
May we never see th
Most Linux veterans do not appreciate KDE's attempts at achieving the integration, uniformity and predictability that Mac and Windows have, because they are too used to the status quo. It's much more than putting a "k" in front of every app. I've addressed this point elsewhere. The trouble is that Red Hat has substituted a different app for every major KDE app. A typical user would use Red Hat's "KDE" and never use any of its apps nor see its potential. KDE-rh is essentially an oversized, overslow window manager. There is no good reason to use KDE in this form: you might as well use a lightweight environment like xfce instead.
Owen's comment that "these goals don't completely coincide" is an understatement. KDE wants to be the desktop. Red Hat wants to own the desktop. They are completely at odds.
They still suck. You got around the suckiness by installing high quality TT fonts from Microsoft. Out of the box, screen fonts are far inferior to Microsoft's. The open source world does not have comparably high quality fonts.
Actually, with the FreeType CVS, the free Postscript fonts look just great. I've currently got the bytecode interpreter disabled because the autohinter in FreeType CVS is so good with Postscript fonts that it looks much better on my LCD than the hinted TrueType fonts. Letter shapes are a whole lot more accurate, while the outlines are still sharp. Read the FreeType mailing list (devel) sometime.
A deep unwavering belief is a sure sign you're missing something...
Yes. Most definately. Font rendering was never one of BeOS's strongest points. It was okay (better than Windows at the time), but the Bitstream renderer was showing its age. I was looking forward to the licensing deal with Bitstream to embed FontFusion (their new renderer) but that never happened. I saw FontFusion in action in QNX RtP and it was incredible. However, the version of FreeType in CVS right now is even better :)
A deep unwavering belief is a sure sign you're missing something...
You totally do not understand the issue. Too much choice is bad for everyone except the people that really love to tweak stuff.
For everyone else, too much choice is bad.
1. It clutters up the interface
2. It complicates the software and introduces bugs, some of which are hard to track down because of the myriad of options available to complicate things.
3. It is extremely confusing.
Choice is not "good period". Choice is good if the options are good.
KDE does have a problem with being too messy. The control center is way too messy. If you like KDE then by all means keep using it, but do not try to call me an idiot for not wanting all the options you like.
BTW: your analogy stinks. Choice about what software to use is good. That is what the analogy works with.
A better analogy is if the cars had big buttons that stated how fast the windows should open or close, or big fat buttons stating what concentration the window cleaning liquid should have. This would be tweakable. Certainly fun for someone but exceptionally bad for the rest.
"Nevermind the fact that if it was too confusing for you to deal with, THEN LEAVE IT ON THE FUCKING DEFAULT BEHAVIOR!!"
This is just insulting. The point is that there are some GOOD options. But how am I going to FIND those options if the interface is cluttered up with loads and loads of crack-options?
You my friend did not understand my point at all. Options DO have costs, which means you have to be careful about what options you give, and what you just decide on.
GNOME will hopefully never be a unification of all the options and features that have ever existed in desktop environments, like "emacs" is for editors.
Red Hat has just told a democratic people that live in constant fear of invasion by one of the most authoritarian and aggresive powers in the world to go fuck themselves--and it's all right because they don't belong to the U.N.?
Nixon did it because he was a shortsided coward. We thought China would make a great ally in the cold war. It didn't.
Exactly why should Taiwan be listed?
Because they are a free and democratic people. Look up a bit. Even Palestine is listed--and that is hardly free and democratic, not even a member of the U.N.
Haven't tried it. It might be nice, and I should look into it before I comment on the messy control center more times.
"because no one person or group can say whether the options are good or bad"
Sure I can. It IS possible to say what options are good or bad.
If these two are fulfilled, then it is a bad option:
1. If the option is wanted by a really small minority.
2. If it is just a minor inconvenience for the minority, and they can change the way they work.
"They're only crack options to YOU. To somebody else, they're good options - that's why they exist."
They are bad too me because they make it seriously difficult to find the good options.
"You should be complaining about the design of configuration tools, not the abundance of options."
Show me ONE SINGLE example of what you are talking about here. I challenge you, show me ONE example of good configuration tools that have the kind of abundance we are talking about here.
It is impossible to create a nice preference dialog for a big system if you do not set a certain limit on the number of options.
You should not need some sort of search-facility just to configure your desktop.
That's true, the KDE project isn't too big on making individual components that separate well from the rest of the project. But the entire point of the project is to have some kind of unity within the system. Say what you will about CORBA, it's hideously complex and overkill for the desktop. The KDE solution is lighter weight (not crippled-by-design any more than XML is crippled in relation to SGML) solution to their particular problem.
The KDE team didn't need something like a standalone signals/slots library because it's built in to the Qt toolkit which the whole project is based on. Why reinvent the wheel?
As for "a powerful C portability/utility library", I don't know what you mean besides libc, which isn't a part of gnome any more than libstdc++ is a part of KDE.
Pango is great, no doubt about it, but on the other hand KDE doesn't have shabby international support. They also didn't have someone from Redhat who was paid to write Pango for them.
On the other hand, why doesn't Gnome give us a complete development environment on the scale of KDevelop? A unified office suite (no, Open Office doesn't count because it's not really part of Gnome).
KDE doesn't focus on providing standalone libraries for uses outside the project, but they provide a hell of a toolset for use inside the project. DCOP is simple and powerful. Kparts was ready and in extensive use well before Bonobo. The C++ object model is inherently easier to work with than the hacked on gtk C pseudo-object model for UI programming. The KDevelop environment is the best GUI development environment on Linux. And that's just for developers. The whole system is very well unified, which is the benefit of the project's focus. Whether or not you like it is a whole other issue. They have done a great job at making a unified system for both developers and users on *NIX. You can't say this as well about Gnome, with its shifting window managers (Metacity is the third standard one in the project's lifetime?)and multiple Office programs.
Perhaps that's why so many of the KDE people are mad about the whole Bluecurve thing. They had already done a great job at making a unified desktop system, and to see it merged with Gnome in the name of unity was perhaps a bit insulting.
"I may not have morals, but I have standards."