Most people at work here hate Open Office and like MS Office.
Well, there are a few factors here:
1.) OpenOffice is different. People always react negatively to change that doesn't necessarily have an immediately tangible benefit.
2.) OpenOffice 1.1 does not offer near as many features as the latest MS Office and still has a lot of rough edges. Anyone who works a heavy paper-pushing desk job and is an expert on MS Office is not a candidate for OO yet. Then again, isn't the fact that such jobs still exist an indication of technology's failure?
3.) OpenOffice is much younger than MS Office. It was reborn as an Open Source project in 2001 and has only really become a quality product since the 1.1 release. Analogous to Mozilla, the work being done contains a lot of re-writing and re-architecting. It will be at least another 2 years before OO has it's own "Firefox phenomenon" unless it can garner some extra commercial backing.
4.) OpenOffice aspires to be more than just a clone of MS Office. Because of that, some of the underlying technology development is getting priority and this is not tangible to end users. What will be tangible is when this technology allows OpenOffice components to be embedded in rich web applications.. say in combination with a web-based document processing system that completely re-defines our current notion of "word processing" and "spreadsheets." Think: MS SharePoint on steroids.. for free.
Having said that, I know several people who have gotten used to OpenOffice and don't mind it at all. Fact is, a very large percentage of Office users don't need more than the very basics. Having followed the test releases, 2.0 is going to be a significant step forward in terms of usability and overall polish.
The apps are generally of better quality than stuff found in the Linux world.
What apps are better quality? Your warez copy of Photoshop vs. Gimp? Sorry, that doesn't count. You can't just make generic statements like this. There is both good and bad software for every platform.
The GUI and applications all use the same user interface, which means that you don't have wildly differing interfaces such as is the case of GTK+ and KDE apps.
Most people generally use either GNOME or KDE apps, so that's not a fair comparison. There is also a trend towards unifying the GUI, such that it won't matter. I could just as easily complain that Apple chose to use a wildly different interface than KDE and GNOME. And worse, their interface is proprietary such that OSX apps can't be easily ported to other platforms.
The OS is incredibly easy to configure compared to the various competing KDE/Gnome distros.
This is a given, but what if all the people who had thrown in the towel and bought Macs had instead contributed to a project aimed at fixing the problem instead.
The OS, apps and hardware are tightly integrated, which means that problems like hardware compatibility don't exist.
It also means the hardware is more expensive and less diverse. Again, this is no excuse for jumping ship instead of fixing the problems.
The software and hardware are both of high quality, which really means something if you've used Dell or no name brands.
Half of my tech-savvy buddies use Macs and they've had no less frequency of hardware problems. Also, the comparison to budget PC brands is unfair. Yes, Macs are better than the "average" PC in terms of quality. But if you know what you're buying, you can get the same quality in a PC.
In the end, Apple is enjoying a temporary boost in the consumer sales due to the Open Source community falling asleep at the wheel in the area of GUIs. But it won't last. And PCs running Linux have far more future potential in the business world..
There's a huge difference between knowing how to mess with the Linux CLI, and wanting to do so. Administrating a Linux box is a terrific learning experience, but once you've learned what you wanted to know the "fun" of maintaining it wears off quickly, and you just want a server that works with a minimum of farting around.
What you say here is a quite valid perspective. However, do you quite realize that if everybody just threw in the towel and bought Macs, the situation would never improve? And then we'd all be stuck paying outrageous prices for hardware and still buying proprietary software. Maybe you don't care about the success of Open Source. Fine.. go buy a Mac for now. But you'll be back. Apple is enjoying a temporary market for *nix enthusiasts while the Open Source community wakes up again and realizes that easy-of-admin is now a much bigger demand than it was in the past.
Say.... you're not dragging out that tired old wives tale about Apple using "proprietary parts" - are you?
If the power supply, motherboard, or CPU in my desktop Mac dies, where precisely, other than Apple, can I get replacement parts? Yes, those are, in fact, still proprietary parts.
Just yesterday, I was thinking about what Sun has in mind with this OpenSolaris thing and CDDL.
Lets look at this realistically. Solaris / OpenSolaris does not have the industry-wide momentum that Linux does. Even if Sun chose GPL and the community rallied around it, it would take several years for OpenSolaris to become a viable product for the mainstream. For example, Linux has an enormous head-start in the area of hardware and architecture support. As a counterexample, look at *BSD. It is very much technically competitive with Linux and is fully open source, but it doesn't have anywhere near the momentum and community that Linux does. I really have a hard time seeing Sun's maneuver as anything but a spoiler for Linux, possibly encouraged by some part of their MS settlement. The truth is, Sun very well could have adopted Linux, as IBM and Novell have. They could have used their experience from Solaris to improve Linux, as IBM has done with their AIX experience. In fact, it seems to me illogical that Sun wouldn't do this from a business perspective.
Now, if Sun had only used a GPL-incompatible copyright license for OpenSolaris, that would almost be understandable, with the perspective of distinguishing it from Linux. But then they started playing the software patent game. Frankly, any company that is a proponent of software patents gets bad marks in my book. Software patents, in addition to being unethical, are the pure antithesis of the spirit and philosophy of open source. There's no middle ground.
We are approaching a point beyond which the art of engineering will be so seriously hindered that only very large companies will be able to approach the creation of software products. Legislative action will be necessary. But will the big companies win that legislative battle? They are winning it so far.
What we probably need is a massive grassroots movement aimed at tearing down software patents in the US. (Similar to the one aimed at preventing them in the EU). We don't have to stand around and accept what we see around us. But that is precisely what most people are doing, based upon the commentary surrounding the Sun/MS patent nonsense. What will it take to get people peeved enough to take action? A combined Sun/MS lawsuit against Linux developers / IBM / etc.?
But what is perhaps more dangerous is the FUD that will arise out of this. And that is why I think we need to destroy software patents now, as a preemptive measure.
The desktop can be fiddled around with, adjusted, etc. Now for his laptop, he probobly wants something that just works and works well.
I would tend to agree on this one.. If I was ever to consider an Apple machine, it would be a laptop.
On the desktop, cost and flexibility reign supreme.. For *nix power users, it's really hard to beat a custom-built $600-700 PC running Debian Linux or FreeBSD. At that price, you're looking at a 2.5Ghz. Athlon 64, 1Gb. RAM, 120Gb. hard drive., and DVD burner. And if something breaks on a PC desktop, it's an inexpensive opportunity to upgrade. With a Mac, you'll pay through the nose for direct replacement parts from Apple.
Of course, the PC laptop market is also improving. Many of today's OEM (Acer, ASUS, etc.) Pentium-M models are cost-effective, reliable, and ACPI-friendly in Linux. I don't think there's a hands-down case for PowerBooks anymore..
Desktop systems are going to be a very interesting market to watch, too, as Microsoft's monopoly gets dented (although probably not broken)..... In the US and most of Europe the numbers will be much different: Microsoft share will drop to probably 80% with Linux and MacOS splitting the difference almost evenly.
You're making one huge assumption here: that the fates of the server and desktop are not inextricable. There is a growing push in the business world for simplified, easier to manage, increasingly server-side solutions. You hit the nail on the head when you said, "the cheapest thing that gets the job done wins," because that is precisely the driving force. Here's why: Linux / Open Source based server-side solutions have the opportunity to make the desktop itself irrelevant. Modern, rich web applications have the capability not only to replace all existing client-side business software, but also to obsolete the very concepts behind traditional, clumsy "office suite" software. Consider, for example, how spreadsheets may be replaced by easily customizable database software and word processing may be replaced by web-based document processing systems. (think: division of content production and typesetting)
If the open source community can pull off this revolutionary paradigm shift instead of chasing MS's desktop-centric taillights, it will dramatically change the future of the industry. Most people are always asking, "What will be the 'killer app' of Open Source on the desktop?" Many think it will be OpenOffice but I disagree. They should be asking, "How can we completely re-define the desktop such that the result is the killer app of Open Source."
This doesn't mean they're attempting to ditch every product Microsoft makes, not by a long shot.
The average, non-technical person uses Windows because that is what came installed on their machine, along with some version of MS Office. If you look at the whole landscape, you'll notice that most people are really not dependent on MS for anything other than Windows and Office. In fact, you could even say that they are only dependent on Windows *because* of Office.
Beating MS has absolutely nothing to do with "what people are used to," because that is always changing anyhow. (ex. compare Win98 to XP) It has everything to do with replacing or obsoleting the need for Office. And beat them we will.
Although another interesting factor is your note that 25% use Macs. It's interesting because this represents the next generation of users. If that number is true, it is a huge indication that people are really fed up with MS nonsense and are even willing to pay more for an alternative -- even an alternative that is far more different than Windows than Linux + KDE.
In a way, perhaps Firefox for Windows is harming Linux.
Mozilla/Firefox is an applications platform that makes the operating system irrelevant. "Oh sure," you say, "that's what everybody predicted in the 90's for the web, but it never happened." But several things have changed since then:
1.) Mature XHTML, CSS, DOM, and JS/ECMAscript standards.
2.) XUL. It is now possible to write very "native feeling" applications that run in Mozilla and do not require any special components or intepreters to be installed first.
3.) Server-side Java/J2EE as applied to web applications is now a mature and efficient development tool thanks to Open Source efforts like Spring, Hibernate, and the Apache/Jakarta project.
So, in the end, what Firefox is doing is preparing the world for a new generation of web applications. It's not just a nicer / safer way to browse the public web.
If you've really been a Linux user since '97 (but I'm thinking this is just bait), you really don't have much incentive to "jump ship" to Apple err.. OSX. You certainly don't need the user-friendly config tools if you survived back then. The slick eye-candy? Not likely.. Your use habits are more CLI-centric than most people. Furthermore, you would know that X.org + KDE/GNOME will have caught up in another 1-2 years but that really doesn't add much value anyhow. (composited window transparency, which is available today, is the only notable exception)
I've been using Linux since '96 and I could care less about OSX, besides adopting a couple of its neat GUI ideas into the Free desktop. I can understand Windows users wanting to switch -- for them it's a huge leap forward in all aspects. But for us long time Linux users, it's just another mildly interesting member of the Unix family tree.
Before Sun, there were no rules. They pioneered the use of open standards.
We're talking about Open Source, not ancient *nix history. There were many contributors toward the pioneering of what we now consider to be "open standards."
What community? All these organisations do what they do for commercial reasons, not part of a worthy charitable motive. Its ALL politics - by IBM, by Novell, by Sun.
The word "community" does not imply charity or lack of commercial motive. When I speak of the "open source community" I care not about what underlying motives members may hold. It is the outcome that matters. IBM and Novell directly compete with each other, yet they are part of the same community because they share a common goal. When you distill it down the most fundamental economics, Open Source is about mutual benefit through agreeing to share the workload.
..not just making a few gestures over the past few years, as IBM and Novell.
Considering that Linux has more future than OpenSolaris, unless Sun changes the license, which gestures are worth more? And incidentally, IBM has contributed far more true open source code than Sun.
Two products alone have changed the IT industry: Open Office and Java. Both were given away FREE.
OpenOffice was most likely given away because Sun really didn't know what to do with the mess of a codebase they acquired from Star Division. It was the last ditch effort, following in the wise footsteps of Netscape with Mozilla. However, did you know that, in the Sun-MS settlement terms made public, there is explicit exclusion of OpenOffice (but not StarOffice) from the legal protection umbrella? Regardless of the fact that MS would probably never actually attempt to sue OO users, this manuever was purposeful and agreed upon by both parties. It effectively gives both Sun and Microsoft a trump card to play should OO become "too popular." Some commitment, eh?
Java is not free and Sun has made numerous efforts to prevent it from becoming free. I'm not talking about Sun's own implementation -- who cares.. it's not that hard to write an alternative. I'm talking about the language specification itself and supposed patents Sun holds to make implementation possible. It would delight me greatly to see IBM release a fully GPL or MPL implementation of *ava and then force Sun into a patent-disarming settlement -- assuming there aren't any workarounds. Java is a wonderful language, but it will not move significantly outside of the enterprise datacenter until there is a fully Open Source implementation that is both more efficient and which can be distributed with every copy of Linux and Mozilla. Remember that.NET/C# is an increasingly capable challenger. Sun would be wise to not let MS "pull an IE" on Java.
Why should Exchange of Ideas mean 'let me have all your code'?
Well, it's optimal for all parties involved, but that's not even the point. If this were merely a matter of incompatible copyright licenses, such that cross-polination with Linux would not be possible, it would not be such a big deal. The big problem is that Sun is only granting use of their supposedly-important patents to those who use *their* copyright license. This is what I mean by trying to block the free exchange of ideas. If somebody comes up with an innovation in Linux, Sun will implement it in their own code. But if Sun comes up with something, they'll patent it to try to prevent its use in Linux. See the problem? They want to create their own little island where they still have control.
but this attitude of 'unless Sun does things exactly our way, they are not to be trusted' is rather immature.
Actually, I think it's Sun's business model of control through software patents that is immature. Most of the rest of the software industry is quite ready to ditch this hideous and unethical relic.
This is plain nonsense. Sun has been a pioneer in open systems, open source and encouraging standards for decades and donating free code and products to the industry. Without them, the IT industry would be a much poorer place. Linux certainly would not be as popular.
The question at hand is trust, not whether Sun has made valuable contributions to the IT industry. (which they have) Of course, as far as I am aware, OpenOffice is the only codebase contribution thus far which falls under the generally accepted definition of open source. Open Source means more than "the code is available to look at" and "you can help us improve it."
But of course, because they are a commercial company and don't follow the pure religion of GPL, and don't understand that Linux is the only true perfect operating system for all situations, we have to distrust them.
There you go again putting words in my mouth. I never mentioned any of those things as reasons why I feel Sun cannot be trusted. The core reason is this: Sun doesn't like to play along without first changing the well-accepted rules. Nearly every move they make contains some type of tactic through which to maintain more control than is comfortable. They absolutely refuse to become a true member of the community, as IBM and Novell very clearly have. The fact that Sun is playing the software patent game with OpenSolaris tells me that they still don't "get it." Meanwhile, IBM is effectively destroying their software patent portfolio to allow more room for true Open Source to flourish. The fact that Sun purposely made its OpenSolaris license incompatible with every other open source license (not just the GPL) signifies to me that they have no intention of encouraging a free exchange of ideas both into and out of their project.
And, nothing but millions of lines of open sourced Solaris and tools code...
You mean: millions of lines of code that are legally useless to any other open source project and are thus effectively proprietary to OpenSolaris.
Open your eyes and look at this situation for what it is. Sun wants to have their cake and eat it too. They want open source developers to flock to OpenSolaris, help develop it, but be legally unable to take anything away from it to other projects (namely Linux). Their lawyers no doubt sat around for days trying to figure out a way to look as "open" as possible but still retain full control. No matter which way you cut it, Sun is still operating under the proprietary control mentality. They don't want the community to have any sway in directing their course. This has been plainly obvious over the years with Java, so it should be no surprise that they're pulling the same tricks with OpenSolaris.
What's the practical outcome of all this? The majority of people who actually help out on OpenSolaris will be those who currently use it. Existing Linux developers certainly aren't going to jump ship. And since Linux is where all the momentum is today, why should somebody new to the game get involved with OpenSolaris instead?
The most disgusting thing about the whole scenario is that Sun is trying to use software patents as their primary means of control. Combined with possible ramifications of the Sun-MS settlement terms, which are under tight wraps, this is simply not a company I believe we can trust.
sounds more like somebody took the most cost-effective option available that did the job at hand
Designing with web standards is the cost effective option. Obviously they hired the wrong people to write their internal apps (or had nobody to give them proper guidance) and now they're paying the price. It is almost always more cost effective long-term to maximize your future options. Flexibility yields efficiency. In the case of web standards, it's cheaper even in the short term to develop with strict conformance because this methodology gives you a way to test the results. "It looks right in IE" works until something breaks or IE gets updated.
And cut the crap about "zealots." If you don't know what you're talking about, don't even bother posting.
Not gonna happen. Seriously, if the new desktop used (for example) Gnome as their base, why would KDE-developers work on it? They would look at it, and think that "We already have a working desktop. Why should we throw it away and work on their desktop instead?". Yes, they would bring along some of their technology, but in the end, they would still abandon their desktop.
Obviously, nothing is going to change overnight. But that doesn't mean that change isn't necessary. I think the realistic answer is this: Both projects need to set a standards base goal for say.. 3 years from now. By that time, there should be no such thing as applications "developed for" either GNOME or KDE. As stated prior, that's really where the big problem is anyhow -- duplicate applications. Instead, applications should be written such that they are fully neutral to the desktop environment and simply use the services made available. (which might not necessarily be provided by GNOME or KDE anyhow!) Incidentally, this is the direction that all cutting edge technologies are moving already, whether from the Java or.NET arena.
And as I said earlier, today's concept of "the desktop" is dying anyhow..
My comment regarding technological differences (language, toolkit etc.) is also relevant to other things besides those two. What about other technologies? KDE uses DCOP, Gnome uses CORBA. KDE uses kparts, Gnome uses Bonobo.
That's true.. they each take a slightly different approach, but the end result is the same. Therefore, they are technologically equivalent. Neither is significantly behind the "state of the art" in any area.
There are simply so much differences between the two that merging them is not really possible. And you can't really merge the developer-base and start a new project, since some developers would prefer Gnome-style, whereas other would prefer KDE-style.
I'm not suggesting a merge because that would, indeed, be impossible. However, there should be a way to compromise between the two. Since there's no dominant technological advantage to developing for either GNOME or KDE, we're talking solely about personal preferences here. Sorry, but that's really not as important as the bigger picture. Developers on both sides should suck it up and agree to disagree but still work together on a single, unified desktop environment. As I mentioned before, it's not so much about KDE vs. GNOME. It's the fact that for every KDE app, somebody wastes time writing a nearly identical GNOME app and vice-versa. In the end, both apps do the exact same thing. They don't really compete with each other. No matter which way you cut it, that's a waste.
So the new desktop would use both GTK AND Qt? And kparts and bonobo? etc. etc.
No, the new desktop would be either KDE or GNOME at its base. For example, GNOME developers could port their HCI-based GUI design to Qt and KDE developers could add a layer of abstraction if necessary allowing the key GUI components to be swapped out.
Isn't that wasted resources? And I would guess that it would make the end-result even more resource-hungry, since you would have two separate, yet functionally similar libraries/technologies running in the background. All that would make the desktop even more bloated and it would be a nightmare to maintain.
And we don't have that today? Honestly, how many people use strictly GNOME or strictly KDE apps? There's always one or two apps that one desktop simply has the better version of. It sure is a shame that all those apps aren't competing within the same desktop environment. For example, I wouldn't give KMail unfair weight over Evolution just because I use KDE and that makes it more efficient for me. See how this wasteful divide can actually *diminish* competition?
Well, they are not THAT disconnected...
Yes, the situation is improving slightly. Of course the ultimate irony is that in the next 10-15 years, the today's concept of a "desktop" is going to be obliterated by both server-centric and distributed technologies. How's that? Well, for instance, users aren't going to be so much worried about "finding / organizing files on their local disks" Most data will be stored in a centralized repository with rich association and context sensitivity. Office suites will be replaced by web-enabled document management systems and extremely rich database tools. Most new applications will be written for the platform that evolves out of today's web browsers. (think: vector graphics, full multimedia capabilities, etc.) Ironically, Open Source may never usurp the desktop from Microsoft. It may simply obsolete it before they can adequately respond. By the time that happens, KDE vs. GNOME will be so irrelevant that either one will fade away or they will have merged into the same technology. In the meantime, it's just a waste, IMO.
Trying to build something specifications-equivalent in a PC involves buying a low-end processor and a small motherboard to match (not to mention the other components), and I don't know if it can be done.
There's one factor that nobody has pointed out yet.. Why does this much-desired SFF machine have to be a standalone powerhouse? I can think of numerous low-powered PC-based solutions that are much smaller and cheaper than the Mac mini. (you could build some *inside* an LCD monitor)
What's the catch? They would be diskless, net-booted thin-client machines and rely upon a standard desktop machine elsewhere in the house. Think Linux + FreeNX + 802.11g. The hardware specs necessary for this are quite minimal.. Pentium-class CPU, 2D video chipset, 32Mb. RAM. We're talking about free recycled parts if you're a DIY hardware hacker. Otherwise, this should be possible at under $100 with economies of scale.
Sure, you wouldn't be able to play 3D games or watch movies on such a terminal, but it would be just fine for internet apps, open office, playing music, viewing photos, etc. Honestly, who is actually going to watch DVD movies or play games on a kitchen counter computer? And otherwise, why buy a Mac Mini for applications where there is plenty of room for a mid-ATX PC that costs $300 and has comparable. specs. (Granted, I only run Linux and don't feel the draw to OSX like many Windows users do.. To me, a PC is just the cheapest Linux box available.)
In case anyone is wondering, NX is a remote desktop protocol similar to but much, much faster (lower bandwidth) than VNC or straight X11. You can run it comfortably over a dial-up connection. Any 802.11 wireless link is more than adequate.
Not really. Gnome is written in C, KDE is written in C++. Gnome uses GTK+, KDE uses Qt. What makes you think that Gnome-hackers would be good KDE-hackers, or vice versa? I mean, the two are technologically quite different.
Despite the fact that GNOME is written in C, they have developed a standard set of extensions that allow developers to use object-oriented design. Some would call that a hack. Others would note that C++ itself was a hack. (: Point is, the language difference is not an issue. There are also GNOME bindings to just about every major language in existance. (same with KDE) Qt vs GTK+? Well, Qt has better API documentation and does a lot more in itself -- GTK+ apps rely on dozens of other libraries. In the end, you can get the same functionality out of both.
Each group has created a desktop according to their vision of what the desktop should be like. And they apparently have quite different visions. How exactly would you merge those two?
Modular design. Use all the same core libraries, but have teams work on pluggable GUI components that suit their taste. Now everybody is happy and the only 'duplicated' effort is that which is actually unique to alternative ideas. This is my contention with the KDE vs. GNOME situation. It's fully possible to accommodate diverse ideas without having two huge, separate, disconnected communities.
Right now Gnome and KDE provide each other some good competition. having one big project with no competition is not necessarily a good thing. Just look at what happened with Xfree! it stagnated for years. and users had no real alternative to it.
I wholly disagree with this argument. There's no reason why competition cannot come from within a community (such as competing sets of pluggable GUI components as prior mentioned). The reason why XFree stagnated was that its development was effectively closed to outsiders. There was no way to get involved and then introduce change from within -- no opportunity for 'new blood' and fresh ideas. Compare to the Linux kernel. There are hundreds of people competing *within* this community to get their improvements accepted into the mainstream codebase. In some cases, the end user can choose which components to use. (think of all the competing filesystems!)
Regarding performance and stability, this is really beyond the scope of reasonable discussion here. It suffices to say that based on what I have seen, there is a lot of room for code optimization in both KDE and GNOME. (Fortunately it is happening, if slowly..) As for stability, I've found dozens of bugs in Kmail and Konqueror. Most don't cause a crash, but they are still annoyingly incorrect operation.
Gnome's still OK - I certainly wouldn't make fun of someone for using it - but understand that quite a few advanced users prefer KDE for reasons that have nothing to do with Windows.
That's exactly it.. advanced users, myself included, usually prefer KDE because it has a richer interface (more widgets, more context sensitivity, more features, more options, more extensibility, etc.) And yet, from what I've seen, most Windows users are more comfortable with using KDE than GNOME because it has a more familiar file dialog box, file manager, etc. So you have a huge community developing for the needs of a very tiny population. It just doesn't make sense. Software today can/should be modular enough that you can plug in alternative components when special needs arise. (such as user-specific sets of HCI guidelines)
I'm certain these developers that volunteer their time are eagerly awaiting your consent as to what projects they may work on.
So, let me get this straight: I'm not allowed to express a valid opinion, but you're allowed to assume that I must be arrogant and must want to tell people what they're allowed to do?
Seeing as how I did nothing of the sort, stop insinuating and come up with a good response.
It's amazing how on Slashdot, the contrary opinion always gets modded up to +5 regardless of how baseless it is.
The really remarkable thing is that in spite of having only a fraction of the corporate support KDE is far more usable. Yes, a few things are clumsier than I would like, but they seem to have avoided the completely idiotic design decisions that GNOME has made (the spatial browser, the hideous file selector, eliminating user-visible preferences to an extreme).
The really pathetic thing is that GNOME and KDE today are pretty much duplicate efforts. This situation has become a terrible waste of community resources. From a technical perspective, there is no significant advantage to either platform. From a user perspective, most people are more comfortable with KDE because it is closer to Windows, which they are used to. GNOME is idealistic; KDE is practical. Guess which more people actually use.
That being said, KDE needs some serious improvement in few performance areas and the stability of apps under its umbrella. Kmail and Konqueror come to mind first..
This is the problem with Linux developers. Developer-centric thinking, not user-centric thinking. Think like an Apple programmer for a few minutes here:
This is the problem with (most) Mac users. Proprietary thinking, not Open Source thinking. so... You've obviously never used a modern Linux distro based on your comments. Besides the ones about not caring about performance, every single one of your points assumes that we're talking about proprietary software from a vendor. In the world of Linux and Open Source, software management is fully automated on all modern distributions. Forget CD's, forget download sites, forget dependancy management, forget compiling, forget having to worry about vendors fixing their software and then sending you manual updates.
If I want a program called "xyz" installed my system, I run one command and (typically) without any further intervention, the following happens: - the latest version of xyz is downloaded from a central repository - any dependancies of xyz are automatically downloaded and installed from the same repository - if those dependancies broke any other software, new versions of that software are installed as well - xyz is installed and configured
It just doesn't get any more "user-centric" than that. All I need is the name of the program and poof.. it's installed. Here's the best part: with one command I can automatically update every piece of software on my system in one fell swoop -- including all components of the OS. Try that on a Mac.
No, for the reference, I'm not anti-Mac / anti-Apple.
"Bundling applications," as defined as giving every app it's own copies of used libraries, is just plain stupid if at all avoidable.
On the other hand, I may have originally misread what you were getting at. So, if that's not what you meant, I'm afraid my answer is different..
Bundling apps / libraries in the sense of giving them their own directory and then symlinking back to some common path like/System/Libraries/KDE or what have you, is not such a bad idea. Check out GoboLinux for a distro aimed at trying this approach.
However point number 5 from the original post is the same.. Let other people share the workload. Building a system completely from source is a waste of your time. You'd do a lot more good to help the GoboLinux folks instead of duplicating their work on your own.
Of course, that being said, a good package management system pretty much makes the administrative end of this discussion moot because you can easily track and remove files as needed even though they are dumped into common directories. Projects like GoboLinux just use the reverse approach. All components get a separate directory and it's the symlinks that get dumped into common directories. I think there are pros and cons to both approaches..
Slashdot Mirror
Most people at work here hate Open Office and like MS Office.
Well, there are a few factors here:
1.) OpenOffice is different. People always react negatively to change that doesn't necessarily have an immediately tangible benefit.
2.) OpenOffice 1.1 does not offer near as many features as the latest MS Office and still has a lot of rough edges. Anyone who works a heavy paper-pushing desk job and is an expert on MS Office is not a candidate for OO yet. Then again, isn't the fact that such jobs still exist an indication of technology's failure?
3.) OpenOffice is much younger than MS Office. It was reborn as an Open Source project in 2001 and has only really become a quality product since the 1.1 release. Analogous to Mozilla, the work being done contains a lot of re-writing and re-architecting. It will be at least another 2 years before OO has it's own "Firefox phenomenon" unless it can garner some extra commercial backing.
4.) OpenOffice aspires to be more than just a clone of MS Office. Because of that, some of the underlying technology development is getting priority and this is not tangible to end users. What will be tangible is when this technology allows OpenOffice components to be embedded in rich web applications.. say in combination with a web-based document processing system that completely re-defines our current notion of "word processing" and "spreadsheets." Think: MS SharePoint on steroids.. for free.
Having said that, I know several people who have gotten used to OpenOffice and don't mind it at all. Fact is, a very large percentage of Office users don't need more than the very basics. Having followed the test releases, 2.0 is going to be a significant step forward in terms of usability and overall polish.
The apps are generally of better quality than stuff found in the Linux world.
What apps are better quality? Your warez copy of Photoshop vs. Gimp? Sorry, that doesn't count. You can't just make generic statements like this. There is both good and bad software for every platform.
The GUI and applications all use the same user interface, which means that you don't have wildly differing interfaces such as is the case of GTK+ and KDE apps.
Most people generally use either GNOME or KDE apps, so that's not a fair comparison. There is also a trend towards unifying the GUI, such that it won't matter. I could just as easily complain that Apple chose to use a wildly different interface than KDE and GNOME. And worse, their interface is proprietary such that OSX apps can't be easily ported to other platforms.
The OS is incredibly easy to configure compared to the various competing KDE/Gnome distros.
This is a given, but what if all the people who had thrown in the towel and bought Macs had instead contributed to a project aimed at fixing the problem instead.
The OS, apps and hardware are tightly integrated, which means that problems like hardware compatibility don't exist.
It also means the hardware is more expensive and less diverse. Again, this is no excuse for jumping ship instead of fixing the problems.
The software and hardware are both of high quality, which really means something if you've used Dell or no name brands.
Half of my tech-savvy buddies use Macs and they've had no less frequency of hardware problems. Also, the comparison to budget PC brands is unfair. Yes, Macs are better than the "average" PC in terms of quality. But if you know what you're buying, you can get the same quality in a PC.
In the end, Apple is enjoying a temporary boost in the consumer sales due to the Open Source community falling asleep at the wheel in the area of GUIs. But it won't last. And PCs running Linux have far more future potential in the business world..
There's a huge difference between knowing how to mess with the Linux CLI, and wanting to do so. Administrating a Linux box is a terrific learning experience, but once you've learned what you wanted to know the "fun" of maintaining it wears off quickly, and you just want a server that works with a minimum of farting around.
What you say here is a quite valid perspective. However, do you quite realize that if everybody just threw in the towel and bought Macs, the situation would never improve? And then we'd all be stuck paying outrageous prices for hardware and still buying proprietary software. Maybe you don't care about the success of Open Source. Fine.. go buy a Mac for now. But you'll be back. Apple is enjoying a temporary market for *nix enthusiasts while the Open Source community wakes up again and realizes that easy-of-admin is now a much bigger demand than it was in the past.
Say.... you're not dragging out that tired old wives tale about Apple using "proprietary parts" - are you?
If the power supply, motherboard, or CPU in my desktop Mac dies, where precisely, other than Apple, can I get replacement parts? Yes, those are, in fact, still proprietary parts.
Just yesterday, I was thinking about what Sun has in mind with this OpenSolaris thing and CDDL.
Lets look at this realistically. Solaris / OpenSolaris does not have the industry-wide momentum that Linux does. Even if Sun chose GPL and the community rallied around it, it would take several years for OpenSolaris to become a viable product for the mainstream. For example, Linux has an enormous head-start in the area of hardware and architecture support. As a counterexample, look at *BSD. It is very much technically competitive with Linux and is fully open source, but it doesn't have anywhere near the momentum and community that Linux does. I really have a hard time seeing Sun's maneuver as anything but a spoiler for Linux, possibly encouraged by some part of their MS settlement. The truth is, Sun very well could have adopted Linux, as IBM and Novell have. They could have used their experience from Solaris to improve Linux, as IBM has done with their AIX experience. In fact, it seems to me illogical that Sun wouldn't do this from a business perspective.
Now, if Sun had only used a GPL-incompatible copyright license for OpenSolaris, that would almost be understandable, with the perspective of distinguishing it from Linux. But then they started playing the software patent game. Frankly, any company that is a proponent of software patents gets bad marks in my book. Software patents, in addition to being unethical, are the pure antithesis of the spirit and philosophy of open source. There's no middle ground.
We are approaching a point beyond which the art of engineering will be so seriously hindered that only very large companies will be able to approach the creation of software products. Legislative action will be necessary. But will the big companies win that legislative battle? They are winning it so far.
What we probably need is a massive grassroots movement aimed at tearing down software patents in the US. (Similar to the one aimed at preventing them in the EU). We don't have to stand around and accept what we see around us. But that is precisely what most people are doing, based upon the commentary surrounding the Sun/MS patent nonsense. What will it take to get people peeved enough to take action? A combined Sun/MS lawsuit against Linux developers / IBM / etc.?
But what is perhaps more dangerous is the FUD that will arise out of this. And that is why I think we need to destroy software patents now, as a preemptive measure.
The desktop can be fiddled around with, adjusted, etc. Now for his laptop, he probobly wants something that just works and works well.
I would tend to agree on this one.. If I was ever to consider an Apple machine, it would be a laptop.
On the desktop, cost and flexibility reign supreme.. For *nix power users, it's really hard to beat a custom-built $600-700 PC running Debian Linux or FreeBSD. At that price, you're looking at a 2.5Ghz. Athlon 64, 1Gb. RAM, 120Gb. hard drive., and DVD burner. And if something breaks on a PC desktop, it's an inexpensive opportunity to upgrade. With a Mac, you'll pay through the nose for direct replacement parts from Apple.
Of course, the PC laptop market is also improving. Many of today's OEM (Acer, ASUS, etc.) Pentium-M models are cost-effective, reliable, and ACPI-friendly in Linux. I don't think there's a hands-down case for PowerBooks anymore..
Desktop systems are going to be a very interesting market to watch, too, as Microsoft's monopoly gets dented (although probably not broken). .... In the US and most of Europe the numbers will be much different: Microsoft share will drop to probably 80% with Linux and MacOS splitting the difference almost evenly.
You're making one huge assumption here: that the fates of the server and desktop are not inextricable. There is a growing push in the business world for simplified, easier to manage, increasingly server-side solutions. You hit the nail on the head when you said, "the cheapest thing that gets the job done wins," because that is precisely the driving force. Here's why: Linux / Open Source based server-side solutions have the opportunity to make the desktop itself irrelevant. Modern, rich web applications have the capability not only to replace all existing client-side business software, but also to obsolete the very concepts behind traditional, clumsy "office suite" software. Consider, for example, how spreadsheets may be replaced by easily customizable database software and word processing may be replaced by web-based document processing systems. (think: division of content production and typesetting)
If the open source community can pull off this revolutionary paradigm shift instead of chasing MS's desktop-centric taillights, it will dramatically change the future of the industry. Most people are always asking, "What will be the 'killer app' of Open Source on the desktop?" Many think it will be OpenOffice but I disagree. They should be asking, "How can we completely re-define the desktop such that the result is the killer app of Open Source."
This doesn't mean they're attempting to ditch every product Microsoft makes, not by a long shot.
The average, non-technical person uses Windows because that is what came installed on their machine, along with some version of MS Office. If you look at the whole landscape, you'll notice that most people are really not dependent on MS for anything other than Windows and Office. In fact, you could even say that they are only dependent on Windows *because* of Office.
Beating MS has absolutely nothing to do with "what people are used to," because that is always changing anyhow. (ex. compare Win98 to XP) It has everything to do with replacing or obsoleting the need for Office. And beat them we will.
Although another interesting factor is your note that 25% use Macs. It's interesting because this represents the next generation of users. If that number is true, it is a huge indication that people are really fed up with MS nonsense and are even willing to pay more for an alternative -- even an alternative that is far more different than Windows than Linux + KDE.
In a way, perhaps Firefox for Windows is harming Linux.
Mozilla/Firefox is an applications platform that makes the operating system irrelevant. "Oh sure," you say, "that's what everybody predicted in the 90's for the web, but it never happened." But several things have changed since then:
1.) Mature XHTML, CSS, DOM, and JS/ECMAscript standards.
2.) XUL. It is now possible to write very "native feeling" applications that run in Mozilla and do not require any special components or intepreters to be installed first.
3.) Server-side Java/J2EE as applied to web applications is now a mature and efficient development tool thanks to Open Source efforts like Spring, Hibernate, and the Apache/Jakarta project.
So, in the end, what Firefox is doing is preparing the world for a new generation of web applications. It's not just a nicer / safer way to browse the public web.
When is the right time to jump on the Apple ship?
If you've really been a Linux user since '97 (but I'm thinking this is just bait), you really don't have much incentive to "jump ship" to Apple err.. OSX. You certainly don't need the user-friendly config tools if you survived back then. The slick eye-candy? Not likely.. Your use habits are more CLI-centric than most people. Furthermore, you would know that X.org + KDE/GNOME will have caught up in another 1-2 years but that really doesn't add much value anyhow. (composited window transparency, which is available today, is the only notable exception)
I've been using Linux since '96 and I could care less about OSX, besides adopting a couple of its neat GUI ideas into the Free desktop. I can understand Windows users wanting to switch -- for them it's a huge leap forward in all aspects. But for us long time Linux users, it's just another mildly interesting member of the Unix family tree.
Before Sun, there were no rules. They pioneered the use of open standards.
..not just making a few gestures over the past few years, as IBM and Novell.
.NET/C# is an increasingly capable challenger. Sun would be wise to not let MS "pull an IE" on Java.
We're talking about Open Source, not ancient *nix history. There were many contributors toward the pioneering of what we now consider to be "open standards."
What community? All these organisations do what they do for commercial reasons, not part of a worthy charitable motive. Its ALL politics - by IBM, by Novell, by Sun.
The word "community" does not imply charity or lack of commercial motive. When I speak of the "open source community" I care not about what underlying motives members may hold. It is the outcome that matters. IBM and Novell directly compete with each other, yet they are part of the same community because they share a common goal. When you distill it down the most fundamental economics, Open Source is about mutual benefit through agreeing to share the workload.
Considering that Linux has more future than OpenSolaris, unless Sun changes the license, which gestures are worth more? And incidentally, IBM has contributed far more true open source code than Sun.
Two products alone have changed the IT industry: Open Office and Java. Both were given away FREE.
OpenOffice was most likely given away because Sun really didn't know what to do with the mess of a codebase they acquired from Star Division. It was the last ditch effort, following in the wise footsteps of Netscape with Mozilla. However, did you know that, in the Sun-MS settlement terms made public, there is explicit exclusion of OpenOffice (but not StarOffice) from the legal protection umbrella? Regardless of the fact that MS would probably never actually attempt to sue OO users, this manuever was purposeful and agreed upon by both parties. It effectively gives both Sun and Microsoft a trump card to play should OO become "too popular." Some commitment, eh?
Java is not free and Sun has made numerous efforts to prevent it from becoming free. I'm not talking about Sun's own implementation -- who cares.. it's not that hard to write an alternative. I'm talking about the language specification itself and supposed patents Sun holds to make implementation possible. It would delight me greatly to see IBM release a fully GPL or MPL implementation of *ava and then force Sun into a patent-disarming settlement -- assuming there aren't any workarounds. Java is a wonderful language, but it will not move significantly outside of the enterprise datacenter until there is a fully Open Source implementation that is both more efficient and which can be distributed with every copy of Linux and Mozilla. Remember that
Why should Exchange of Ideas mean 'let me have all your code'?
Well, it's optimal for all parties involved, but that's not even the point. If this were merely a matter of incompatible copyright licenses, such that cross-polination with Linux would not be possible, it would not be such a big deal. The big problem is that Sun is only granting use of their supposedly-important patents to those who use *their* copyright license. This is what I mean by trying to block the free exchange of ideas. If somebody comes up with an innovation in Linux, Sun will implement it in their own code. But if Sun comes up with something, they'll patent it to try to prevent its use in Linux. See the problem? They want to create their own little island where they still have control.
but this attitude of 'unless Sun does things exactly our way, they are not to be trusted' is rather immature.
Actually, I think it's Sun's business model of control through software patents that is immature. Most of the rest of the software industry is quite ready to ditch this hideous and unethical relic.
heh.. Speaking of MMX, did anyone ever figure out what that *really* stood for? I mean..
.. or perhaps something more creative like:
Matrix Math eXtensions
MultiMedia eXtensions
Massive Marketing eXperiment
Masks Moore's eXtenuation
And if I recall correctly, the official Intel answer was something along the lines of "it doesn't mean anything.. it's just a name." Right.... (:
This is plain nonsense. Sun has been a pioneer in open systems, open source and encouraging standards for decades and donating free code and products to the industry. Without them, the IT industry would be a much poorer place. Linux certainly would not be as popular.
The question at hand is trust, not whether Sun has made valuable contributions to the IT industry. (which they have) Of course, as far as I am aware, OpenOffice is the only codebase contribution thus far which falls under the generally accepted definition of open source. Open Source means more than "the code is available to look at" and "you can help us improve it."
But of course, because they are a commercial company and don't follow the pure religion of GPL, and don't understand that Linux is the only true perfect operating system for all situations, we have to distrust them.
There you go again putting words in my mouth. I never mentioned any of those things as reasons why I feel Sun cannot be trusted. The core reason is this: Sun doesn't like to play along without first changing the well-accepted rules. Nearly every move they make contains some type of tactic through which to maintain more control than is comfortable. They absolutely refuse to become a true member of the community, as IBM and Novell very clearly have. The fact that Sun is playing the software patent game with OpenSolaris tells me that they still don't "get it." Meanwhile, IBM is effectively destroying their software patent portfolio to allow more room for true Open Source to flourish. The fact that Sun purposely made its OpenSolaris license incompatible with every other open source license (not just the GPL) signifies to me that they have no intention of encouraging a free exchange of ideas both into and out of their project.
And, nothing but millions of lines of open sourced Solaris and tools code...
You mean: millions of lines of code that are legally useless to any other open source project and are thus effectively proprietary to OpenSolaris.
Open your eyes and look at this situation for what it is. Sun wants to have their cake and eat it too. They want open source developers to flock to OpenSolaris, help develop it, but be legally unable to take anything away from it to other projects (namely Linux). Their lawyers no doubt sat around for days trying to figure out a way to look as "open" as possible but still retain full control. No matter which way you cut it, Sun is still operating under the proprietary control mentality. They don't want the community to have any sway in directing their course. This has been plainly obvious over the years with Java, so it should be no surprise that they're pulling the same tricks with OpenSolaris.
What's the practical outcome of all this? The majority of people who actually help out on OpenSolaris will be those who currently use it. Existing Linux developers certainly aren't going to jump ship. And since Linux is where all the momentum is today, why should somebody new to the game get involved with OpenSolaris instead?
The most disgusting thing about the whole scenario is that Sun is trying to use software patents as their primary means of control. Combined with possible ramifications of the Sun-MS settlement terms, which are under tight wraps, this is simply not a company I believe we can trust.
sounds more like somebody took the most cost-effective option available that did the job at hand
Designing with web standards is the cost effective option. Obviously they hired the wrong people to write their internal apps (or had nobody to give them proper guidance) and now they're paying the price. It is almost always more cost effective long-term to maximize your future options. Flexibility yields efficiency. In the case of web standards, it's cheaper even in the short term to develop with strict conformance because this methodology gives you a way to test the results. "It looks right in IE" works until something breaks or IE gets updated.
And cut the crap about "zealots." If you don't know what you're talking about, don't even bother posting.
Not gonna happen. Seriously, if the new desktop used (for example) Gnome as their base, why would KDE-developers work on it? They would look at it, and think that "We already have a working desktop. Why should we throw it away and work on their desktop instead?". Yes, they would bring along some of their technology, but in the end, they would still abandon their desktop.
.NET arena.
Obviously, nothing is going to change overnight. But that doesn't mean that change isn't necessary. I think the realistic answer is this: Both projects need to set a standards base goal for say.. 3 years from now. By that time, there should be no such thing as applications "developed for" either GNOME or KDE. As stated prior, that's really where the big problem is anyhow -- duplicate applications. Instead, applications should be written such that they are fully neutral to the desktop environment and simply use the services made available. (which might not necessarily be provided by GNOME or KDE anyhow!) Incidentally, this is the direction that all cutting edge technologies are moving already, whether from the Java or
And as I said earlier, today's concept of "the desktop" is dying anyhow..
My comment regarding technological differences (language, toolkit etc.) is also relevant to other things besides those two. What about other technologies? KDE uses DCOP, Gnome uses CORBA. KDE uses kparts, Gnome uses Bonobo.
That's true.. they each take a slightly different approach, but the end result is the same. Therefore, they are technologically equivalent. Neither is significantly behind the "state of the art" in any area.
There are simply so much differences between the two that merging them is not really possible. And you can't really merge the developer-base and start a new project, since some developers would prefer Gnome-style, whereas other would prefer KDE-style.
I'm not suggesting a merge because that would, indeed, be impossible. However, there should be a way to compromise between the two. Since there's no dominant technological advantage to developing for either GNOME or KDE, we're talking solely about personal preferences here. Sorry, but that's really not as important as the bigger picture. Developers on both sides should suck it up and agree to disagree but still work together on a single, unified desktop environment. As I mentioned before, it's not so much about KDE vs. GNOME. It's the fact that for every KDE app, somebody wastes time writing a nearly identical GNOME app and vice-versa. In the end, both apps do the exact same thing. They don't really compete with each other. No matter which way you cut it, that's a waste.
So the new desktop would use both GTK AND Qt? And kparts and bonobo? etc. etc.
No, the new desktop would be either KDE or GNOME at its base. For example, GNOME developers could port their HCI-based GUI design to Qt and KDE developers could add a layer of abstraction if necessary allowing the key GUI components to be swapped out.
Isn't that wasted resources? And I would guess that it would make the end-result even more resource-hungry, since you would have two separate, yet functionally similar libraries/technologies running in the background. All that would make the desktop even more bloated and it would be a nightmare to maintain.
And we don't have that today? Honestly, how many people use strictly GNOME or strictly KDE apps? There's always one or two apps that one desktop simply has the better version of. It sure is a shame that all those apps aren't competing within the same desktop environment. For example, I wouldn't give KMail unfair weight over Evolution just because I use KDE and that makes it more efficient for me. See how this wasteful divide can actually *diminish* competition?
Well, they are not THAT disconnected...
Yes, the situation is improving slightly. Of course the ultimate irony is that in the next 10-15 years, the today's concept of a "desktop" is going to be obliterated by both server-centric and distributed technologies. How's that? Well, for instance, users aren't going to be so much worried about "finding / organizing files on their local disks" Most data will be stored in a centralized repository with rich association and context sensitivity. Office suites will be replaced by web-enabled document management systems and extremely rich database tools. Most new applications will be written for the platform that evolves out of today's web browsers. (think: vector graphics, full multimedia capabilities, etc.) Ironically, Open Source may never usurp the desktop from Microsoft. It may simply obsolete it before they can adequately respond. By the time that happens, KDE vs. GNOME will be so irrelevant that either one will fade away or they will have merged into the same technology. In the meantime, it's just a waste, IMO.
I trust you reported these on bugs.kde.org?
The repeatable ones, yes. (:
Trying to build something specifications-equivalent in a PC involves buying a low-end processor and a small motherboard to match (not to mention the other components), and I don't know if it can be done.
There's one factor that nobody has pointed out yet.. Why does this much-desired SFF machine have to be a standalone powerhouse? I can think of numerous low-powered PC-based solutions that are much smaller and cheaper than the Mac mini. (you could build some *inside* an LCD monitor)
What's the catch? They would be diskless, net-booted thin-client machines and rely upon a standard desktop machine elsewhere in the house. Think Linux + FreeNX + 802.11g. The hardware specs necessary for this are quite minimal.. Pentium-class CPU, 2D video chipset, 32Mb. RAM. We're talking about free recycled parts if you're a DIY hardware hacker. Otherwise, this should be possible at under $100 with economies of scale.
Sure, you wouldn't be able to play 3D games or watch movies on such a terminal, but it would be just fine for internet apps, open office, playing music, viewing photos, etc. Honestly, who is actually going to watch DVD movies or play games on a kitchen counter computer? And otherwise, why buy a Mac Mini for applications where there is plenty of room for a mid-ATX PC that costs $300 and has comparable. specs. (Granted, I only run Linux and don't feel the draw to OSX like many Windows users do.. To me, a PC is just the cheapest Linux box available.)
In case anyone is wondering, NX is a remote desktop protocol similar to but much, much faster (lower bandwidth) than VNC or straight X11. You can run it comfortably over a dial-up connection. Any 802.11 wireless link is more than adequate.
Not really. Gnome is written in C, KDE is written in C++. Gnome uses GTK+, KDE uses Qt. What makes you think that Gnome-hackers would be good KDE-hackers, or vice versa? I mean, the two are technologically quite different.
Despite the fact that GNOME is written in C, they have developed a standard set of extensions that allow developers to use object-oriented design. Some would call that a hack. Others would note that C++ itself was a hack. (: Point is, the language difference is not an issue. There are also GNOME bindings to just about every major language in existance. (same with KDE) Qt vs GTK+? Well, Qt has better API documentation and does a lot more in itself -- GTK+ apps rely on dozens of other libraries. In the end, you can get the same functionality out of both.
Each group has created a desktop according to their vision of what the desktop should be like. And they apparently have quite different visions. How exactly would you merge those two?
Modular design. Use all the same core libraries, but have teams work on pluggable GUI components that suit their taste. Now everybody is happy and the only 'duplicated' effort is that which is actually unique to alternative ideas. This is my contention with the KDE vs. GNOME situation. It's fully possible to accommodate diverse ideas without having two huge, separate, disconnected communities.
Right now Gnome and KDE provide each other some good competition. having one big project with no competition is not necessarily a good thing. Just look at what happened with Xfree! it stagnated for years. and users had no real alternative to it.
I wholly disagree with this argument. There's no reason why competition cannot come from within a community (such as competing sets of pluggable GUI components as prior mentioned). The reason why XFree stagnated was that its development was effectively closed to outsiders. There was no way to get involved and then introduce change from within -- no opportunity for 'new blood' and fresh ideas. Compare to the Linux kernel. There are hundreds of people competing *within* this community to get their improvements accepted into the mainstream codebase. In some cases, the end user can choose which components to use. (think of all the competing filesystems!)
Regarding performance and stability, this is really beyond the scope of reasonable discussion here. It suffices to say that based on what I have seen, there is a lot of room for code optimization in both KDE and GNOME. (Fortunately it is happening, if slowly..) As for stability, I've found dozens of bugs in Kmail and Konqueror. Most don't cause a crash, but they are still annoyingly incorrect operation.
Gnome's still OK - I certainly wouldn't make fun of someone for using it - but understand that quite a few advanced users prefer KDE for reasons that have nothing to do with Windows.
That's exactly it.. advanced users, myself included, usually prefer KDE because it has a richer interface (more widgets, more context sensitivity, more features, more options, more extensibility, etc.) And yet, from what I've seen, most Windows users are more comfortable with using KDE than GNOME because it has a more familiar file dialog box, file manager, etc. So you have a huge community developing for the needs of a very tiny population. It just doesn't make sense. Software today can/should be modular enough that you can plug in alternative components when special needs arise. (such as user-specific sets of HCI guidelines)
I'm certain these developers that volunteer their time are eagerly awaiting your consent as to what projects they may work on.
So, let me get this straight: I'm not allowed to express a valid opinion, but you're allowed to assume that I must be arrogant and must want to tell people what they're allowed to do?
Seeing as how I did nothing of the sort, stop insinuating and come up with a good response.
It's amazing how on Slashdot, the contrary opinion always gets modded up to +5 regardless of how baseless it is.
The really remarkable thing is that in spite of having only a fraction of the corporate support KDE is far more usable. Yes, a few things are clumsier than I would like, but they seem to have avoided the completely idiotic design decisions that GNOME has made (the spatial browser, the hideous file selector, eliminating user-visible preferences to an extreme).
The really pathetic thing is that GNOME and KDE today are pretty much duplicate efforts. This situation has become a terrible waste of community resources. From a technical perspective, there is no significant advantage to either platform. From a user perspective, most people are more comfortable with KDE because it is closer to Windows, which they are used to. GNOME is idealistic; KDE is practical. Guess which more people actually use.
That being said, KDE needs some serious improvement in few performance areas and the stability of apps under its umbrella. Kmail and Konqueror come to mind first..
This is the problem with Linux developers. Developer-centric thinking, not user-centric thinking. Think like an Apple programmer for a few minutes here:
This is the problem with (most) Mac users. Proprietary thinking, not Open Source thinking. so... You've obviously never used a modern Linux distro based on your comments. Besides the ones about not caring about performance, every single one of your points assumes that we're talking about proprietary software from a vendor. In the world of Linux and Open Source, software management is fully automated on all modern distributions. Forget CD's, forget download sites, forget dependancy management, forget compiling, forget having to worry about vendors fixing their software and then sending you manual updates.
If I want a program called "xyz" installed my system, I run one command and (typically) without any further intervention, the following happens:
- the latest version of xyz is downloaded from a central repository
- any dependancies of xyz are automatically downloaded and installed from the same repository
- if those dependancies broke any other software, new versions of that software are installed as well
- xyz is installed and configured
It just doesn't get any more "user-centric" than that. All I need is the name of the program and poof.. it's installed. Here's the best part: with one command I can automatically update every piece of software on my system in one fell swoop -- including all components of the OS. Try that on a Mac.
No, for the reference, I'm not anti-Mac / anti-Apple.
"Bundling applications," as defined as giving every app it's own copies of used libraries, is just plain stupid if at all avoidable.
/System/Libraries/KDE or what have you, is not such a bad idea. Check out GoboLinux for a distro aimed at trying this approach.
On the other hand, I may have originally misread what you were getting at. So, if that's not what you meant, I'm afraid my answer is different..
Bundling apps / libraries in the sense of giving them their own directory and then symlinking back to some common path like
However point number 5 from the original post is the same.. Let other people share the workload. Building a system completely from source is a waste of your time. You'd do a lot more good to help the GoboLinux folks instead of duplicating their work on your own.
Of course, that being said, a good package management system pretty much makes the administrative end of this discussion moot because you can easily track and remove files as needed even though they are dumped into common directories. Projects like GoboLinux just use the reverse approach. All components get a separate directory and it's the symlinks that get dumped into common directories. I think there are pros and cons to both approaches..