Linked by Thom Holwerda on Sun 3rd Feb 2008 23:09 UTC, submitted by Moochman
KDE This article details the story of a KDE-loving software engineer who was forced to use Windows for his job. "His only hope was that he knew Qt was cross compatible with Windows Linux and Mac, and there was talk that someday, KDE was to be ported to Windows. So he waited. Well, KDE4 was announced and there was much joy. Betas were released and there was much bitching. KDE4.0.0 was released and there again was much joy (and still a little bitching). More importantly an actual honest to goodness Windows port is released. Here follows that engineer's report."
Order by: Score:
v Poor combination
by Liquidator on Mon 4th Feb 2008 00:05 UTC
djame
Member since:
2005-07-08

Come on.... It's shipped on 90 if not 95% computer sold each year.
Anyway, except from the famous joel spolky's "how microsoft lost the api war", I have yet to see any good arguments about microsoft fall

Reply Score: 16

FishB8 Member since:
2006-01-16

Granted, I don't know what the author meant by this, but I personally see windows as a dying platform. Not because of it's user base (which is obviously very dominate) but because there is not much interesting happening as far as development of new and cutting edge ideas are concerned. Quite honestly it has become boring and lethargic. Cutting edge has become too risky and they cater more towards maintaining backwards compatibility than towards the cutting edge.

Reply Score: 11

Chicken Blood Member since:
2005-12-21

Granted, I don't know what the author meant by this, but I personally see windows as a dying platform. Not because of it's user base (which is obviously very dominate) but because there is not much interesting happening as far as development of new and cutting edge ideas are concerned. Quite honestly it has become boring and lethargic. Cutting edge has become too risky and they cater more towards maintaining backwards compatibility than towards the cutting edge.


To contrast, can you tell me about the "cutting edge" ideas that are developing on other platforms?

Reply Score: 5

elsewhere Member since:
2005-07-13

To contrast, can you tell me about the "cutting edge" ideas that are developing on other platforms?


Cross-platform application development, open protocols and open standards?

Ok, not cutting edge, but a refreshing change, none the less... ;)

Reply Score: 9

gustl Member since:
2006-01-19

Very simple:

1. Get yourselves a Kubunto or Knoppix CD.

2. Load KDE

3. Try to find out what you can do with Konqueror (without Konqi ever bloating!):

- File Browsing
- File Browsing, split Window
- Additionally open a directory in a new tab
- now type a http:// URL into the URL bar
- now split the view again
- now type man:ls in the URL bar
- create another tab
- enter ftp://username@your.next.ftp.server.com
- if there is a pdf file, click on it.

You should now have 3 tabs, first with split windows for file browsing, second with web browsing and man page displaying, third with ftp file browsing or an open pdf document.

Now count the buttons in the button bar. Are they more than you have in Windows Explorer? Sure, but not by much, for sure the icon count is low enough to make the statement "it's not bloated" a true one.

And now USE this setup. Copy files here and there - ftp to local, local to ftp.

You will find out, that Konqueror is to Windows Explorer what a Mercedes is to a Yugo. Both can get you from Vienna to Rom, but the Mercedes is less noisy, safer and can play CDs - the Yugo can't.

That is probably the reason why the Author meant Windows is a dying system.

Reply Score: 10

autumnlover Member since:
2007-04-12

Now try to uninstall Konqueror without removing the KDE itself. Then you will see the difference between IE-driven Explorer and Konqueror.

Reply Score: 2

Erunno Member since:
2007-06-22

Now try to uninstall Konqueror without removing the KDE itself. Then you will see the difference between IE-driven Explorer and Konqueror.


That's pretty easy on a distribution which doesn't mark Konqueror as a hard dependency (e.g. Gentoo if my memory doesn't fail me).

Reply Score: 2

glarepate Member since:
2006-01-04

Are you claiming 'Doze is improved in functionality with IE removed? Or that KDE, and possibly Linux itself rather than one of it's several user interfaces, is benefited by removing Konq?

Didn't think so. But if I did run 'Doze I would want to remove the largest vector for infection and compromise of security and possibly even replace the whole Explorer GUI/interface with something more secure, usable and stable. But I don't.

What would your point be if you were to state it without obfuscation by ambiguity?

Reply Score: 1

Almafeta Member since:
2007-02-22

And now USE this setup.


... so you intended to prove the superiority of Windows Explorer with that post?

Reply Score: 4

wirespot Member since:
2006-06-21

He obviously did. If you're the type of person that enjoys fewer features and lower quality software, that s.

Reply Score: 2

google_ninja Member since:
2006-02-05

imo .net 3.5 is by far the most modern and interesting platform to develop on. it isnt even comparable to obj-c on apple, or even worse, c/c++ on linux.

Reply Score: 9

leos Member since:
2005-09-21

imo .net 3.5 is by far the most modern and interesting platform to develop on. it isnt even comparable to obj-c on apple, or even worse, c/c++ on linux.


Hmm.. No. Qt and C++ are just as good as .NET to work with. And the difference is that you can write a whole platform in them, which you can't realistically do with .NET, because your performance will be terrible. Not to mention that you'll be locked in to windows (mono doesn't really count).

Reply Score: 9

tomcat Member since:
2006-01-06

Hmm.. No. Qt and C++ are just as good as .NET to work with. And the difference is that you can write a whole platform in them, which you can't realistically do with .NET, because your performance will be terrible.


Sorry, but that's just so wrong. .NET apps are compiled to native code. "Terrible" perf? Uh, I don't think so.

Not to mention that you'll be locked in to windows (mono doesn't really count).


Huh? Why doesn't mono count? It's open source, it's freely available, and it works. Now. It sounds to me like you're basing your objections primarily on ideology, not reality.

Reply Score: 7

leos Member since:
2005-09-21

Sorry, but that's just so wrong. .NET apps are compiled to native code. "Terrible" perf? Uh, I don't think so.


They still need the runtime. Yes, technically they are JIT compiled, but you still need to load the runtime, which means they're not really native code. Just look at the memory footprint of .NET apps compared to native ones.

Huh? Why doesn't mono count? It's open source, it's freely available, and it works. Now. It sounds to me like you're basing your objections primarily on ideology, not reality.


Mono doesn't really count because it's not an official port of .NET. Microsoft doesn't acknowledge it, and it's not quite compatible with .NET on windows. So for any serious development, it's not a real option.

Reply Score: 6

tomcat Member since:
2006-01-06

They still need the runtime. Yes, technically they are JIT compiled, but you still need to load the runtime, which means they're not really native code.


The .NET runtime is also compiled so, wrong, they ARE native code. It's no different than any other statically or linked native code dependency (ie. C Runtime, Win32, etc).

Just look at the memory footprint of .NET apps compared to native ones.


You can't just make blanket statements like that. Sure, if your native app doesn't do anything, it has a lower memory footprint, but just try doing common things that most native apps do -- like bringing up a common open/save dialog, interacting with common Windows controls -- and then see how quickly your app's footprint increases in size due to DLL dependencies. Try running DEPENDS (http://www.dependencywalker.com/) against your native app and evaluate its true memory footprint. I think that you'll be really surprised and, by contrast, it's a wash between .NET and native apps most of the time.

Reply Score: 1

leos Member since:
2005-09-21

The .NET runtime is also compiled so, wrong, they ARE native code. It's no different than any other statically or linked native code dependency (ie. C Runtime, Win32, etc).


Educate yourself: http://en.wikipedia.org/wiki/Common_Language_Runtime

The key part is: "At compile-time, a .NET compiler converts such code into CIL code. At runtime, the CLR's just-in-time compiler (JIT compiler) converts the CIL code into code native to the operating system."

That is fundamentally different than running native code. If you don't understand the difference I can't help you. Even though you can compile .NET languages to native code (just like Java can), it's rarely done and still not the same as running native code that doesn't require a virtual machine.

You can't just make blanket statements like that.


That's true. And I really can't be bothered to find detailed comparisons for your (because no comparison would be perfect anyway). I just know from experience (and common sense/logic) that applications written on a virtual machine require more resources. It's pretty much unavoidable, since the virtual machine itself (which is a complex piece of software) requires some memory and processing power, no matter what you do.

Sure, if your native app doesn't do anything, it has a lower memory footprint, but just try doing common things that most native apps do -- like bringing up a common open/save dialog, interacting with common Windows controls -- and then see how quickly your app's footprint increases in size due to DLL dependencies.


Yes yes I've written plenty of apps, and I know how to check dependencies and measure memory consumption.
You are partially correct in that the larger your app is, the less of a difference the overhead of the VM will make. But for the apps I write as an independent consultant, it makes a big difference.

Reply Score: 4

google_ninja Member since:
2006-02-05

First off, there is nothing stopping you from doing a full system on .net, or any modern managed runtime. .Net will Just In Time (JIT) compile to native code on the first run (or at install time), in some cases giving you better performance then C++ due to cpu specific optimizations. You do get some overhead from stuff like garbage collection and security checks, but that is vastly outweighed by the complete elimination of whole classes of bugs and exploits.

.net has alot of really cool features Qt doesn't. LINQ is a single syntax to query virtually anything, from an RDBMS to an XML file to a collection to amazon. LINQ2SQL is a built in, really well done OR/M. We have had pretty much the best implementation of generics out there for awhile now (allowing you to define type as an argument). On other languages, they just do some compile time magic to do the boxing/unboxing for you. With .net, it is actually a part of the platform, so foreaching through a List<string> is actually more performant then using ye olde ArrayList. We just got lambdas, which is a really concise way of doing inline anonymous functions. On the VB.net side we got XML literals, which while is really weird at first glance, really makes it quite possibly the best language out there for XML work now. On the asp side, we just got a rails style MVC framework (for those who don't dig the MVP model), we've got a REST framework that you can plug in on top of the LINQ2SQL OR/M called astoria that should be coming out fairly soon. And to my knowledge, doing web services is way easier then any other platform out there.

We are also getting a really nice functional language with this release called F#. That also brings up the whole language agnostic aspect of the framework. You can combine any number of languages you like into a single solution, meaning that if you have heavy duty recursion needs, there is nothing simpler then just adding a library in F# to your C# application, with no interop headaches or overhead. It is really nice to be able to pull out the best tool for the job, and to be able to immediately plug into any code published on the net.

There is WPF, which is a nice separation of concern, generating the UI from a separate xml file. While not exactly a new idea, IMO XAML beats the competition in terms of features and flexibility, and MS has actually started releasing XAML tools specifically targeted at designers, which has not been done on other platforms. That is a huge deal for corporate development, instead of the designers generating a mockup and then the developers copying it, the designers can create the actual UI, and the developers just plug into it.

Not to mention an incredibly rich platform with far more scope then anything out there (except possibly java). On Qt you have desktop and mobile support. With .net, there isn't much that you can't do with it. One of my favorite parts is how transferable the knowledge is. There is very little difference from coding a ASP.net webform, to a winform, to an app for your cell phone. There are difference, but generally only in the front end, and even then for the most part the ideas are all the same.

The toolset is a joy to work with and has an absolutely insane amount of documentation and resources. the much maligned tight integration in the stack ends up giving you a hell of alot of power. To my knowledge, there isnt anything like the asp.net caching api on other server side languages, and with httpmodules you can plug into the IIS pipeline programmatically.

Last but not least, mono does count. If you don't believe me, give it a try. They are pretty much completely .net 2.0 compatible (including ASP.net). Not only that, but its performance is suprisingly good. They are already being used as a scripting engine in a whole bunch of commercial applications (Second Life is the latest one), and it is a really solid platform in its own right. The big two things for me is that there is no Visual Studio yet, and WPF is going to take years to implement. Other then those two issues (and they are big ones), the experience is very close to what you get on the microsoft implementation.

All that to say, Qt is probably the best C++ platform out there, but its still C++. The future is in managed code, and imo .net is the best managed solution on the market.

Reply Score: 16

gustl Member since:
2006-01-19

Question:

Would KDE in cross-Platform C# (=Mono) be easier to implement than in Qt and C++?

This is not trolling, I honestly have no Idea, as I never learned C#.

And by the way: Only answers from People who used BOTH languages for a time count in my book!

Reply Score: 2

google_ninja Member since:
2006-02-05

I have used .net, mono, and qt before, but I have never used the qt bindings for mono (http://cougarpc.net/qyoto/).

That being said, I would say it is probably easier using C++/Qt at this time simply due to the tools available. KDevelop and QtDesigner are full featured, mature apps. Monodevelop hasn't hit 1.0 yet. If you are a "real men code in vi" kind of guy though, I would give it a try. The Qt bindings are generated off of the Qt header files, so you use it in pretty much the same way. The big differences would be the managed aspects, different style of libraries, the loads of syntactical sugar in C#, and the whole managed aspect.

Reply Score: 2

hobgoblin Member since:
2005-07-06

microsoft have always been good with dev systems. hell, even their office pack can be considered a rad environment. i have seen some corporate offices that base their internal solutions on ms office with special scripted files.

its even how microsoft got started, by providing basic interpreters to the different microcomputers that was being sold.

the question is, is the future for microsoft in dev tools for different platforms, or is it in making new versions of ms windows, ms office, and active directory and friends?

Reply Score: 2

Moochman Member since:
2005-07-06

Java has most of the things you mention (language interoperability, managed-ness, little adjustment needed between enterprise, desktop and mobile, support for XML layouts (see <a href="http://www.devarticles.com/c/a/XML/UI-Design-with-Java-and-XML-Tool...), but with one big advantage: it actually delivers on the promise of being cross-platform. Not cross-platform as in "we'll create a version for our OS and let everyone else make a crippled copy of it", cross-platform as in... cross-platform. It seems to me that the whole point of running things via virtual machine doesn't really *do* a whole lot if it doesn't serve to make apps cross-platform. So as long as .NET remains purely MS-controlled (and as long as MS keeps limiting its support for platforms other than its own), don't expect to see much support for it outside of the Windows-only world.

Reply Score: 4

google_ninja Member since:
2006-02-05

I like Java, I was a Java guy for years before I was a .net guy. The problem with java is it is too "ivory tower" in its design paradigms. Sometimes fully implementing a pattern makes sense, at other time it just means more code you need to write for no real reason. I find in general, .net takes a far more pragmatic approach, which ends up in less code and a more plesent experience.

The other problem with java is that it has been stagnating for years now. .Net showing up has forced movement, but what has been done has really not been that well thought out (like their generics implementation).

Last issue is that java is truely cross platform for only trivial apps. As someone who has worked with it for years, there are plenty of platform specific bugs you have to deal with, and optimization is completely different depending on where you are deploying to. That is not minizing what it does have, but in reality for a big project, you are pretty much operating like the rest of the world and doing plenty of porting work.

Apart from the piss-poor UI toolkit, I would choose Java over C++ any day. But I am enjoying myself much more since I have started working on .net.

Reply Score: 2

rajj Member since:
2005-07-06

There isn't a single feature you mentioned that isn't available in some form on OSS platforms. The fact that bring up Mono sorta blows the 'Windows is the most interesting platform' argument out of the water does it not?

The only thing interesting about .NOT, as a technology, is that all the languages compile to a common byte-code. I have to wonder what the point of having multiple languages targeting the same VM is if they have to all match the underlying semantics anyway. It amounts really to nothing more than syntactical sugar. On the other hand, I wonder what all the GCC languages do -- oh yeah, C is the common "byte-code" there; perhaps not as sexy, but it works.

The only thing left is the library, and frankly, you can keep your 600# bloat to yourself. If I wanted that, I'd just write for Java (or Mono).

There hasn't been a single thing I haven't been able to get done --quickly-- using Python and C in years.

Reply Score: 3

leos Member since:
2005-09-21

First off, there is nothing stopping you from doing a full system on .net, or any modern managed runtime.


Technically true, but you'll still run into walls with performance. This may eventually solve itself with faster computers, but we're not at that point yet. Just imagine the resource requirements of Vista if it was written on top of .NET.

.Net will Just In Time (JIT) compile to native code on the first run (or at install time), in some cases giving you better performance then C++ due to cpu specific optimizations.


Yeah this got bandied about a lot when Java came out too, but it wasn't true then and it still isn't true now in the general case. Sure you can find a benchmark where managed code is faster, but overall you're taking a hit.

.net has alot of really cool features Qt doesn't. LINQ is a single syntax to query virtually anything, from an RDBMS to an XML file to a collection to amazon.


LINQ is cool, but it remains to be seen if it's actually useful for general apps. It's cool in demos so far, and not much else.

We have had pretty much the best implementation of generics out there for awhile now (allowing you to define type as an argument). On other languages, they just do some compile time magic to do the boxing/unboxing for you. With .net, it is actually a part of the platform, so foreaching through a List is actually more performant then using ye olde ArrayList. We just got lambdas, which is a really concise way of doing inline anonymous functions. On the VB.net side we got XML literals, which while is really weird at first glance, really makes it quite possibly the best language out there for XML work now. On the asp side, we just got a rails style MVC framework (for those who don't dig the MVP model), we've got a REST framework that you can plug in on top of the LINQ2SQL OR/M called astoria that should be coming out fairly soon. And to my knowledge, doing web services is way easier then any other platform out there.


For web services, yes. Obviously Qt is not useful for that (but Java is). The rest is more or less a wash. A few random language features don't actually make much difference to writing code. In the end the quality of the class library brings the most benefit.

We are also getting a really nice functional language with this release called F#. That also brings up the whole language agnostic aspect of the framework. You can combine any number of languages you like into a single solution, meaning that if you have heavy duty recursion needs, there is nothing simpler then just adding a library in F# to your C# application, with no interop headaches or overhead.


Yikes. Good luck maintaining that. Realistically, a company is going to standardize on one language for each project (and probably in the company). F# might be slightly better than C# for certain cases, but the extra maintenance and training headache nullifies all those advantages. Realistically all you want is a base language and scripting support within apps (like QtScript does).

There is WPF, which is a nice separation of concern, generating the UI from a separate xml file.


Once again, nice idea but totally unproven as of yet. We'll see how this turns out in a few years when real apps start using it.

Not to mention an incredibly rich platform with far more scope then anything out there (except possibly java). On Qt you have desktop and mobile support. With .net, there isn't much that you can't do with it.


True, you can do web services easily for example. But if your focus is desktop/mobile, it just isn't as good. So you're a jack of all trades, but not really the best tool for what I do.

All that to say, Qt is probably the best C++ platform out there, but its still C++. The future is in managed code, and imo .net is the best managed solution on the market.


Maybe. In a few years I might agree with you. Maybe in a few years hardware will have caught up to the point where managed code is really the way to go. But right now, performance is still a big deal. Vista is still mostly native code, and even that is too big and slow. Just imagine if the entire userspace was built on .NET/WPF.

The bottom line for me right now is that I can use Qt to write a desktop application that is completely cross platform across Windows, Mac, and Linux with a single codebase, and the resulting app will appear native on all platforms (with the odd exception), with great performance. I can't do that with .NET, so there's not really much point in any of its other interesting features because it fails the basic requirement.

Edited 2008-02-04 20:48 UTC

Reply Score: 5

google_ninja Member since:
2006-02-05

Yeah this got bandied about a lot when Java came out too, but it wasn't true then and it still isn't true now in the general case. Sure you can find a benchmark where managed code is faster, but overall you're taking a hit.


Alot has changed since java came out. In general, java has given managed languages a really bad name due to its horrible UI toolkit(s). While you are completely correct that you are taking a hit, that hit is going to reducing code complexity by eliminating memory management, and to improving security by sandboxing and the runtime checks. Considering that ten years ago, what you can buy for 800$ at best buy would have been considered a super computer, I don't consider that a bad tradeoff.

Technically true, but you'll still run into walls with performance. This may eventually solve itself with faster computers, but we're not at that point yet. Just imagine the resource requirements of Vista if it was written on top of .NET.


If we aren't at that point yet, we are damn close. It is definately acceptable for the vast majority of applications though.

LINQ is cool, but it remains to be seen if it's actually useful for general apps. It's cool in demos so far, and not much else.


I was actually just listing off cool 3.5 stuff I have been reading up on recently ;-) It is true that this stuff isn't proven yet, but the ideas behind them are quite solid. What I was trying to illustrate is that .net is moving forward, and incorporating alot of really modern ideas. C++ is and always will be C++.

For web services, yes. Obviously Qt is not useful for that (but Java is). The rest is more or less a wash. A few random language features don't actually make much difference to writing code. In the end the quality of the class library brings the most benefit.


It kinda shows I am a web guy, when I am pulling cool features out of the air, at least half are web specific. You are right about that, but that is something that would be a hell of alot harder to illustrate. It is alot easier to say .net does x, where x has no real equivalent in other languages.

True, you can do web services easily for example. But if your focus is desktop/mobile, it just isn't as good. So you're a jack of all trades, but not really the best tool for what I do.


I agree with you here to a point. I'm a web guy, so while it is cool to be able to hack up some small app for my phone, it isn't exactly something that is billable for me. But what is really cool is that even though I have zero background in windows desktop apps, it was the easiest thing in the world for me to write a rich client that consumed a web service. In Java, or anything else for that matter, that would have been something I would have had to outsource. In my spare time I'm reading up on WPF (mostly cause I find the whole idea cool), and those skills will be directly transferable to writing silverlight RIAs, as silverlight is a subset of WPF. I was a J2EE guy for years, and being able to do this stuff so easily is just the coolest thing.


The bottom line for me right now is that I can use Qt to write a desktop application that is completely cross platform across Windows, Mac, and Linux with a single codebase, and the resulting app will appear native on all platforms (with the odd exception), with great performance. I can't do that with .NET, so there's not really much point in any of its other interesting features because it fails the basic requirement.


And that's the rub. Honestly, if a requirement for me to do business was being cross platform, I would probably make the same choices. Mono gives you some options, but there is still porting work that needs to happen. But just because it has a deal breaker for you, doesn't mean that there is a ton of stuff that it can do that qt/c++ wouldn't really be that good at.

Reply Score: 2

leos Member since:
2005-09-21

Alot has changed since java came out. In general, java has given managed languages a really bad name due to its horrible UI toolkit(s).


Partially, but mostly the slowness did it for me. .NET languages are faster (but so is Java 1.6), but they're still not where I'd like them to be.

While you are completely correct that you are taking a hit, that hit is going to reducing code complexity by eliminating memory management, and to improving security by sandboxing and the runtime checks.


The thing is though, with Qt, I don't worry about that stuff either. In my current project, I have some 7000 odd lines of C++ code. Guess how many delete statements I have in there. 7. 7 delete statements in 7000 lines of code. So for all intents and purposes, I don't have to bother with memory management. (No I don't have memory leaks ;)
On the sandboxing issue you have a point (although all the usual suspects like Qt string classes are also checked at runtime)

Which kind of brings me to another point. Those 7000 lines of code may not seem like much, but that makes up a flow chart editor with different shapes and objects, automatic connectors, lines, infinite zoom, undo, multiple sheets, printing, exporting to PDF or various image formats, a help browser, automatic update checker, sqlite database for the files, etc, etc.. And that same code runs on Windows 98 to Vista, OS X, and just about any unix derivative you can think of.
I've just never used any other language/toolkit that can come close to that kind of features/code density.

Errmm. Sorry about that completely unrelated raving ;)

If we aren't at that point yet, we are damn close. It is definately acceptable for the vast majority of applications though.


That I would argue with. A lot of the applications I use aren't exactly fast. Microsoft Office runs at a decent speed, but I definitely wouldn't want to sacrifice any of that. Same goes for Visual Studio, Firefox, the file manager, autocad, photoshop, etc.. The programs I spend the most time with are either very complex and thus need all the speed they can get, or very small and light, which means I really don't want them taking any extra resources.

What I was trying to illustrate is that .net is moving forward, and incorporating alot of really modern ideas. C++ is and always will be C++.


True. .NET is advancing at a very high rate. And I'm exited to see what will come next. But C++ isn't as static as you might think. Have a look here for the new stuff coming in 2009. http://en.wikipedia.org/wiki/C%2B%2B0x

And that's the rub. Honestly, if a requirement for me to do business was being cross platform, I would probably make the same choices. Mono gives you some options, but there is still porting work that needs to happen. But just because it has a deal breaker for you, doesn't mean that there is a ton of stuff that it can do that qt/c++ wouldn't really be that good at.


I guess we have a very different set of needs, which means there are no wrong answers, everything is shades of gray, and the world of internet debates implodes in a cloud of common sense. ;)

Edited 2008-02-05 08:02 UTC

Reply Score: 4

djame Member since:
2005-07-08

>Cutting edge has become too risky and they cater more towards >maintaining backwards compatibility than towards the cutting edge.

This is maybe the biggest strenght of windows, come on !
Just try to run any games (except q3 and ut) made by Loki on any recent distribution. Or any commercial apps from Wordperfect to applixware which used to run just fine on any distro. What would you say to those who bought these programs and can't use them anymore or whithout incredible pains (such as installing another dynamic linker)...


As long as linux won't respect people investment, windows won't be a dying platform..



Djamé

Reply Score: 6

Shade Member since:
2005-07-07

Most of the old Loki games work fine here... SMAC, SimCity 3000, and might and magic anyway. If you're having problems try:

(Google-fu)

http://gentoo-wiki.com/HOWTO_Running_Old_Loki_Games

Yes, it's from Gentoo but it works on any distro... Most of the Loki games use a symbolic link to the executable, so it's quite easy to add the instructions on the wiki to a script, remove the symbolic link, and your done.

So if the game doesn't work it's a google search for "old Loki games", 4 lines of text, the ability to delete a file, and the ability to save a file. Failing that you could go play with a pile of rocks. :p

(And before you say: "Ohh that's too much work for Joe User!" this is the same Joe User that'll download a DOS emulator on Windows to play vintage DOS games so the point is moot. You can't handle an emulator if you can't handle Google and a text file.)

Reply Score: 5

djame Member since:
2005-07-08

Yes, it's from Gentoo but it works on any distro... Most of the Loki games use a symbolic link to the executable, so it's quite easy to add the instructions on the wiki to a script, remove the symbolic link, and your done.

Come on ;) you made my day, I just wrote before that having to tweak the dynamic linker was not considered as straithforward and guess what I found on the link you provided ?
The solution is to use an old version of glibc such as 2.2.5. However all the other libraries used by the game must also be compiled against an equally old (or older) version of glibc. (..)The game is then run by prepending the game binary to run the game with the old ld-linux.so.2 dynamic linker/loader

So 1) could you tell me what you taught me ? 2) Could you tell me how easy it is compared to just run an installer ?

Last time, I was upset about the reactivity of my mail reader (thunderbird) so i wanted to install an old version of Netscape communicator just to check if my memory was right about its speed and even with all the tricks I know (such as making work genuine version of Houdini and Wordperfect on mandriva 2006) I couldn't manage to get it work and I think that not being able to install apps I 've been working with for ages sucks like hell.

Cheers,
Djamé

Reply Score: 5

Shade Member since:
2005-07-07

Oh come on, you've cited your own examples there-- The big closed legacy apps are: Netscape Communicator, Wordperfect 8 and 9, the Loki games, and the Transgaming release of the Sims. OK-- That's 2001 (and earlier). Tell you daughter what you like, but odds are she wasn't born or eating solid food when this stuff came out-- and there is a theororetical (and somwhat contrived) probability you might need Netscape, but come on.

Compare that what to how many hours you spend hunting software on a virgin Windows install. Let's see-- first of all you want to keep that virgin Windows box of the net. Then you probably need to find: A decent firewall, virus scanner, spyware remover, Web browser, and your fav plugins and codecs. Then you remove the crap the vendor installed. Then you connect the machine to the web and have to do microsoft and your vendor updates. Then you have to the hardware vendor's sites and and grab the even more recent than your system vendor's drivers. Then you get to use your computer, and still put up with all manner of new bugs in legacy apps. Then, unless your family and friends stay on top of all of this stuff you get to repeat the process in 6 months. (Don't I know it...)

I'll take the pain of google and a 4 line script to run 7 year old software any day... I fix enough Windows boxes for family and friends as it is... (Which is sad because I haven't worked on Windows boxes in years, and it left my desktop in 98 or so.)

Reply Score: 2

djame Member since:
2005-07-08

Don't mix everything, we're not talking about security fix or whatever, we're talking about binary compability.
And no, the apps you're talking about are not the only one.
How about the version of Allegro Common lisp 6.2 which my Lab bought and can only be run on glibc2.1 (50 000 euros, the site licence), or sictus prolog, or Maya 6 (I don't remember the version) or tons of other apps ?
I know that oss version of similar tools exist but if you want a really good lisp compiler, you've got no choice.
The point is not about security or 4 lines of scripts, the point is about being able to use the software I bought longer than 2 years.

I used to be a strong supporter of OSS, and still I am, but seriously, changing API, ABI or drivers interface is not the way to ease wider adoption on linux, whatever desktop environment you like.
It's directly related to the time you have to devote to maintain the tools you want to use in order to do your job. And being older, having a job, a life, you just have less time for this kind of lame stuff such as
"Oh, sh.t I forgot to install ld compat, libstd++5.co and ld.5 and so I'll have to spend I don't know how many time to fix that.
You do that once, maybe twice, but the third time, you just say no more.

Reply Score: 1

Soulbender Member since:
2005-08-18

Just try to run any games (except q3 and ut) made by Loki on any recent distribution


Try running a Windows 9x game on XP without being an administrator.

Reply Score: 4

dagw Member since:
2005-07-06

At least you can easily get them to run, and since just about every windows user in the world as Admin rights on their gaming box it's hardly like it's a huge problem.

Edited 2008-02-04 15:33 UTC

Reply Score: 4

Soulbender Member since:
2005-08-18

At least you can easily get them to run


Really. Try the first Turok game or Resident Evil.

Reply Score: 2

glarepate Member since:
2006-01-04

As long as linux won't respect people investment ...

Pardon my descent into grammar-nazi-hood but:

Linux is *based* on people investment much more so than, but not exclusive of, capital investment.

As long as you value [old, but classic] games over the bulk of available productive, stable, but not easily backwards compatible functionality you should be running Donkey Kong or some other Mario-based system on WII, no?

Reply Score: 3

djame Member since:
2005-07-08

Actually this is a bad example if you look at all the support consoles producers brougth to their old consoles
PS 2 : compatible with ps 1
PS 3 : compatible with ps 2 (and ps1 ? i don't know)

I may remember that there was some sort of compatibilty between supernes and n64 if you bought the proper cartridge adapter and actually
the wii can play old game through emulation (it's not properly talking binary compatibility but, the same can be said for mac os X ppc which was able to launch classic application and for mac os classic which was able to run 68k code (and of course, see Rosetta for mactel machine).
What does it mean ? it means that an os provider actually care about software (games or others) bought by users and doesn't threw away the baby with the bath's water every time he makes an update to the os (or a major change for that matter).
But this is an old debate so...

Reply Score: 1

vimh Member since:
2006-02-04

I'd argue more that the desktop is a dying 'platform.' No seriously. You suggest that MS has done nothing to bring anything new to the desktop. I won't disagree. But I suggest that neither has the Macintosh world or Linux.

We have been sitting in the same place for years. Choose your poison.

Reply Score: 2

trenchsol Member since:
2006-12-07

Some people have serious perception problems, or keep living in imaginary world they made for themselves, which is, by definition, unhealthy mental condition. Microsoft Windows have been installed on more than 90% PC's in the world during last ten years. That is the fact, hard, unavoidable fact. And, in those 10 years I have seen dozens of articles describing the end of Microsoft. None of them offered any sound argument for that except authors own wishful thinking.

Windows will go away someday, there is no doubt of it, but some substantial change is going to take them away. Such change has not happened yet, so please, keep your daydreaming for yourselves until then.

In the 90's Windows offered a compromise of user friendly and affordable system at expense of quality, stability and usability. Since then the way of using PC is the Windows way. And no platform is better in being Windows than Windows itself. It has nothing to do with quality, usability, security or anything else. It is a market leader position.

Reply Score: 3

elsewhere Member since:
2005-07-13

Come on.... It's shipped on 90 if not 95% computer sold each year.


If you're not growing, you're dying. ;)

It's tongue-in-cheek, anyways.

Reply Score: 4

MrCopilot Member since:
2008-02-04

Come on.... It's shipped on 90 if not 95% computer sold each year.
Anyway, except from the famous joel spolky's "how microsoft lost the api war", I have yet to see any good arguments about microsoft fall

Response:
http://mrcopilot.blogspot.com/2008/02/windows-platform-death-revisi...

Reply Score: 2

google_ninja Member since:
2006-02-05

I don't get it. Windows is dying because some guys daughter like KDE?

Reply Score: 3

RandomGuy Member since:
2006-07-30

Yes, we KDE users have some amazing mental powers. ;-)

Reply Score: 4

StephenBeDoper Member since:
2005-07-06

To summarize the response:

"When I wrote that Windows is 'dying,' I actually meant something completely different than either the literal or commonly-accepted meaning of the word 'dying.'"

Reply Score: 2

Thoughts on KDE for Windows ...
by WorknMan on Mon 4th Feb 2008 00:46 UTC
WorknMan
Member since:
2005-11-13

Hope they get sound working... I wanna piece of that Simon clone ;)

Other than that, I don't see much that interests me, cept for k3b (which wasn't even mentioned) and maybe Amarok. As for Konquerer (the file manager part) and Dolphin, check out Directory Opus on Windows. Mops the floor with both of those apps.

Oh, and taking a look at the second screenshot (the one with the packages) reminds me of one of the reasons why I don't use Linux. What a huge clusterf**k.

Reply Score: 5

Verunks Member since:
2007-04-02

Hope they get sound working... I wanna piece of that Simon clone ;)

in the latest unstable version of kdewin( http://download.cegit.de/kde-windows/releases/unstable/kde-4.00.80 ) you can install phonon and amarok2 although it doesn't work perfectly but maybe it's the same on linux i haven't tried yet

Reply Score: 1

RE: Thoughts on KDE for Windows ...
by miscz on Mon 4th Feb 2008 01:16 UTC in reply to "Thoughts on KDE for Windows ..."
miscz Member since:
2005-07-17

Oh, and taking a look at the second screenshot (the one with the packages) reminds me of one of the reasons why I don't use Linux. What a huge clusterf**k.

No, that's the reason not to use Windows. Cygwin installer is a "huge clusterf**k", I wouldn't compare it to Linux package managers.

Reply Score: 10

WorknMan Member since:
2005-11-13

I wouldn't compare it to Linux package managers.

Why not? Doesn't look much different:

http://www.nongnu.org/synaptic/images/0.53-upgrades-only.png

Reply Score: 5

miscz Member since:
2005-07-17

Cygwin installer is way overcomplicated for stuff it does. Synaptic is much more than just installer. It's comparing apples to oranges because on regular Linux desktop you'd have base system installed by default. Also, Synaptic has search functionality unlike Cygwin so you don't have to actually scroll through hundreds of arcane-sounding packages. Not to mention more user-friendly alternatives like the Gnome application installer.

http://img131.imageshack.us/img131/6027/zrzutekranudodajusuprogka9....

Reply Score: 7

RE: Thoughts on KDE for Windows ...
by klimg on Mon 4th Feb 2008 10:49 UTC in reply to "Thoughts on KDE for Windows ..."
klimg Member since:
2007-08-03

For 80 bucks directory opus better does mop the floor - don't care what it does it with though.....

Reply Score: 3

v diening windows?
by netdur on Mon 4th Feb 2008 01:20 UTC
RE: diening windows?
by NotInterested on Mon 4th Feb 2008 02:11 UTC in reply to "diening windows?"
NotInterested Member since:
2008-01-02

it feels more like KDE is dieing...

Either that or your spelling, or my eyes, or... or...

Reply Score: 3

v RE: diening windows?
by marcusesq on Mon 4th Feb 2008 05:54 UTC in reply to "diening windows?"
RE[2]: diening windows?
by netdur on Mon 4th Feb 2008 09:18 UTC in reply to "RE: diening windows?"
netdur Member since:
2005-07-07

or I suck trolling!? am just newbie, don't be harsh

Reply Score: 1

Shell Replacement
by computrius on Mon 4th Feb 2008 02:12 UTC
computrius
Member since:
2006-03-26

This doesnt seem to be so much a port of kde to windows, as much as it is a port of a bunch of QT apps to windows. In which case, what exactly is new here? This has been done for a long long time before kde 4 was even planned..

Id be more interested to see a replacement shell for explorer. A real port of kde to windows.

There used to be tons of projects like this (litestep, etc). Are any of them still active? or better yet, working with vista?

Reply Score: 2

RE: Shell Replacement
by KugelKurt on Mon 4th Feb 2008 10:58 UTC in reply to "Shell Replacement"
KugelKurt Member since:
2005-07-06

This doesnt seem to be so much a port of kde to windows, as much as it is a port of a bunch of QT apps to windows. In which case, what exactly is new here? This has been done for a long long time before kde 4 was even planned..

The following is new:
Past KDE releases depended lot on X11 and POSIX. Previous porting efforts to Windows and Mac OS X resulted in quite big unofficial patches. KDE 4 however uses Qt and other abstraction layers to a larger degree.
This is the result of officially embracing non-X11 platforms.

Reply Score: 6

RE: Shell Replacement
by anda_skoa on Mon 4th Feb 2008 13:15 UTC in reply to "Shell Replacement"
anda_skoa Member since:
2005-07-07

This doesnt seem to be so much a port of kde to windows, as much as it is a port of a bunch of QT apps to windows.


There is a big difference between a port of "a bunch of Qt apps" and a port of "a bunch of KDE apps".

The applications mentioned in this blog are using the KDE application framework, which means that this framework is now also available for developers of multiplatform applications.

Obviously, since KDE itself also uses Qt, a Qt application could deliver all features a KDE application can, but it would require huge efforts since the application developers would basically have to implement it all by themselves and even then it would be totally new, untested code.

Reply Score: 4

RE: Shell Replacement
by polaris20 on Mon 4th Feb 2008 14:50 UTC in reply to "Shell Replacement"
polaris20 Member since:
2005-07-06

Agreed. It's not going to be that interesting (IMO) until KDE replaces the actual shell completely, like Aston Shell does (http://www.astonshell.com/).

Otherwise, it's a collection of ported Qt apps, some working, some not.

Reply Score: 3

RE[2]: Shell Replacement
by Soulbender on Mon 4th Feb 2008 15:23 UTC in reply to "RE: Shell Replacement"
Soulbender Member since:
2005-08-18

like Aston Shell does (http://www.astonshell.com/).


Wow. Yeah, I for one can hardly wait for KDE to become that lame....

Reply Score: 3

RE[3]: Shell Replacement
by polaris20 on Mon 4th Feb 2008 17:49 UTC in reply to "RE[2]: Shell Replacement"
polaris20 Member since:
2005-07-06

You can customize Aston's looks however you want, and it runs faster than the default shell. This is bad because.....?

Whereas the KDE implementation for Windows is a bunch of Qt apps. Yay.

Reply Score: 2

RE[4]: Shell Replacement
by superstoned on Tue 5th Feb 2008 14:31 UTC in reply to "RE[3]: Shell Replacement"
superstoned Member since:
2005-07-07

KDE doesn't intend to replace the windows shell - ever. If you want the full KDE experience, use Linux/BSD/Solaris/whateverUnix. Just the apps (at least most of them) will become available for Windows. We'll try to ensure they are stable and usable, but our focus is and will be FREE SOFTWARE platforms.

Trying to fix Windows is useless, MS will just screw you some other way. Give up and use a proper, open platform.

Reply Score: 3

RE[2]: Shell Replacement
by Morty on Mon 4th Feb 2008 18:49 UTC in reply to "RE: Shell Replacement"
Morty Member since:
2005-07-06

Agreed. It's not going to be that interesting (IMO) until KDE replaces the actual shell completely,


A desktop shell is rather boring and uninteresting, and not very important when it comes to preforming your tasks eg. work.

It's in the applications you spend your time. A collection of high quality best of breed applications KDE (and the use of kdelibs) provides, are very interesting.

Reply Score: 4

RE[3]: Shell Replacement
by polaris20 on Mon 4th Feb 2008 23:20 UTC in reply to "RE[2]: Shell Replacement"
polaris20 Member since:
2005-07-06

Huh? the speed and stability of the shell doesn't have an impact on work?

Reply Score: 2

How times change...
by grat on Mon 4th Feb 2008 03:01 UTC
grat
Member since:
2006-02-02

I thought *BSD was dying?

:)

Seriously-- Windows isn't even ill. The fact that other OS's are getting more attention is the sign of a healthy market.

Reply Score: 10

tomcat
Member since:
2006-01-06

... when he prefaces his remarks with indefensible bombast about a "dying Windows platform". Oh, please.

Reply Score: 4

KugelKurt Member since:
2005-07-06

Yeah, humor is officially forbidden in the blogosphere...

Reply Score: 3

StephenBeDoper Member since:
2005-07-06

That would explain why the blogger opted for juvenile flamebait instead.

Reply Score: 3

Dying Windows Platform
by _mikk on Mon 4th Feb 2008 03:19 UTC
_mikk
Member since:
2005-10-19

Wow!!!

I've never read such a bullshit headline.
Amazing!

:-D

Reply Score: 12

v RE: I've never read such...
by glarepate on Mon 4th Feb 2008 08:16 UTC in reply to "Dying Windows Platform"
Windows dying ?
by raver31 on Mon 4th Feb 2008 06:51 UTC
raver31
Member since:
2005-07-06

It is in my house, it also is in my friends, family and work colleagues houses, but this is a drop in the ocean.
Windows will be around for a long long time to come, and I think this is another flamebait article put on here to drive up traffic, and thus revenue.

Next article I see about Windows dying, Linux dying, OS X dying, BSD dying will get completely ignored. The comments will also get ignored as there will be fanboys from all platforms arguing.

Please stop posting articles of this nature.

Reply Score: 3

RE: Windows dying ?
by tomcat on Mon 4th Feb 2008 07:16 UTC in reply to "Windows dying ? "
tomcat Member since:
2006-01-06

Actually, it's really just the title that I object to. I have no problem with the assertion that KDE has something interesting to offer on Windows (or Linux).

Reply Score: 4

RE: Windows dying ?
by KugelKurt on Mon 4th Feb 2008 11:04 UTC in reply to "Windows dying ? "
KugelKurt Member since:
2005-07-06

In this context I think it's meant more humorous than flamebaiting. The article is actually interesting for whoever is interested in KDE on Windows.

Reply Score: 3

So what?
by Buck on Mon 4th Feb 2008 07:28 UTC
Buck
Member since:
2005-06-29

Well, who'd want to save Windows anyway. Let it die.

Reply Score: 0

autumnlover
Member since:
2007-04-12

Windows is "dying" ? Where ? What ? When you will be a candidate to a parliament or for a presidency and your party/yourself will gain a 0,5% of votes, and your main competitor will reach 95% of votes, will you still be telling people that you win the elections ? get real!

ps. fanboys please don't tell me about servers, because KDE is certainly not a server-environment.

Reply Score: 2

superstoned Member since:
2005-07-07

Just read the article. The title was not meant to be funny.

Reply Score: 2

Oh, joy...a new version of Windows!
by mlauzon on Mon 4th Feb 2008 18:03 UTC
mlauzon
Member since:
2005-07-25

So, I see the new version of Windows is named:

Windows Linux

Wonder what it's going to be like.... ;)

Reply Score: 1

Windows is not dying...
by kajaman on Mon 4th Feb 2008 21:02 UTC
kajaman
Member since:
2006-01-06

I just don't understand why some people like spawning stupid flame wars with such titles. Anyway, I like to comment ;) .

Windows is not dying, it's growing and getting more mature. I'm saying this being GNU/Linux user myself - because of many reasons - but first of all usability (second, philosophy).

But I see how significant changes have been made to Windows, and especially server side, and programmer's development environment like .net/VisualStudio. I stopped treating this platform like unusable piece of crap, because it's not any more. You might disagree if it's "done right", but lots of programmers like it.

But - here we go - Windows is still lagging behind on desktop usability (sic!). When I compare Windows Vista, which I use along with Linux for some work, to my KDE (3.5) or shiny Fedora's/Ubuntu's Gnome or OSX I feel like these are two different worlds. Windows still not easy to use, it's full of quirks and it's not very logical (tried to change MAC address of my ethernet card - what a hell to find it in settings). It's great mystery of our times why such unusable system is used on ~90% desktops across world. In this sense - some KDE apps would make great difference - but again - they are inconsistent with rest of clumsy windows GUI so we have more mess on our desktops. Better switch to real alternatives than install KDE on windows IMHO, but good to have it there.

Reply Score: 1