apt-get, kpackage, gnome-apt, the new Debian installer that is soon to come out, that rpm-getter tool
I'd say getting and installing packages is far easier in many Linux systems, even for a less-knowledgable user.
Actually, many KDE programmers are from Trolltech, and support Opensource idealogy.
Trolltech are open out of:
A) some idealogy
B) it pushes Qt, which strngthens their grip in the commercial world, and allows them to 'show off' samples of running code, such as KDE.
Closing Qt in light of the two is unlikely, as most of their profit is from their closed-source buyers, who pay a Single-fee for Qt, with yearly upgrade fees. No need to pay per-sale
(Knowing this, as I worked for a Qt-using propriarity software company).
Unless they whole business model, they would have to sell Qt to the developers, because buyers don't pay for Qt.
Ironically, many of those developers work in Trolltech, and Trolltech know would never pay.
They also know KDE has no chance of succeeding in a closed pay-per-copy license, in the opensource world.
I use Konqueror every day, but there are just a few things I feel are missing. The articles didn't mention these, though. heh, same here.
2. It would be nice if I could put my favorite links on the menu bar, like with Navigator. Settings->Show bookmark toolbar
if that's what you mean.
3. The bookmarks menu demands that I hold down the mouse button while swishing through my bookmark folders. If I accidentally let go, I end up with the wrong site, or all too commonly, get the "edit bookmarks" page. This behaviour depends on the Style/theme you've chosen (look&feel styles). In the default one, you can simply click (and let go immediately) on the bookmarks menu, and wander without having hold the button.
4. There is noooo rule four. I don't get it:)
5. Konqueror still croaks on various web sites. I don't know if it is the complexity, or maybe something to do with managing the color palettes. (My xterms are fixed - graphics upgrades are impossible...) It happens very rarely that web pages cause trouble. Once in a few days of intensive browsing it crashes, but I never had it "croak up".
6. They did mention the loading time, but I'll still mention that it is slow. Sure, maybe my P150 was not up to snuff, but an AMD 800 with 256 MB of DDR? It seems to load fast here (a matter of 1 second or 2 at most), but maybe its because I'm running KDE.
What the fuck have you been smoking? This is not acceptable. Free software should be free to everyone, for any use. Those Trolltech fucks keep weaseling around, trying to avoid admitting they're still a proprietary software company, using open source only for advertisement. Fuck them, fuck KDE. Heh, you could have had a legitimate point back in QPL days, but now they use the GPL, so you're a complete troll.
Why do we need another web browser? We don't, we have Konqueror:P
Do you have some problem with Mozilla that we should know about? Besides for taking 25 seconds to load for the first time (in a while) on my Athlon 750 (for comparison, Konqueror takes at most 4 seconds on the first load)?
Oh yes, it also seg faults when I try to load it now:
ii mozilla-browse 0.9.3+0-3 Mozilla Web Browser - core and browser
Yes, it also doesn't display Hebrew pages as nicely as Konqueror, which automagically selects the proper encodings and fonts.
Its also not using the very very powerful KDE IO slaves and KPart architectures, which allows centrally writing code to support protocols, format interpretation, etc.
I'm sure some asshole will moderate this post troll or flamebait, but I'm 100% in earnest here. You sure do deserve a 'troll' moderation or at least 'misinformed' (theoretic moderation) for reasons I will outline below.
Did you ever wonder why Mozilla is continually falling behind schedule? Because people like the Konqueror team decide to go off on their own instead of working for the good of the community. Mozilla was there first, and it deserves the support of the community. Konqueror is *NOT* a web browser. It can be *used* as a web browser, but it is much more. It merely encapsulates the KParts/IOSlaves of KDE, with some nice toolbars, menus, and cute features on the side. As an example, Konqueror can use Gecko for rendering instead of KHTML.
KHTML is the actual rendering engine. KHTML was there BEFORE mozilla. KHTML was there before Netscape opened its code.
Mozilla people decided to REWRITE the rendering engine, rather than using the already-existing KHTML code!
They should have joined the KHTML/Konqueror project.
If the free software community wants to make a good impression on the business world (and it may already be too late), we must, at all costs, avoid splitting into tiny, useless factions working on useless, duplicate projects. Konqueror is very useful. Mozilla is very useful.
Who cares about the impression of the business world?
Here's an idea: before starting your new project, check to see if someone is already working on a similar project. Had the Konqueror team observed this little suggestion, the whole Konqueror fiasco could have been avoided. You mean the Mozilla fiasco could have been avoided, as KHTML was there first:P
Maybe you should check on your facts, or read about the behind-the-scenes architecture of Konqueror and KDE, and that of Mozilla, and compare which is more powerful and flexible, and allows for more code-reuse/etc.
Konq being a browser simply results from its use of the KPart for HTML viewing.
Just as it views an image or a PDF file, or a system directory, using the proper viewing KPart.
Its ability to get web pages via the HTTP protocol is a simple result of the HTTP IO slave which is part of the powerful KDE architecture. Konqueror merely encapsulates KDE technologies, which are so powerful, that a subset of them form a web browser.
Thats the problem with the KDE developers. Since day one they closed they're ears and eyes to reasonable answers. The fact they started with a proprietory/not free toolkit underscores the prevelant 'we dont give a shit about you' attitude the KDE project has.
Qt was always open.
The only difference between the QPL and the GPL is that QPL has some section meant to prevent the forking of Qt into a free project that will compete with Trolltech.
Its a very reasonable license in its opensource terms.
They could never accept a browser like Mozilla for one reason: They believe they are better.
Mozilla has matured into a stable and fast web browser with an unbelievable amount of power. But the KDE project just blows it off and keeps on believing their browser is a superior product. Again, as someone else said, Konqueror is *NOT* a browser. It is a framework, encapsulating the KParts.
The actual browser is KHTML.
KHTML was usable before Mozilla was, so the question should be reversed, and truly, I believe that Mozilla people *should join the development of Konqueror and KHTML*, which *IS* a better product.
Sorry to break this to you , but Konquerer is a cheap imitation of the real thing. Immitation or not, Konqueror is *usable as* a fast usable web browser, with powerful and quick viewing of many file formats from anywhere (of which HTML is simply one of), including the local file system, FTP, Samba, etc.
(am I biased ? I dont know. I use KDE for my desktop.. but use mozilla/evolution daily.)
When was the last time you *seriously* tried using Konqueror instead?
That's what the non-us Debian server is for:)
Seriously, there's a workaround for every problem.
With new algorithm-generator software, that spits out many many patented algorithms, we may see a change in the field of software patenting.
Also, Debian could claim prior art to most patented allegations, and/or opensource authors could put the projects under people who are not restricted by the patent, etc. etc.
Are you kidding? Object foo knowing how to bar itself is quintessential OO. Then you can have object foo2 knowing how to bar itself, or if they're derived from the same superclass & bar can be abstracted a level, move bar up to the superclass. In smalltalk to print an integer, you'd do:
(5 + 4) printNl!
5+4 evaluates to an Integer object 9, which then prints itself. If object foo knows how to Stringize itself, then foo2 also does!
print (5+4)
and its not that 'print' itself knows how to handle various objects, it merely calls the object's __str__ method, which is clearly polymorphic.
Making a (to the user) generic print function to which you pass a string/integer/character may be passing around objects internally, but if you do that (as in the snipped of python you post) the language might as well be procedural from the user's point of view - the benefits of OO were used in language design, but are lost during the actual language use. The only 'advantage lost' is the syntatic sugar of using:
OBJECT-IDENTIFIER . FUNCTION-IDENTIFIER (Extra Parameters)
rather than:
FUNCTION-IDENTIFIER ( OBJECT-IDENTFIER, Extra Parameters)
This is an issue in Python (And Smalltalk).
The reason I said its not very OO'ish for an object to know how to print itself, because an object should not implement EVERY specific feature anyone might have thought of.
Do you think every object should implement a 'describe vocally' method, so that you could: '(5+4) say' too?
What about '(5+4) encrypt' or '(5+4) encodeMp3Sample'?
If you simply remove the difference between the notaitons:
f(a,b)
and:
a.f(b)
as it is in LISP, you get the consistency and the power.
However, it makes it a bit harder to find the list of available methods (which can be resolved easily).
Anyhow, Python allows you to:
object.print = my_print
so it can print itself:P
But the generic print is more OO'ish, because an object should have a minimal set of things it can do, one of them is represent its state as a string. In Python, another is returning a state object describing the state uniquely, in a way that can be reconstructed.
its dumb, because a print function would simply be better.
The fact its quite hard to mix strings with variable names and the need to convert between types shows you - not that you need a specific 'print' that helps you convert everything to a specific type - string, but that you need a general method of converting everything to a specific type (in this case string).
And this method is already available:
def print(*vars):
for var in vars:
sys.stdout.write(str(var) + " ")
sys.stdout.write("\n")
Or something of this sort.
Is: print ("Var:", var) so much worse? I think its more powerful, allowing things such as nesting prints in lambda's, etc.
I think that it would be cooler to steal the one cool Perl feature here:
print ("Var: $var") and allow an escape code such as '$' to address the non-literal namespace of variables, and convert whatever to string via __str__.
- Smalltalk's method call syntax
- LISP's macros
- The language that first introduced lambda's.
- The language that first introduced closures.
- Simula's code structuring
- Python's array slicing syntax (not sure about this)
- Python's indentation-is-syntatic
- etc. etc. etc.
Don't let Microsoft convince you innovating is collecting pieces from all around and sticking them together in a single bunch.
And just for the record - I hate Java, note it hasn't invented anything either:)
I agree that the built-in print is probably a dumb addition to Python. But dumb additions are harmless - whereas the lack of useful additions is harmful.
Trying to be pure usually results in the latter.
Built-in types do not stink, as they allow for better performance. You could always use the class-wrappers of the built-in types in Python - so NO functionality was ever missing.
is this truly OO?:)
and in Python, its: print 'Hello World'
If you want to describe how your objects 'print themselves', in Python its better done in a more generic way, you write an __str__ method in your object, that knows to represent it as a string for printing, but as an added bonus - it can be represented as a string for ANY purpose!:)
Python has many advantages over Smalltalk or LISP.
Its code is highly readable even by newbies, and it has many syntatic sugars that make the code elegant and clean.
Python saves every bit of redundancy:
- Variables aren't declared beyond their assignment, whereas in Smalltalk, they are.
- There are no explicit indentation-markers, indentation itself is used. This means there is no redundant and error-prone representation of the code which may lose sync with the real representation of the code hierarchy, whereas in Smalltalk (and LISP, btw) this redundancy is there.
And there are various other points where Python saves you, and increases the elegance of the code - at the expense of making the language perhaps less 'pure'.
Python also has MUCH more practical code available - graphics, 3d, gaming engines, COM object interaction, etc.
LISP is a whole other story - its probably more powerful - but its mostly unreadable by code illiterate - and I'm not sure about the practical library availability either.
Microsoft was out there INNOVATING Name a single feature in C# that hasn't appeared in at least 3 languages prior to it.
Anyhow, I'd take Python, C++ or LISP before I'd take C#, for any task. (The first for elegance and ease, the second for performance, and the third for its power).
I agree Ruby and Perl seem to be like the latter type.
Python, however, is a truly clean, well-designed, quite-pure, strong-typed language.
I wouldn't say it has a Basic odor at all, besides for a built-in 'print', perhaps:)
Basic is braindead and has no library - instead, it has its entire library built into its syntax in a horrid way.
Python is the exact opposite - many many libraries separated into modular modules, with only flow constructs, OO organization, functions and exception handling built in.
I think you mean "Apparently very few reading skills". But then what do I know? All I can do is spell and use proper grammar. You probably speak better English than I -- not that it means much, English being a second language I don't speak often.
That's not what you said. Maybe that's what you meant to say, but it's not what you actually said. Sounds like you're the one with the reading problem. Try reading your own messages before hitting the "Submit" button. I just re-read my original message - and its blatantly trivial I was bitching about the judging of Linux according to a specific configuration. I said:
"It seems you have a common disease - assuming Linux has One look, One way, One policy - as you have in Windows.
No, there is no 'general' way distriubtions handle errors and no 'default' graphical mode that distributions always select. "
Aka: Linux is not an OS, its a kernel, and GNU/Linux is a minimal OS that serves as the core OS of distributions.
It does not contain the management code - therefore you cannot bitch about its management sucking - you'd have to bitch about a distribution's management - but noone seemed to do that.
Really, is that so? Can you back up that statement with facts? I didn't think so. Anyone who thinks that a GNU/Linux system has to be as insanely complex and disorganized as it is just to provide the functionality it can offer is simply too lazy to learn anything about usability or human-machine interfaces. There are numerous specific aspects of any given GNU/Linux distribution that simply wreak from an unnecessary lack of usability, consistency, and simplicity. Want some examples? Just look at the recent corporate analysis of GNOME's usability that was posted on Slashdot and you'll see plenty. You obviously have some basic misconception about the terminology of this discussion.
Linux is a kernel.
GNU/Linux is a core OS.
A distribution is a whole OS, bundled with all the software you may need.
When saying 'Linux is complex' - you are claiming that the kernel is complex. Or even 'GNU/Linux is complex' is claiming that the kernel, plus the set of simple command line utilities are complex.
These systems are very simple. The kernel is very clean and simple, and the tools are clean and simple.
What you mean to say - is that the Linux user interfaces are 'insanely complex'.
I have little experience with Redhat or Mandrake, but from that little experience - they seemed to have auto-configured hardware/software/etc. They also set up a trivial-to-use KDE desktop, where all I need to do is point&click, and all my hardware is supported.
I don't see where you claim the UI has any insane complexity, but perhaps you're referring to distributions such as Debian and Slackware - where the user is *expected* to know how to handle the lower-level interfaces.
If you try to argue that Gnome is not very usable - I'd agree. I don't use gnome, although many others find it usable.
I use KDE, my parents use KDE, and they are VERY computer illeterate.
Really? That's amazing. Then apparently Windows 2000 doesn't actually exist, and is a violation of some previously undiscovered universal law of reality. It offers the same overall functionality in a much more usable package--and that is they key reason people use it over any UNIX-like system. Win2K is much much more complex than Linux -- using the correct terminology ofcourse.
If you look under the hood - you'll find a MUCH cleaner system under Linux, than under the Windows hood.
When comparing the user interfaces - KDE is more configurable, better looking (to most viewers), and much more quickly evolving.
Why do you incorrectly assume that I can't be speaking about more than either of those? While both the filesystem and "package layout" are poorly organized, GNU/Linux systems suffer from numerous other disorganizations. For instance, many applications hard-code filesystem paths into their source, so you actually have to recompile an application from source code if you simply want to install it to a directory of your choice. Another example? There is no distinction between a user's environment variables and system environment variables as there is in NT, so if a user unwittingly blows away a necessary system environment variable with their.bashrc file or whatnot, suddenly certain applications don't work and they have no idea why. More examples? Every application stores its configuration files in a different file format, and often times you have to learn a new scripting or programming language just to accomplish something as simple as getting X-Windows to work right with your video adaptor. Actually, automatic X configuration with Redhat/Mandrake is quite easy, and from my exprience - they autodetect the hardware.
If not - tough luck - you have to work through some admittedly more complex UI's to configure them.
Appearnatly you're missing that basic configuration is a one-time task that, like the hardware installation - can be done by a techie who knows what he's doing.
Using the machine is where UI usability should be compared.
Anyhow - why the hell would you be learn a programming language when configuring X?
About the 'standard configuration format', called the registry, that must be a hilarrious joke - because locating registry keys and interpreting their representations is at least as difficult than editting the dot text files in the home directory - not that you have to, given that applications provide a UI to edit their configurations.
I've never encountered any difficulties with my option to override system variable values.
Hardcoded paths are not a major problem - because the system is so well organized, that unlike Windows - files *DO NOT MOVE AROUND*. You are *NOT SUPPOSED* to install it in a 'directory of your choice'. Organization *leaves you no choice*, because choice, not based on policies is simply *unorganized*.
There is only ONE right location for a file under any given policy. This is actually far better organized and easier on the system than the Windows configuration of organize by-package, that can be moved around anywhere.
Yes--it's called "Search". And if you don't like using search, you know that an application's configuration is always in the registry, its files are generally always under "C:\Program Files\", and your documents are generally always under "My Documents". Pretty easy.
The fact you have to *SEARCH* for files shows poor organization. Had it been as well organized as *nix, you would be able to follow the simple policy rules to find your files.
Really? I'd sure like to see it. But all I've seen is that some things are available only as source, some are available as RPMs, some are newer RPMs that won't install on older versions of Red Hat, some require additional libraries to already be installed on your system that you might have to go track down, some are only available as slackware modules, and sometimes even when you get the source it won't even compile unless you manually hack away on some makefiles. You call that uniform and consistent? I call apt-get install a uniform and consistent way to install things.
Once in around.. a year I have to look for a package out of the Debian repositries, and get a tarball/rpm of the package, and compile/debianize it. I didn't have to manually edit a Makefile or so in a year, and probably won't have to in a long while.
Typical Windows installation:
- Windows user opens his repositry, browses it/searches it for the wanted package. (~2 minutes)
- Windows user reads installation procedure and downloads the file (~4 minutes, not including download time which is usually large due to typical static linking and including all dependencies in the package)
- Windows user runs the installer, answers questions, chooses installation location, etc. (~3 minutes, or ~6 minutes, depending on whether he needs a reboot or not. Not to mention possible loss of time due to reboot)
Typical Debian user installation:
- apt-cache/auto-apt search for the package (~2 minutes)
- apt-get install package-name (~15 seconds, not including download time, which is small, because dependencies are separated and only downloaded if required).
Compare the ~10-15 minute Windows installation with the simple, and *uniform* 3 minute Debian installation.
You can also consider the fact Windows Update is a bad joke compared to apt-get's dist-upgrade functionality, but that's another issue.
Which way to manage software is simpler, uniform and more productive?
Wow--I guess you've never heard of http://www.winfiles.com, http://www.zdnet.com, http://www.cnet.com, http://www.windowsupdate.com, or any of the other numerous central repositories of Win32 software that are out there. But why should that surprise me--you've obviously chosen to stick your head in a hole, obssess about GNU/Linux, and ignore the rest of the world. Those repositries are just links to various software packages. They are NOT central repositries, that I was talking about.
By *central*, I mean there is a *central* organization or entity, making sure they all work right with each other, and function as a *whole system*, not as a bunch of packages.
Like Redhat with its collection of RPM's, Mandrake, and the best in the regard: Debian.
Another unbiased, fact-based response from an obviously educated and objective Slashdot reader.
To summarize the points that you call a 'bias':
- I agree Linux configuration - when not automatically successful, can be complex for users, however, the hardware installation should be accomodated with the configuration - as it is with Windows (coming preinstalled from the OEM, and people installing the proper drivers for computer illiterate).
- Computer illiterate are successfully using KDE, everywhere. I haven't used Gnome in a while so in your comparisons, try to use KDE.
- Linux has official, central, and automatically managed package installation and upgrade systems, whereas Windows is typically using a bunch of EXE's and very weak registry settings to register installed apps - a poor excuse for a package system. Windows has no automated way of upgrading the system (No, not just Office and a few core componenets, the whole system - for its thousands of packages).
- Linux implements file system policies that make file locationing a predictable, simple, and senseful decision, whereas with Windows, although Microsoft published the hierarchy standards, nobody is following them, apart for My Documents and Program Files, which are a small part of the standard - go read it.
- Linux is more configurable and doesn't get in the way of users who want to do out-of-the-ordinary things, or fix problems that authors wouldn't fix on short notice (examples: strace, minikernels without support for devices you don't need, manual access to ALL of the running systems and their configurations, and of course source availability).
If you can't wait a few days for unstable packages in the unstable tree to be fixed - use stable or testing.
Console font problems, as well as specific software not working are easy to fix - and have little to do with Debian. In the worst case, its an installation script that went wrong.
If you can't handle a broken package or automatic configuration file generation going wrong - maybe you should consider using stable as well.
If you want the absoloute newest - you can usually get stable distro packages for it.
I've installed Debian on countless different machines at different people's, and it has always worked.
Sure, a package fails here or there (what you seemed to have going), but getting it fixed was as hard as fixing it in any other distribution. Most other distro's have broken packages much much more often in my, and others' experience.
What I said was that Linux has various distribitions, each with its own policies, friendliness, usability for different people, and stability.
Thus, it is ignorant and stupid to judge Linux by the newest Mandrake, or Debian, or whatever, because they *DO NOT REPRESENT Linux in its entirety*, and you *CANNOT concluce* that their faults are Linux faults and will be found in all distributions.
KDE *HAS* made Linux much easier for the average Joe, and I've *SEEN* average Joe's use it easily and successfully.
it is still insanely complex, poorly organized, inconsistent
If you think Linux is more complex than it should be, then you don't have a clue. Linux is designed pretty much as simply as ANY system with such functionalities can be designed. Grasping how Linux works is *MUCH* easier than grasping how, for example, the Windows kernel works, and involves much less details, because it is implemented much much more cleanly.
If you think it is 'poorly organized', then you are either referring to the file system - or to package layout.
In both of these fields, at least some distributions of Linux take the competitors hands down. Debian Linux, for example, has a VERY VERY organized file system, where each file belongs to a certain package, and its location is DICTATED by a STRICT SIMPLE policy. Can you say you can easily locate a file in Windows?
Hell, can you even tell what most of the files thrown around disorganized on a Windows HDD are? What rules dictate files' locations?
When it comes to software packages - Linux is also MUCH more organized - as there is a uniform and consistent way to install and manage packages, and often even CENTRAL repositries where those packages can be found, where you also know they will work with others. Windows, on the other hand, has packages thrown all over the Internet, without any central repositries or such centralization.
Think twice before comparing Linux's organization to others, its one of the best organized systems, in almost all distributions.
What the hell are you talking about when you mention inconsistency - I have no clue, because Debian GNU/Linux is the most consisten and uniform-way system I've ever seen.
and is more like an ongoing project for developers than a usable piece of software. Yes, Linux is an ongoing project.
Like Windows - almost all distributions have a 'stable' release that is 0..2 years older than the current unstable package versions.
Unlike Windows - the unstable packages that are more interesting to developers you seem to be mentioning, *ARE* available to Users.
I find Qt much prettier than Gtk+ in its default looks.. and it has wonderful styles.. Gtk+ has nice themes too, but it all looks very.. hmm.. dull, after moving to KDE/Qt:)
Actualy it would be nice if the whole concept of dot-files didn't exist..
A ~/Settings directory would be perfect, with subdirectories for specific application settings..
No need to hide any files, just organize them properly..
Slashdot Mirror
Windows is not much easier to install, it simply comes pre-installed :P
apt-get, kpackage, gnome-apt, the new Debian installer that is soon to come out, that rpm-getter tool
I'd say getting and installing packages is far easier in many Linux systems, even for a less-knowledgable user.
Wasn't Parrot a hoax?
Linux is only free if your time is worthless.
Free samples in the market are only free if your time is worthless (It takes time to take the sample and test it).
Linux saves time, on the long-run, so your claim is just a provactive idiocy.
Actually, many KDE programmers are from Trolltech, and support Opensource idealogy.
Trolltech are open out of:
A) some idealogy
B) it pushes Qt, which strngthens their grip in the commercial world, and allows them to 'show off' samples of running code, such as KDE.
Closing Qt in light of the two is unlikely, as most of their profit is from their closed-source buyers, who pay a Single-fee for Qt, with yearly upgrade fees. No need to pay per-sale
(Knowing this, as I worked for a Qt-using propriarity software company).
Unless they whole business model, they would have to sell Qt to the developers, because buyers don't pay for Qt.
Ironically, many of those developers work in Trolltech, and Trolltech know would never pay.
They also know KDE has no chance of succeeding in a closed pay-per-copy license, in the opensource world.
I use Konqueror every day, but there are just a few things I feel are missing. The articles didn't mention these, though.
:)
heh, same here.
2. It would be nice if I could put my favorite links on the menu bar, like with Navigator.
Settings->Show bookmark toolbar
if that's what you mean.
3. The bookmarks menu demands that I hold down the mouse button while swishing through my bookmark folders. If I accidentally let go, I end up with the wrong site, or all too commonly, get the "edit bookmarks" page.
This behaviour depends on the Style/theme you've chosen (look&feel styles). In the default one, you can simply click (and let go immediately) on the bookmarks menu, and wander without having hold the button.
4. There is noooo rule four.
I don't get it
5. Konqueror still croaks on various web sites. I don't know if it is the complexity, or maybe something to do with managing the color palettes. (My xterms are fixed - graphics upgrades are impossible...)
It happens very rarely that web pages cause trouble. Once in a few days of intensive browsing it crashes, but I never had it "croak up".
6. They did mention the loading time, but I'll still mention that it is slow. Sure, maybe my P150 was not up to snuff, but an AMD 800 with 256 MB of DDR?
It seems to load fast here (a matter of 1 second or 2 at most), but maybe its because I'm running KDE.
What the fuck have you been smoking? This is not acceptable. Free software should be free to everyone, for any use. Those Trolltech fucks keep weaseling around, trying to avoid admitting they're still a proprietary software company, using open source only for advertisement. Fuck them, fuck KDE.
Heh, you could have had a legitimate point back in QPL days, but now they use the GPL, so you're a complete troll.
Why do we need another web browser? :P
:P
We don't, we have Konqueror
Do you have some problem with Mozilla that we should know about?
Besides for taking 25 seconds to load for the first time (in a while) on my Athlon 750 (for comparison, Konqueror takes at most 4 seconds on the first load)?
Oh yes, it also seg faults when I try to load it now:
ii mozilla-browse 0.9.3+0-3 Mozilla Web Browser - core and browser
Yes, it also doesn't display Hebrew pages as nicely as Konqueror, which automagically selects the proper encodings and fonts.
Its also not using the very very powerful KDE IO slaves and KPart architectures, which allows centrally writing code to support protocols, format interpretation, etc.
I'm sure some asshole will moderate this post troll or flamebait, but I'm 100% in earnest here.
You sure do deserve a 'troll' moderation or at least 'misinformed' (theoretic moderation) for reasons I will outline below.
Did you ever wonder why Mozilla is continually falling behind schedule? Because people like the Konqueror team decide to go off on their own instead of working for the good of the community. Mozilla was there first, and it deserves the support of the community.
Konqueror is *NOT* a web browser. It can be *used* as a web browser, but it is much more. It merely encapsulates the KParts/IOSlaves of KDE, with some nice toolbars, menus, and cute features on the side. As an example, Konqueror can use Gecko for rendering instead of KHTML.
KHTML is the actual rendering engine. KHTML was there BEFORE mozilla. KHTML was there before Netscape opened its code.
Mozilla people decided to REWRITE the rendering engine, rather than using the already-existing KHTML code!
They should have joined the KHTML/Konqueror project.
If the free software community wants to make a good impression on the business world (and it may already be too late), we must, at all costs, avoid splitting into tiny, useless factions working on useless, duplicate projects.
Konqueror is very useful. Mozilla is very useful.
Who cares about the impression of the business world?
Here's an idea: before starting your new project, check to see if someone is already working on a similar project. Had the Konqueror team observed this little suggestion, the whole Konqueror fiasco could have been avoided.
You mean the Mozilla fiasco could have been avoided, as KHTML was there first
Maybe you should check on your facts, or read about the behind-the-scenes architecture of Konqueror and KDE, and that of Mozilla, and compare which is more powerful and flexible, and allows for more code-reuse/etc.
Konq being a browser simply results from its use of the KPart for HTML viewing.
Just as it views an image or a PDF file, or a system directory, using the proper viewing KPart.
Its ability to get web pages via the HTTP protocol is a simple result of the HTTP IO slave which is part of the powerful KDE architecture. Konqueror merely encapsulates KDE technologies, which are so powerful, that a subset of them form a web browser.
One Word: Arrogance.
Thats the problem with the KDE developers. Since day one they closed they're ears and eyes to reasonable answers. The fact they started with a proprietory/not free toolkit underscores the prevelant 'we dont give a shit about you' attitude the KDE project has.
Qt was always open.
The only difference between the QPL and the GPL is that QPL has some section meant to prevent the forking of Qt into a free project that will compete with Trolltech.
Its a very reasonable license in its opensource terms.
They could never accept a browser like Mozilla for one reason: They believe they are better.
Mozilla has matured into a stable and fast web browser with an unbelievable amount of power. But the KDE project just blows it off and keeps on believing their browser is a superior product.
Again, as someone else said, Konqueror is *NOT* a browser. It is a framework, encapsulating the KParts.
The actual browser is KHTML.
KHTML was usable before Mozilla was, so the question should be reversed, and truly, I believe that Mozilla people *should join the development of Konqueror and KHTML*, which *IS* a better product.
Sorry to break this to you , but Konquerer is a cheap imitation of the real thing.
Immitation or not, Konqueror is *usable as* a fast usable web browser, with powerful and quick viewing of many file formats from anywhere (of which HTML is simply one of), including the local file system, FTP, Samba, etc.
(am I biased ? I dont know. I use KDE for my desktop.. but use mozilla/evolution daily.)
When was the last time you *seriously* tried using Konqueror instead?
That's what the non-us Debian server is for :)
Seriously, there's a workaround for every problem.
With new algorithm-generator software, that spits out many many patented algorithms, we may see a change in the field of software patenting.
Also, Debian could claim prior art to most patented allegations, and/or opensource authors could put the projects under people who are not restricted by the patent, etc. etc.
Are you kidding? Object foo knowing how to bar itself is quintessential OO. Then you can have object foo2 knowing how to bar itself, or if they're derived from the same superclass & bar can be abstracted a level, move bar up to the superclass. In smalltalk to print an integer, you'd do:
:P
(5 + 4) printNl!
5+4 evaluates to an Integer object 9, which then prints itself.
If object foo knows how to Stringize itself, then foo2 also does!
print (5+4)
and its not that 'print' itself knows how to handle various objects, it merely calls the object's __str__ method, which is clearly polymorphic.
Making a (to the user) generic print function to which you pass a string/integer/character may be passing around objects internally, but if you do that (as in the snipped of python you post) the language might as well be procedural from the user's point of view - the benefits of OO were used in language design, but are lost during the actual language use.
The only 'advantage lost' is the syntatic sugar of using:
OBJECT-IDENTIFIER . FUNCTION-IDENTIFIER (Extra Parameters)
rather than:
FUNCTION-IDENTIFIER ( OBJECT-IDENTFIER, Extra Parameters)
This is an issue in Python (And Smalltalk).
The reason I said its not very OO'ish for an object to know how to print itself, because an object should not implement EVERY specific feature anyone might have thought of.
Do you think every object should implement a 'describe vocally' method, so that you could: '(5+4) say' too?
What about '(5+4) encrypt' or '(5+4) encodeMp3Sample'?
If you simply remove the difference between the notaitons:
f(a,b)
and:
a.f(b)
as it is in LISP, you get the consistency and the power.
However, it makes it a bit harder to find the list of available methods (which can be resolved easily).
Anyhow, Python allows you to:
object.print = my_print
so it can print itself
But the generic print is more OO'ish, because an object should have a minimal set of things it can do, one of them is represent its state as a string. In Python, another is returning a state object describing the state uniquely, in a way that can be reconstructed.
its dumb, because a print function would simply be better.
The fact its quite hard to mix strings with variable names and the need to convert between types shows you - not that you need a specific 'print' that helps you convert everything to a specific type - string, but that you need a general method of converting everything to a specific type (in this case string).
And this method is already available:
def print(*vars):
for var in vars:
sys.stdout.write(str(var) + " ")
sys.stdout.write("\n")
Or something of this sort.
Is: print ("Var:", var) so much worse? I think its more powerful, allowing things such as nesting prints in lambda's, etc.
I think that it would be cooler to steal the one cool Perl feature here:
print ("Var: $var") and allow an escape code such as '$' to address the non-literal namespace of variables, and convert whatever to string via __str__.
- Smalltalk's method call syntax
:)
- LISP's macros
- The language that first introduced lambda's.
- The language that first introduced closures.
- Simula's code structuring
- Python's array slicing syntax (not sure about this)
- Python's indentation-is-syntatic
- etc. etc. etc.
Don't let Microsoft convince you innovating is collecting pieces from all around and sticking them together in a single bunch.
And just for the record - I hate Java, note it hasn't invented anything either
I agree that the built-in print is probably a dumb addition to Python. But dumb additions are harmless - whereas the lack of useful additions is harmful.
Trying to be pure usually results in the latter.
Built-in types do not stink, as they allow for better performance. You could always use the class-wrappers of the built-in types in Python - so NO functionality was ever missing.
is this truly OO? :)
:)
and in Python, its: print 'Hello World'
If you want to describe how your objects 'print themselves', in Python its better done in a more generic way, you write an __str__ method in your object, that knows to represent it as a string for printing, but as an added bonus - it can be represented as a string for ANY purpose!
Python has many advantages over Smalltalk or LISP.
Its code is highly readable even by newbies, and it has many syntatic sugars that make the code elegant and clean.
Python saves every bit of redundancy:
- Variables aren't declared beyond their assignment, whereas in Smalltalk, they are.
- There are no explicit indentation-markers, indentation itself is used. This means there is no redundant and error-prone representation of the code which may lose sync with the real representation of the code hierarchy, whereas in Smalltalk (and LISP, btw) this redundancy is there.
And there are various other points where Python saves you, and increases the elegance of the code - at the expense of making the language perhaps less 'pure'.
Python also has MUCH more practical code available - graphics, 3d, gaming engines, COM object interaction, etc.
LISP is a whole other story - its probably more powerful - but its mostly unreadable by code illiterate - and I'm not sure about the practical library availability either.
Microsoft was out there INNOVATING
Name a single feature in C# that hasn't appeared in at least 3 languages prior to it.
Anyhow, I'd take Python, C++ or LISP before I'd take C#, for any task. (The first for elegance and ease, the second for performance, and the third for its power).
I agree Ruby and Perl seem to be like the latter type. :)
Python, however, is a truly clean, well-designed, quite-pure, strong-typed language.
I wouldn't say it has a Basic odor at all, besides for a built-in 'print', perhaps
Basic is braindead and has no library - instead, it has its entire library built into its syntax in a horrid way.
Python is the exact opposite - many many libraries separated into modular modules, with only flow constructs, OO organization, functions and exception handling built in.
Konqy :)
I think you mean "Apparently very few reading skills". But then what do I know? All I can do is spell and use proper grammar.
.bashrc file or whatnot, suddenly certain applications don't work and they have no idea why. More examples? Every application stores its configuration files in a different file format, and often times you have to learn a new scripting or programming language just to accomplish something as simple as getting X-Windows to work right with your video adaptor.
You probably speak better English than I -- not that it means much, English being a second language I don't speak often.
That's not what you said. Maybe that's what you meant to say, but it's not what you actually said. Sounds like you're the one with the reading problem. Try reading your own messages before hitting the "Submit" button.
I just re-read my original message - and its blatantly trivial I was bitching about the judging of Linux according to a specific configuration. I said:
"It seems you have a common disease - assuming Linux has One look, One way, One policy - as you have in Windows.
No, there is no 'general' way distriubtions handle errors and no 'default' graphical mode that distributions always select. "
Aka: Linux is not an OS, its a kernel, and GNU/Linux is a minimal OS that serves as the core OS of distributions.
It does not contain the management code - therefore you cannot bitch about its management sucking - you'd have to bitch about a distribution's management - but noone seemed to do that.
Really, is that so? Can you back up that statement with facts? I didn't think so. Anyone who thinks that a GNU/Linux system has to be as insanely complex and disorganized as it is just to provide the functionality it can offer is simply too lazy to learn anything about usability or human-machine interfaces. There are numerous specific aspects of any given GNU/Linux distribution that simply wreak from an unnecessary lack of usability, consistency, and simplicity. Want some examples? Just look at the recent corporate analysis of GNOME's usability that was posted on Slashdot and you'll see plenty.
You obviously have some basic misconception about the terminology of this discussion.
Linux is a kernel.
GNU/Linux is a core OS.
A distribution is a whole OS, bundled with all the software you may need.
When saying 'Linux is complex' - you are claiming that the kernel is complex. Or even 'GNU/Linux is complex' is claiming that the kernel, plus the set of simple command line utilities are complex.
These systems are very simple. The kernel is very clean and simple, and the tools are clean and simple.
What you mean to say - is that the Linux user interfaces are 'insanely complex'.
I have little experience with Redhat or Mandrake, but from that little experience - they seemed to have auto-configured hardware/software/etc. They also set up a trivial-to-use KDE desktop, where all I need to do is point&click, and all my hardware is supported.
I don't see where you claim the UI has any insane complexity, but perhaps you're referring to distributions such as Debian and Slackware - where the user is *expected* to know how to handle the lower-level interfaces.
If you try to argue that Gnome is not very usable - I'd agree. I don't use gnome, although many others find it usable.
I use KDE, my parents use KDE, and they are VERY computer illeterate.
Really? That's amazing. Then apparently Windows 2000 doesn't actually exist, and is a violation of some previously undiscovered universal law of reality. It offers the same overall functionality in a much more usable package--and that is they key reason people use it over any UNIX-like system.
Win2K is much much more complex than Linux -- using the correct terminology ofcourse.
If you look under the hood - you'll find a MUCH cleaner system under Linux, than under the Windows hood.
When comparing the user interfaces - KDE is more configurable, better looking (to most viewers), and much more quickly evolving.
Why do you incorrectly assume that I can't be speaking about more than either of those? While both the filesystem and "package layout" are poorly organized, GNU/Linux systems suffer from numerous other disorganizations. For instance, many applications hard-code filesystem paths into their source, so you actually have to recompile an application from source code if you simply want to install it to a directory of your choice. Another example? There is no distinction between a user's environment variables and system environment variables as there is in NT, so if a user unwittingly blows away a necessary system environment variable with their
Actually, automatic X configuration with Redhat/Mandrake is quite easy, and from my exprience - they autodetect the hardware.
If not - tough luck - you have to work through some admittedly more complex UI's to configure them.
Appearnatly you're missing that basic configuration is a one-time task that, like the hardware installation - can be done by a techie who knows what he's doing.
Using the machine is where UI usability should be compared.
Anyhow - why the hell would you be learn a programming language when configuring X?
About the 'standard configuration format', called the registry, that must be a hilarrious joke - because locating registry keys and interpreting their representations is at least as difficult than editting the dot text files in the home directory - not that you have to, given that applications provide a UI to edit their configurations.
I've never encountered any difficulties with my option to override system variable values.
Hardcoded paths are not a major problem - because the system is so well organized, that unlike Windows - files *DO NOT MOVE AROUND*. You are *NOT SUPPOSED* to install it in a 'directory of your choice'. Organization *leaves you no choice*, because choice, not based on policies is simply *unorganized*.
There is only ONE right location for a file under any given policy. This is actually far better organized and easier on the system than the Windows configuration of organize by-package, that can be moved around anywhere.
Yes--it's called "Search". And if you don't like using search, you know that an application's configuration is always in the registry, its files are generally always under "C:\Program Files\", and your documents are generally always under "My Documents". Pretty easy.
The fact you have to *SEARCH* for files shows poor organization. Had it been as well organized as *nix, you would be able to follow the simple policy rules to find your files.
Really? I'd sure like to see it. But all I've seen is that some things are available only as source, some are available as RPMs, some are newer RPMs that won't install on older versions of Red Hat, some require additional libraries to already be installed on your system that you might have to go track down, some are only available as slackware modules, and sometimes even when you get the source it won't even compile unless you manually hack away on some makefiles. You call that uniform and consistent?
I call apt-get install a uniform and consistent way to install things.
Once in around.. a year I have to look for a package out of the Debian repositries, and get a tarball/rpm of the package, and compile/debianize it. I didn't have to manually edit a Makefile or so in a year, and probably won't have to in a long while.
Typical Windows installation:
- Windows user opens his repositry, browses it/searches it for the wanted package. (~2 minutes)
- Windows user reads installation procedure and downloads the file (~4 minutes, not including download time which is usually large due to typical static linking and including all dependencies in the package)
- Windows user runs the installer, answers questions, chooses installation location, etc. (~3 minutes, or ~6 minutes, depending on whether he needs a reboot or not. Not to mention possible loss of time due to reboot)
Typical Debian user installation:
- apt-cache/auto-apt search for the package (~2 minutes)
- apt-get install package-name (~15 seconds, not including download time, which is small, because dependencies are separated and only downloaded if required).
Compare the ~10-15 minute Windows installation with the simple, and *uniform* 3 minute Debian installation.
You can also consider the fact Windows Update is a bad joke compared to apt-get's dist-upgrade functionality, but that's another issue.
Which way to manage software is simpler, uniform and more productive?
Wow--I guess you've never heard of http://www.winfiles.com, http://www.zdnet.com, http://www.cnet.com, http://www.windowsupdate.com, or any of the other numerous central repositories of Win32 software that are out there. But why should that surprise me--you've obviously chosen to stick your head in a hole, obssess about GNU/Linux, and ignore the rest of the world.
Those repositries are just links to various software packages. They are NOT central repositries, that I was talking about.
By *central*, I mean there is a *central* organization or entity, making sure they all work right with each other, and function as a *whole system*, not as a bunch of packages.
Like Redhat with its collection of RPM's, Mandrake, and the best in the regard: Debian.
Another unbiased, fact-based response from an obviously educated and objective Slashdot reader.
To summarize the points that you call a 'bias':
- I agree Linux configuration - when not automatically successful, can be complex for users, however, the hardware installation should be accomodated with the configuration - as it is with Windows (coming preinstalled from the OEM, and people installing the proper drivers for computer illiterate).
- Computer illiterate are successfully using KDE, everywhere. I haven't used Gnome in a while so in your comparisons, try to use KDE.
- Linux has official, central, and automatically managed package installation and upgrade systems, whereas Windows is typically using a bunch of EXE's and very weak registry settings to register installed apps - a poor excuse for a package system. Windows has no automated way of upgrading the system (No, not just Office and a few core componenets, the whole system - for its thousands of packages).
- Linux implements file system policies that make file locationing a predictable, simple, and senseful decision, whereas with Windows, although Microsoft published the hierarchy standards, nobody is following them, apart for My Documents and Program Files, which are a small part of the standard - go read it.
- Linux is more configurable and doesn't get in the way of users who want to do out-of-the-ordinary things, or fix problems that authors wouldn't fix on short notice (examples: strace, minikernels without support for devices you don't need, manual access to ALL of the running systems and their configurations, and of course source availability).
If you can't wait a few days for unstable packages in the unstable tree to be fixed - use stable or testing.
Console font problems, as well as specific software not working are easy to fix - and have little to do with Debian. In the worst case, its an installation script that went wrong.
If you can't handle a broken package or automatic configuration file generation going wrong - maybe you should consider using stable as well.
If you want the absoloute newest - you can usually get stable distro packages for it.
I've installed Debian on countless different machines at different people's, and it has always worked.
Sure, a package fails here or there (what you seemed to have going), but getting it fixed was as hard as fixing it in any other distribution. Most other distro's have broken packages much much more often in my, and others' experience.
Appearantly you have very little reading skills.
What I said was that Linux has various distribitions, each with its own policies, friendliness, usability for different people, and stability.
Thus, it is ignorant and stupid to judge Linux by the newest Mandrake, or Debian, or whatever, because they *DO NOT REPRESENT Linux in its entirety*, and you *CANNOT concluce* that their faults are Linux faults and will be found in all distributions.
KDE *HAS* made Linux much easier for the average Joe, and I've *SEEN* average Joe's use it easily and successfully.
it is still insanely complex, poorly organized, inconsistent
If you think Linux is more complex than it should be, then you don't have a clue. Linux is designed pretty much as simply as ANY system with such functionalities can be designed. Grasping how Linux works is *MUCH* easier than grasping how, for example, the Windows kernel works, and involves much less details, because it is implemented much much more cleanly.
If you think it is 'poorly organized', then you are either referring to the file system - or to package layout.
In both of these fields, at least some distributions of Linux take the competitors hands down. Debian Linux, for example, has a VERY VERY organized file system, where each file belongs to a certain package, and its location is DICTATED by a STRICT SIMPLE policy. Can you say you can easily locate a file in Windows?
Hell, can you even tell what most of the files thrown around disorganized on a Windows HDD are? What rules dictate files' locations?
When it comes to software packages - Linux is also MUCH more organized - as there is a uniform and consistent way to install and manage packages, and often even CENTRAL repositries where those packages can be found, where you also know they will work with others. Windows, on the other hand, has packages thrown all over the Internet, without any central repositries or such centralization.
Think twice before comparing Linux's organization to others, its one of the best organized systems, in almost all distributions.
What the hell are you talking about when you mention inconsistency - I have no clue, because Debian GNU/Linux is the most consisten and uniform-way system I've ever seen.
and is more like an ongoing project for developers than a usable piece of software.
Yes, Linux is an ongoing project.
Like Windows - almost all distributions have a 'stable' release that is 0..2 years older than the current unstable package versions.
Unlike Windows - the unstable packages that are more interesting to developers you seem to be mentioning, *ARE* available to Users.
Get a clue.
I find Qt much prettier than Gtk+ in its default looks.. and it has wonderful styles.. Gtk+ has nice themes too, but it all looks very.. hmm.. dull, after moving to KDE/Qt :)
Actualy it would be nice if the whole concept of dot-files didn't exist..
A ~/Settings directory would be perfect, with subdirectories for specific application settings..
No need to hide any files, just organize them properly..