The articel itself is standard Gnome propaganda again: "blabla, your file organisation is wrong, you are wrong, but our filemanager just must be right, blabla". Gnome has a whole lot of great ideas in it, but all to often Gnome maintainer seem to think they know better then the user what is good for them. Removing or hiding configurabilty in GConf makes this even worse.
Anyway, back to another issue in the article which I don't get at all, at the end Eugenia talks about Spartial Mode and DB-based filesystem with no folders at all and how they 'mix' well. Well, how do they mix? Isn't spartial actually the completly opposite of a DB-based filesystem? In spatial mode you represent folders as windows, one window represents exactly one folder, its size and position is safed, so that the window actually becomes the folder from a users point of view. With a DB-based filesystem however there wouldn't be folders, just data sorted in whatever way you need it at the point, so you would have lots and lots of different views onto your data, no windowfolder corelation which seems to be what spatial is all about.
Re:The hell? Linux isn't different at all
on
Ten Years of BeOS
·
· Score: 1
It's still getting a whole bunch of new applications that are mere shells of the commercial implementations they're trying to emulate. You may as well not even consider them in the equation.
The important property of Open Source is that it doesn't have to be better than its commercial counter parts to be successfull, it just has to be good enough for whatever the users are using it. And seriously, all my word processing needs had been fullfilled by AmiPro in the days where 386er where still normal, there is nothing on this earth that even could get me remotly motivated to even start thinking about buying MSOffice 200X, OpenOffice, AbiWorld or LaTex are more than good enough for my needs.
OpenSource might not be up to commercial applications nor might it be dominant on the marked (how could it?, since its free...), I however still prefer my Linux system any day, about throwing out a few thousand Euros for what I would need to get the commercial equivalent.
Is it really the harddrive? Sure it might be the hardware part that could give the best boost in todays system, but how about fixing the boot issue at the software level? I mean boot speed hasn't been much of an issue big in the past, yet, today it suddenly is, even while harddrives and CPUs and RAM are much faster then a few years or even decades ago.
What went wrong on the software level that systems now take so long to boot? Are it generic kernels that probe for a lot of non existing hardware and thus basically just spend their time waiting for timeouts? Or is it todays software with all its dynamic linking and its dependecy (heard the slowness of KDE was in most part a fault of ld.so)? Or is it just bad organisation of boot scripts, ie. lack of paralelims while starting them, useless stuff first, so that the important stuff has to wait and such?
I can second that, the spiral tool seems to be rather useless and especially rather limited. Nothing wrong with having one, but it shouldn't be limited to spirals. A lot of other drawing programms have stuff like that too, but they don't only provide spirals, but also pseudo 3d blocks, arrows and other stuff.
About the missing layers support, I don't consider it much needed in a vector graphic application, yes it should be there and is a nice additional feature, but with grouping one can already get pretty similar things done.
However one very important feature currently missing is blending (ie. morphing one polygon shape into another), without it its almost impossible to get good looking shadows and highlights into vector graphics. Sketch already had blending for quite a while.
I would actually do it the other way around, skip the whole theory completly for the beginning and only come back to it at a much later point if there is still interest in programming. Sure theory is important if one wants to be a real programmer, but just for having a little fun with a programming language its not needed at all and theory isn't really thu fun part of it anyway for most people. Stuff like sorting, linked lists and such all come build-in with todays programming language, so not knowing about how they are implemented doesn't even hurt at all, since one can write a lot of usefull programms even without knowing the details.
Speaking about programming language I would start with Python, which I find a quite a bit easier to understand then Scheme from a beginners point of view (no need to declare variables, familiar infix syntax) and it also has an interactive interpreter like Scheme:
>>> a = 5 >>> b = 10 >>> a + b 15 >>> def add(x, y):... return x + y... >>> add(a, b) 15
Above code should be easy enough to understand and should already be enough to have some fun.
The only things I miss in Python, which for example QBasic provided, is the IDE (not sure if there if there are any easy ones for Python) and more important a easy way to get access to the screen, in QBasic it was just:
SCREEN 9 LINE (10,10)-(100,100)
and you had a line on the screen, with Python I don't think there is anything that simple, most GUI packages always require a whole lot of initialisation, which is really not beginner friendly.
why do you NOT have a problem with your web browser being able to view a myriad of image formats but not a myriad of video formats?
There actually is an important difference between and image format and an embedded video/flash/javascript thingy. A image is 'passive' it just stays there, it doesn't waste CPU resources, it doesn't jump at you or anything, its just passive. A video, flash or javascript (or all together) on the other side is 'active', they do things on their own, often are impossible to stop, let new windows popup, deadlock the browser, or just waste CPU resources and other not so nice stuff.
Sure with image formats such as mng or animated gif, things aren't so clear any more, anyway, point is that active content is evil most of the time and does things the user does not want. This is however not so much the problem of the active content itself, but more of the webbrowsers that often provide little or no help for the user to stop or control the active content, an external viewer does most of the time a lot better job here then some half-broken browser plugin and thats the major reason why some people prefer external viewers over plugin in stuff.
Well, yes, we are at a point at which we could make small computers, but the computers for the masses are still rather large. Not really because the users need bigger and faster computers (for a bit of Office computers a few years ago were already more then fast enough), reliability and silence would be far more important, but just because they still sell better, luckily thats slowly changing now.
Getting a tiny, noiseless computer that you can plug into your normal monitor, tv and USB keyboard are still far away from being the norm and in most cases either quite expensive or very hard to get.
The main driving force behind computer development the last 10-20 years where just making them faster, but not better or more userfriendly, luckily we are slowly reaching a point where computers are almost fast enough for almost everything, especially when it comes to normal use, so that the arguments for a new computer no langer can be just 'faster'.
Just consider the last 15-20 years, back then computer where very small, directly build into the keyboard, almost made no noise beside the drives, could be plugged into every TV and 'just worked', no need to maintain or reinstall the OS really, just insert another floppy and viola it worked.
Today computers are in most cases big noisy grey boxes. People have to reinstall or maintain their OS, manually install security patches and every once in a while a nice internet worms does funny things with your machine. GUIs are still pretty much the same as the first GUIs that apeared, just a bit more colorfull here and there, but fundamentally still pretty much the same.
Beside getting faster, bigger and louder and more throublesome in maintaince there really wasn't much change in the computers you normally use in the last 15-20 years, sure memory has grown bigger and now we can do some cool new stuff, but fundamentally not much has changed, some issues have even getting worse.
That said, technology might improve expotentially, user interfaces however do not. Just look at something basic like 'undo', neither Linux, Windows or MacOS provides an 'undo' on filesystem level, instead they provide this ugly workaround of a trashcan which doesn't help on a plain 'rm' or if you overwrite a file instead of deleting it. Versioned filesystems however are nothing new and should be easily doable with todays machines, however non OS does it that way, which just shows that hardware might be evolving fast, software however is not.
One of the major benefits of '10' is that the human has 10 fingers. And as a lot of people use their fingers to do everydays calculation, using a base 10 system helps a whole lot here making the calculations easier and feel more natural.
Well, a file manager is a pretty damn important part of a desktop environment, if that doesn't work and isn't easy to make work than for sure thats a big problem.
From my experince the removal of preferences or the move of them into GConf without any visiblity in the configuration dialog in Gnome2.x has been a whole LOT more confusing than just another setting in the configuration dialog.
If I don't know what it does, I don't touch it, if I messed things up I reset to default, very easy. With todays Gnome2.x I have to search google to first find out that a setting exists at all and then start browsing through gconf settings to tune them, thats nothing what I would call userfriendly.
In general I agree that config dialogs are often a bit overfull, but the way to pretty much remove everything that the Gonme developers are going is far worse.
The major problem here is that Gnome doesn't run in a Gnome-only environment. I use Gnome together with a whole bunch of non-Gnome apps, which in turn pretty much requirse to tweak a few Gnome-settings to run sane (color-scheme, panel behaviour, focus behaviour, etc.). Since Gnome2.0 a lot of settings necesarry for this have been gone into GConfi-only, no more GUI to tweak them, I wasted quite a few hours to get back to Gnome1.4 useablity since then.
Same with the spatial Nautilus, in a few-files/directory environment it might actually work quite good, after all on the Amiga it worked very well. But back then harddisks where small and storage were rare. Today I have a harddrive with deep directory trees and half millions of files, spatial doesn't work very well for that. Since Nautilus is also missing a useable smallicon/column view or tab-completion (Rox like for example), it makes it almost impossible to navigate through such an environment, since I have a lot to scroll, click windows away and such.
In general I really have the impression that Gnome is designed to be a cute desktop environment, that when run alone in a 'clean' environment might give a good experince, but which pretty much falls apart in most real world cases where people work with more than a handfull of files.
Am not the one asked, but lets try to answer anyway:
- ugly colorfull default icons, tuning for colorfullness, not for usefullness - to much mimicing of Windows all over the place - horizontal scrolling in konquor - bad redraw behaviour in konqueror - same stupid way to handle overfull icon panels as windows - to crowded menus (Configure Shortcuts, Configure foo, Configure bar, Configure baz, Configure foo), makes it hard to find the right thing - extreme long app startup times if one doesn't already run KDE
And here some screenshots which illustrate some of the problems:
http://pingus.seul.org/~grumbel/tmp/kdesucks/
Anyway, about your points I fully agree, Nautilus is a joke and configuration options are also a bit lacking. I however really like the default look, since it doesn't try to make itself 'unreadable' by using gradients all over the place.
I like the review. Is it flaimbait? Maybe. But is it true? I would say yes, he raises some very good issues that I know a lot of people have with Gnome. As with my day to day experince in using Gnome (I don't like KDE much, and a plain WindowMaker just isn't enough) I can fully agree with him.
Nautilus is in most part still a rather unpleasent piece of a file browser, it has been getting faster over the time, but thats it, its still missing some very basic features, smallicon-view being one, a simple button to go into the parent folder being another, but well, there are lots of more things that are just horrible broken in Nautilus (Mime-Handling and friends come to mind...). I really don't get why Gnome people got started with Nautilus in the first place, its after all just a very under-featured filemanager that was unusablily slow in its first few releases and only slowly got better. In the time they spend fixing it up, they might have at well written two other filemanager. And in its current stage I really doubt that it will reach a stage were I would consider it useable any time soon (maybe in gnome 3.0 or 4.0 or so, but it will take time...).
And well, the GConf issue, he is correct has well. Gnome Devolper have moved a lot stuff from Config Dialogs into obscure GConf settings once Gnome went from 1.4 to 2.x, what should that we good for? Basic settings that are needed to tweak Gnome into a useable stage are well hidden in some GConf settings now. I still prefer Gnome1.4 over Gnome2.x for exactly that reason. And well, this isn't only a Nautilus problem, it has tortured many Gnome apps, especially Galeon has suffered a lot, which again I consider the Gnome1.4 based builds a lot better then the Gnome2.x ones.
Anyway, as long as Gnome developers will just ignore the issues the users have I don't see much room for improvement in Gnome itself. They have multiple times in the past shown that they don't seem to care about who actually uses Gnome, but instead have optimized it for some 'hypotectical newbie user' which I have yet to meet in reality.
And how does catching this guy help here at all really? It doesn't stop the virus, it doesn't undo the damage the virus has created, neither does the guy have enough money to pay the once that got damaged. At best it might stop this guy the write another virus, but thats it, there are thousands of people out there that happily will jump in and create new viruses, this guy wasn't some special super brain, writing that virus wasn't that difficult.
I get the feeling that this is really just a huge PR stunt so that Microsoft can say "Hey, we did something and catched the guy!", thus hiding the fact that the real problem is Microsoft that can't get their OS secure. Viruses and Virus writer are just the symptom, not the cause of this whole issue. If people wants to stop this happening again they need to start at Microsoft, not at some lonly virus writer somewhere out there.
Sounds like a missunderstanding, lets quote a bit:
The development of a Multi-Platform 3D Graphics Rendering Engine and the creation of a hardware accelerated Ignalum Linux OS based on OpenGL allows applications/games developed for the engine to run using OpenGL or DirectX, running on Linux or Windows, without having to change a single line of code.
This sounds like they are just developing yet another 3D engine that runs under OpenGL or DirectX, like there are already dozens of them out there, nothing special. This doesn't sound to me like they would try something like WineX that allows to run DirectX games directly under Linux, far from it.
Beside that WineX already does run a lot of games on Linux, yet, no breakthrough. While I agree that games are a very important factor in Linux, we are not at a point were they are really that important. As long as I still have to dive through weird config files to get my printer working and have to update or patch my Kernel for a more or less common piece of hardware (graphic tablet, 3d card), Linux won't go mainstream. Linux needs one standard and userfriendly way to handle and configure hardware, not one or two ways to do it for each distro. Linux has all the capabilities there, it just needs some more agreements on standards across distros.
Well, we might not have ToyStory graphics, but we are damn close to it. The problem however isn't graphics, but gameplay, camera movement and such. Until we get the graphics to ToyStory level it is just a matter of time, however there hasn't really been any major improvments in gameplay or camera movement over the last years, just little improvements here and there. And as long as you just have your little 3d puppet with a camera behind it and jump through some jump'n run levels, it will neither look nor feel like the ToyStory movie, no matter if the graphics itself match in polygon count and ligthing.
The problem is that even with story and acting up to Pixar's level you still wouldn't have a game, you would have a movie, nothing more. And how does one fit good gameplay into a heavily story/acting dependend game anyway? So far the only genre that did that reasonably well where adventure games, but well, they can life quite without using cutting edge realtime graphics and instead be based on prerendered stuff without to much problem. The games that need fast realtime graphics just too often end up being rather dumb fast action games, where story, if at all, just takes place in cutscenes just too often just slow down the gameplay instead of improving it.
While I agree that fun is really important and should always be number one goal, the story line can be extremly important as well. Sure with Mario type games its not really important, but with adventure games, some rpgs or other dialog-heavy games it can play a major rule, ruin the story line and you ruin the game, no matter if the gameplay itself is good.
Gimp has been around since 1996, which makes around 8 years of development, nothing that I would call a 'short time'. That said looking back at earlier Gimps the progress isn't really that impressive its still pretty much the same Gimp as back then, just with cleanup here and there. On the code side I think they have rewritten it almost completly now, but from the userinterface it has still many of the same faults as years ago.
If you read the article you will find out that the features Gimp provided where not the main problem, he gives a little chart which looks like:
Features 80% Ease of Use 25% Value for Money 10% 'Must Have' Factor 1%
And I can pretty much with that. Gimp provides heaploads of plugins for a lot of stuff, but usebility really reduces the overall picture by a huge amount. Things as lack of preview for example make a lot of plug-ins alot harder to use if not even completly impossible to use effectivly (ie. Plugin->Ok->Undo->Plugin... cycles where in other apps you get a realtime preview when you move the slider).
The problem is that Gimp doesn't need yet another little contribution, but a major redesign in quite a few points. Sure newer and cleaner dialogs are a good thing, but they don't remove this 'GUI is broken' issue that Gimp has when comparted to other Apps.
See MDI for example, every time it is mentioned in the mailinglist, IRC or whereever it ends up in huge flamewars, developers mentioning that it is useless and stuff like that. Getting something MDI like implemented in the current Gimp is something that I consider very difficult if not impossble, after all the issue is there for quite some years. Doing it in a fork would be quite easier, since it removes the whole time-wasting flamewar aspect of it.
I have though about that, but I would actually go another step further and start from scratch instead of basing stuff on the Gimp Code base. I haven't reviewed the codebase in depth, however looking at the not really existing progress over the last years something in the basics just has to be very wrong. I mean Gimp0.99 is pretty much the same as Gimp2.0, yes, it has a heapload more plug-ins on board, but thats basically it, the interface hasn't changed so much beside some minor cleanup and the basic tools are still the same (no funky Painter like brushes and such), ScriptFu is still as limited as it ever was and creating complete Tools as PlugIns is still not possible, Layers are not structured, still no Macro Recorder and stuff like that that already was missing back then in the very first Gimp I touched. There is also advanced stuff like OpenGL integration that could help to speed up a few task a whole lot with todays graphics card, which you won't get any time soon in Gimp.
If somebody thinks that would be a heapload of work, he might be right, but on the other side, getting Gimp into a state where all the wished features would be possible could be even more work and after all one could always write a compability layer that allows to reuse Gimp plugins, which are after all the only thing in Gimp that is quite good.
Actually I think the reason for menu-bar at the top of the screen was introduced because the first Apple Computers had a very tiny screen, so screen-space was a real problem, a single menubar on top could help to save some space.
With todays rather large screen resolutions I find the menubar at top quite a bit confusing, since it can be quite a bit away from where the real 'action' happens. Especially with apps that don't pop-up a default window/document it can be quite problematic, you launch and app and nothing happens, until you find out that the tiny menubar at the other end of the screen has changed. I also don't really get it why the menubar is on top, while the toolbar isn't, both seem to be more or less the same thing, ie. toolbar simply groups often needed item from the menubar.
So yet another console generation passed without a Thornado release? I have been waiting for that game since the N64 days, kind of sad to see that it won't get a GameCube release.
Slashdot Mirror
The articel itself is standard Gnome propaganda again: "blabla, your file organisation is wrong, you are wrong, but our filemanager just must be right, blabla". Gnome has a whole lot of great ideas in it, but all to often Gnome maintainer seem to think they know better then the user what is good for them. Removing or hiding configurabilty in GConf makes this even worse.
Anyway, back to another issue in the article which I don't get at all, at the end Eugenia talks about Spartial Mode and DB-based filesystem with no folders at all and how they 'mix' well. Well, how do they mix? Isn't spartial actually the completly opposite of a DB-based filesystem? In spatial mode you represent folders as windows, one window represents exactly one folder, its size and position is safed, so that the window actually becomes the folder from a users point of view. With a DB-based filesystem however there wouldn't be folders, just data sorted in whatever way you need it at the point, so you would have lots and lots of different views onto your data, no windowfolder corelation which seems to be what spatial is all about.
OpenSource might not be up to commercial applications nor might it be dominant on the marked (how could it?, since its free...), I however still prefer my Linux system any day, about throwing out a few thousand Euros for what I would need to get the commercial equivalent.
Is it really the harddrive? Sure it might be the hardware part that could give the best boost in todays system, but how about fixing the boot issue at the software level? I mean boot speed hasn't been much of an issue big in the past, yet, today it suddenly is, even while harddrives and CPUs and RAM are much faster then a few years or even decades ago.
What went wrong on the software level that systems now take so long to boot? Are it generic kernels that probe for a lot of non existing hardware and thus basically just spend their time waiting for timeouts? Or is it todays software with all its dynamic linking and its dependecy (heard the slowness of KDE was in most part a fault of ld.so)? Or is it just bad organisation of boot scripts, ie. lack of paralelims while starting them, useless stuff first, so that the important stuff has to wait and such?
I can second that, the spiral tool seems to be rather useless and especially rather limited. Nothing wrong with having one, but it shouldn't be limited to spirals. A lot of other drawing programms have stuff like that too, but they don't only provide spirals, but also pseudo 3d blocks, arrows and other stuff.
About the missing layers support, I don't consider it much needed in a vector graphic application, yes it should be there and is a nice additional feature, but with grouping one can already get pretty similar things done.
However one very important feature currently missing is blending (ie. morphing one polygon shape into another), without it its almost impossible to get good looking shadows and highlights into vector graphics. Sketch already had blending for quite a while.
I would actually do it the other way around, skip the whole theory completly for the beginning and only come back to it at a much later point if there is still interest in programming. Sure theory is important if one wants to be a real programmer, but just for having a little fun with a programming language its not needed at all and theory isn't really thu fun part of it anyway for most people. Stuff like sorting, linked lists and such all come build-in with todays programming language, so not knowing about how they are implemented doesn't even hurt at all, since one can write a lot of usefull programms even without knowing the details.
... return x + y ...
Speaking about programming language I would start with Python, which I find a quite a bit easier to understand then Scheme from a beginners point of view (no need to declare variables, familiar infix syntax) and it also has an interactive interpreter like Scheme:
>>> a = 5
>>> b = 10
>>> a + b
15
>>> def add(x, y):
>>> add(a, b)
15
Above code should be easy enough to understand and should already be enough to have some fun.
The only things I miss in Python, which for example QBasic provided, is the IDE (not sure if there if there are any easy ones for Python) and more important a easy way to get access to the screen, in QBasic it was just:
SCREEN 9
LINE (10,10)-(100,100)
and you had a line on the screen, with Python I don't think there is anything that simple, most GUI packages always require a whole lot of initialisation, which is really not beginner friendly.
Sure with image formats such as mng or animated gif, things aren't so clear any more, anyway, point is that active content is evil most of the time and does things the user does not want. This is however not so much the problem of the active content itself, but more of the webbrowsers that often provide little or no help for the user to stop or control the active content, an external viewer does most of the time a lot better job here then some half-broken browser plugin and thats the major reason why some people prefer external viewers over plugin in stuff.
Well, yes, we are at a point at which we could make small computers, but the computers for the masses are still rather large. Not really because the users need bigger and faster computers (for a bit of Office computers a few years ago were already more then fast enough), reliability and silence would be far more important, but just because they still sell better, luckily thats slowly changing now.
Getting a tiny, noiseless computer that you can plug into your normal monitor, tv and USB keyboard are still far away from being the norm and in most cases either quite expensive or very hard to get.
The main driving force behind computer development the last 10-20 years where just making them faster, but not better or more userfriendly, luckily we are slowly reaching a point where computers are almost fast enough for almost everything, especially when it comes to normal use, so that the arguments for a new computer no langer can be just 'faster'.
Just consider the last 15-20 years, back then computer where very small, directly build into the keyboard, almost made no noise beside the drives, could be plugged into every TV and 'just worked', no need to maintain or reinstall the OS really, just insert another floppy and viola it worked.
Today computers are in most cases big noisy grey boxes. People have to reinstall or maintain their OS, manually install security patches and every once in a while a nice internet worms does funny things with your machine. GUIs are still pretty much the same as the first GUIs that apeared, just a bit more colorfull here and there, but fundamentally still pretty much the same.
Beside getting faster, bigger and louder and more throublesome in maintaince there really wasn't much change in the computers you normally use in the last 15-20 years, sure memory has grown bigger and now we can do some cool new stuff, but fundamentally not much has changed, some issues have even getting worse.
That said, technology might improve expotentially, user interfaces however do not. Just look at something basic like 'undo', neither Linux, Windows or MacOS provides an 'undo' on filesystem level, instead they provide this ugly workaround of a trashcan which doesn't help on a plain 'rm' or if you overwrite a file instead of deleting it. Versioned filesystems however are nothing new and should be easily doable with todays machines, however non OS does it that way, which just shows that hardware might be evolving fast, software however is not.
One of the major benefits of '10' is that the human has 10 fingers. And as a lot of people use their fingers to do everydays calculation, using a base 10 system helps a whole lot here making the calculations easier and feel more natural.
Well, a file manager is a pretty damn important part of a desktop environment, if that doesn't work and isn't easy to make work than for sure thats a big problem.
From my experince the removal of preferences or the move of them into GConf without any visiblity in the configuration dialog in Gnome2.x has been a whole LOT more confusing than just another setting in the configuration dialog.
If I don't know what it does, I don't touch it, if I messed things up I reset to default, very easy. With todays Gnome2.x I have to search google to first find out that a setting exists at all and then start browsing through gconf settings to tune them, thats nothing what I would call userfriendly.
In general I agree that config dialogs are often a bit overfull, but the way to pretty much remove everything that the Gonme developers are going is far worse.
The major problem here is that Gnome doesn't run in a Gnome-only environment. I use Gnome together with a whole bunch of non-Gnome apps, which in turn pretty much requirse to tweak a few Gnome-settings to run sane (color-scheme, panel behaviour, focus behaviour, etc.). Since Gnome2.0 a lot of settings necesarry for this have been gone into GConfi-only, no more GUI to tweak them, I wasted quite a few hours to get back to Gnome1.4 useablity since then.
Same with the spatial Nautilus, in a few-files/directory environment it might actually work quite good, after all on the Amiga it worked very well. But back then harddisks where small and storage were rare. Today I have a harddrive with deep directory trees and half millions of files, spatial doesn't work very well for that. Since Nautilus is also missing a useable smallicon/column view or tab-completion (Rox like for example), it makes it almost impossible to navigate through such an environment, since I have a lot to scroll, click windows away and such.
In general I really have the impression that Gnome is designed to be a cute desktop environment, that when run alone in a 'clean' environment might give a good experince, but which pretty much falls apart in most real world cases where people work with more than a handfull of files.
Am not the one asked, but lets try to answer anyway:
- ugly colorfull default icons, tuning for colorfullness, not for usefullness
- to much mimicing of Windows all over the place
- horizontal scrolling in konquor
- bad redraw behaviour in konqueror
- same stupid way to handle overfull icon panels as windows
- to crowded menus (Configure Shortcuts, Configure foo, Configure bar, Configure baz, Configure foo), makes it hard to find the right thing
- extreme long app startup times if one doesn't already run KDE
And here some screenshots which illustrate some of the problems:
http://pingus.seul.org/~grumbel/tmp/kdesucks/
Anyway, about your points I fully agree, Nautilus is a joke and configuration options are also a bit lacking. I however really like the default look, since it doesn't try to make itself 'unreadable' by using gradients all over the place.
I like the review. Is it flaimbait? Maybe. But is it true? I would say yes, he raises some very good issues that I know a lot of people have with Gnome. As with my day to day experince in using Gnome (I don't like KDE much, and a plain WindowMaker just isn't enough) I can fully agree with him.
Nautilus is in most part still a rather unpleasent piece of a file browser, it has been getting faster over the time, but thats it, its still missing some very basic features, smallicon-view being one, a simple button to go into the parent folder being another, but well, there are lots of more things that are just horrible broken in Nautilus (Mime-Handling and friends come to mind...). I really don't get why Gnome people got started with Nautilus in the first place, its after all just a very under-featured filemanager that was unusablily slow in its first few releases and only slowly got better. In the time they spend fixing it up, they might have at well written two other filemanager. And in its current stage I really doubt that it will reach a stage were I would consider it useable any time soon (maybe in gnome 3.0 or 4.0 or so, but it will take time...).
And well, the GConf issue, he is correct has well. Gnome Devolper have moved a lot stuff from Config Dialogs into obscure GConf settings once Gnome went from 1.4 to 2.x, what should that we good for? Basic settings that are needed to tweak Gnome into a useable stage are well hidden in some GConf settings now. I still prefer Gnome1.4 over Gnome2.x for exactly that reason. And well, this isn't only a Nautilus problem, it has tortured many Gnome apps, especially Galeon has suffered a lot, which again I consider the Gnome1.4 based builds a lot better then the Gnome2.x ones.
Anyway, as long as Gnome developers will just ignore the issues the users have I don't see much room for improvement in Gnome itself. They have multiple times in the past shown that they don't seem to care about who actually uses Gnome, but instead have optimized it for some 'hypotectical newbie user' which I have yet to meet in reality.
I get the feeling that this is really just a huge PR stunt so that Microsoft can say "Hey, we did something and catched the guy!", thus hiding the fact that the real problem is Microsoft that can't get their OS secure. Viruses and Virus writer are just the symptom, not the cause of this whole issue. If people wants to stop this happening again they need to start at Microsoft, not at some lonly virus writer somewhere out there.
Beside that WineX already does run a lot of games on Linux, yet, no breakthrough. While I agree that games are a very important factor in Linux, we are not at a point were they are really that important. As long as I still have to dive through weird config files to get my printer working and have to update or patch my Kernel for a more or less common piece of hardware (graphic tablet, 3d card), Linux won't go mainstream. Linux needs one standard and userfriendly way to handle and configure hardware, not one or two ways to do it for each distro. Linux has all the capabilities there, it just needs some more agreements on standards across distros.
Well, we might not have ToyStory graphics, but we are damn close to it. The problem however isn't graphics, but gameplay, camera movement and such. Until we get the graphics to ToyStory level it is just a matter of time, however there hasn't really been any major improvments in gameplay or camera movement over the last years, just little improvements here and there. And as long as you just have your little 3d puppet with a camera behind it and jump through some jump'n run levels, it will neither look nor feel like the ToyStory movie, no matter if the graphics itself match in polygon count and ligthing.
The problem is that even with story and acting up to Pixar's level you still wouldn't have a game, you would have a movie, nothing more. And how does one fit good gameplay into a heavily story/acting dependend game anyway? So far the only genre that did that reasonably well where adventure games, but well, they can life quite without using cutting edge realtime graphics and instead be based on prerendered stuff without to much problem. The games that need fast realtime graphics just too often end up being rather dumb fast action games, where story, if at all, just takes place in cutscenes just too often just slow down the gameplay instead of improving it.
While I agree that fun is really important and should always be number one goal, the story line can be extremly important as well. Sure with Mario type games its not really important, but with adventure games, some rpgs or other dialog-heavy games it can play a major rule, ruin the story line and you ruin the game, no matter if the gameplay itself is good.
Gimp has been around since 1996, which makes around 8 years of development, nothing that I would call a 'short time'. That said looking back at earlier Gimps the progress isn't really that impressive its still pretty much the same Gimp as back then, just with cleanup here and there. On the code side I think they have rewritten it almost completly now, but from the userinterface it has still many of the same faults as years ago.
If you read the article you will find out that the features Gimp provided where not the main problem, he gives a little chart which looks like:
Features 80%
Ease of Use 25%
Value for Money 10%
'Must Have' Factor 1%
And I can pretty much with that. Gimp provides heaploads of plugins for a lot of stuff, but usebility really reduces the overall picture by a huge amount. Things as lack of preview for example make a lot of plug-ins alot harder to use if not even completly impossible to use effectivly (ie. Plugin->Ok->Undo->Plugin... cycles where in other apps you get a realtime preview when you move the slider).
The problem is that Gimp doesn't need yet another little contribution, but a major redesign in quite a few points. Sure newer and cleaner dialogs are a good thing, but they don't remove this 'GUI is broken' issue that Gimp has when comparted to other Apps.
See MDI for example, every time it is mentioned in the mailinglist, IRC or whereever it ends up in huge flamewars, developers mentioning that it is useless and stuff like that. Getting something MDI like implemented in the current Gimp is something that I consider very difficult if not impossble, after all the issue is there for quite some years. Doing it in a fork would be quite easier, since it removes the whole time-wasting flamewar aspect of it.
I have though about that, but I would actually go another step further and start from scratch instead of basing stuff on the Gimp Code base. I haven't reviewed the codebase in depth, however looking at the not really existing progress over the last years something in the basics just has to be very wrong. I mean Gimp0.99 is pretty much the same as Gimp2.0, yes, it has a heapload more plug-ins on board, but thats basically it, the interface hasn't changed so much beside some minor cleanup and the basic tools are still the same (no funky Painter like brushes and such), ScriptFu is still as limited as it ever was and creating complete Tools as PlugIns is still not possible, Layers are not structured, still no Macro Recorder and stuff like that that already was missing back then in the very first Gimp I touched. There is also advanced stuff like OpenGL integration that could help to speed up a few task a whole lot with todays graphics card, which you won't get any time soon in Gimp.
If somebody thinks that would be a heapload of work, he might be right, but on the other side, getting Gimp into a state where all the wished features would be possible could be even more work and after all one could always write a compability layer that allows to reuse Gimp plugins, which are after all the only thing in Gimp that is quite good.
Actually I think the reason for menu-bar at the top of the screen was introduced because the first Apple Computers had a very tiny screen, so screen-space was a real problem, a single menubar on top could help to save some space.
With todays rather large screen resolutions I find the menubar at top quite a bit confusing, since it can be quite a bit away from where the real 'action' happens. Especially with apps that don't pop-up a default window/document it can be quite problematic, you launch and app and nothing happens, until you find out that the tiny menubar at the other end of the screen has changed. I also don't really get it why the menubar is on top, while the toolbar isn't, both seem to be more or less the same thing, ie. toolbar simply groups often needed item from the menubar.
So yet another console generation passed without a Thornado release? I have been waiting for that game since the N64 days, kind of sad to see that it won't get a GameCube release.