The worst that will happen is that your app won't look like the user's other apps
The worst that happens is that KDE and GNOME and Motif and etc apps don't work well together becaue of very different toolkist. Copy-n-Pate, Drag-n-Drop, IPC, etc all have very serious interoperation issues. Just because you can run a KDE app under Gnome or a Gnome app under KDE doesn't mean it is fully functional. The smallest issue of running non-KDE apps under KDE or non-Gnome apps under Gnome is the look-n-feel. The biggest issues are interoperabilty.
I don't see why one could not do what you are suggesting. It may be a lot of work though. It was built to around a proprietary ATL, so you would be basically be starting from scratch and just keeping the same interface.
Not really. Anyone can download Java freely, even though it may not be Free as in speech. Try to search google for "ATL download" and see what you get. I couldn't find anywhere to get the developer files needed to _develop_ with ATL. To the best of my knowledge, you need to buy MS Visual Studio to get the developer files needed to program with ATL. That is not the case with Tomcat, JBoss, etc, both of which I use and never had to buy a license from Sun.
and went with Linux on Apple Xserve's. I know I would fee much safer in a submarine running on more robust PPC hardware then x86. And just about any OS is more stable then NT 4. What were these people thinking? It had to be a big political stunt. Some MS goons lobby some Swedish politicians, and wham, you have yourself a new sub "powered by NT"(tm).
What? What closed source underpininings are you talking about? What part of Linux is closed source? And what closed source code/API's/etc does Abiword depend on?
I do agree with your statement, however, as I stated, WTL still needs the non-OSS ATL to function. As far as I know, Quake does not have any limitations like this.
This WTL may be OSS, however it depends on the ATL, which is not OSS. So again, this looks like a little marketing PR to me. MS making the WTL OSS while depending on the non-OSS ATL is Like having the OSS Linux kernel needing some non-OSS code from SCO. It doesn't make sense, accept for PR since many people will over look the important detail of depending on the non-OSS ATL.
Also, just like the last OSS release from MS, this is not what I would call an important piece of software. In MS's new world.Net is king. So the WTL can go away. However, it is pretty clever to get a little PR out of this release instead of just letting the WTL fade away.
All MS needs to do to make a big positive change of their image, is release ONE important piece of their software as OSS and actaully let the community contribute. I am not talking about a bread-n-butter app like MS Office. Some non-revenue generating app like IE or even explorer.exe. Those two apps could use some serious fixin.
The better approach would have been to make this new feature an _option_ and not the default. I think it is pretty dumb to make a major change to how one browses the file system and then make it the new default. Also, to get back to the old method is not very intuitive or easy to find, unless one spends their time following Gnome development. If the Gnome developers were adamant about making the new spatial mode the default, then the first time the user runs the new Nautilus, they should be prompted with detailed instructions on how to turn it off or an option to switch to the old method. A friend of mine uses Dreamweaver and he was showing it to me. The first time he ran it, he was prompted to pick the layout he wanted. One for programmers or one for designers. To me this is very user friendly and should have been the approach used by the Nautilus team.
Learn how to use it. Use Shift-doubleclick or middle-doubleclick to automatically close the parent. Alternatively, go wherever you want, then use "Close all parents" from the menu.
I thought Nautilus was going for being more intuitive? How is it intuitive if it requires someone to learn how to do or in this case _not_ to do something? There are no options for it in Nautilus. The new spatial feature requires more work then before to accoplish the same task.
and _I_ think, 15 levels is braindead
Hmm, here are some pretty long paths on my Linux desktop:/. lameness filter may break up the paths.
Yes, that is very intuitive right there. Most actions are handled by double-clicking in Gnome. After, looking through the opotions for Nautilus, I did not find anything that said to middle-click or use gconf-editor to change the behaviour, or what key to change.
And frankly, I consider 15 levels deep file hierarchies as braindead as you find nautilus
Well, I easily have more then 50,000 personal files. I guess I could just use ONE BIG FOLDER! Maybe it is brain dead to you, however I can find any of those 50,000+ files in 3 seconds or less based on a directory structure I use. I don't need to remember file names. I just need to have an idea of what I am looking for and I know where to go.
For example, personal code:
/home/foo/Documents/code/personal/$PROJECT_NAME/
Or maybe I want an email from some project I was working on for a company?
The number of people who have never used a computer before and that _will_ use one in the near future is extreamly small. Driving a car with a steering wheel may not be what "experts" call the most intuitve way of driving. However, anyone that drives or has been exposed to a car have learned that way and now consider it intuitive. Just like the _vast_ majority of computer users have been exposed to MS Windows and now consider the "MS Way" to be intuitive. The biggest hurdle I have in getting most of my fellow programmers to use Linux is that they all say it is not easy to use or is too different from what they have spent many years learning.
Basically a few Nautilus developers considered the spatial method to be more "intuitive" and forced it to end users of Gnome, dispite a lot of controversy on Gnome mailing lists. The smart thing would have been to leave Nautilus the way it was and make the new spatial method an _option_.
You can call it "spatial" or whatever you like. The fact is, is that when I double click on a folder I get a new window. When I need to get down to a subfolder that is about 15 levels deep, I have 15 windows open that clutter my desktop. I then need to move around/close 15 windows. I think this is brain dead. The Gnome guys should have left nautilus the way it was and made this new spatial option just that, an option. You don't go and make a major change like this without extremely positive user feedback, wich the new spatial mode did not have. Gnome needs to focus on leaving options in and not taking them out. Instead of stripping out options, the Gnome people should just have sane defaults.
Having a choice between different OSes is a good thing. However, within the OS one picks, there needs to be standards. I have been a long time user of Linux. However, my professional career is as a developer with 99% of all the work I do being for MS Windows. For Linux to get more user acceptence there _needs_ to be standards. Imagine if instead of TCP/IP, everyone used their own transmission protocol, or instead of HTTP, you needed a custom app for each web site? Right now, we have tons of very talented OSS developers constantly reinventing the wheel. KDE guys use mcop for IPC, Gnome uses orbit, different sound libraries, different everything to accoplish the same tasks. This moving target is not a good place to develop commercial software for.
Lets face reality, commercial software will never go away since it would be impossible to have all the software that the world needs to be developed in the spare time of OSS developers. So we have a problem of commercial companies not even looking at Linux on the desktop because it is such a moving target. What toolkit should they use? QT, GTK+, lesstif, motif, xlib? If they pick QT, will they lose sales to Gnoem users? None of these toolkits work well together. Drag-N-Drop, copy-n-paste, etc are all a pain when working with apps from different toolkits. This all leads to a non-integrated end-user experience. Integration is very important in a desktop OS. Yes, IMO you can over-integrate such as MS has done, however there needs to be _some_ integration. For example, I don't like Konq as a web browser and use Firebird. Firebird uses GTK+ and because of that, I lose functionality if I use it with KDE with simple tasks like copy-n-paste. I like K3B, but it just doesn't fit under my Gnome desktop. This is similar to how Mac OS X users don't care for running X11 apps under Mac OS X and prefer native OS X apps.
I personally think a community approach is the best way to develop applications. However, I think a corporate/organization/commercial approach is the best way to deliver an OS. A community approach for a full OS is too fragmented, where as a corporate/organization/commercial approach could make sound choices as the final product from a great pool of community developed applications. I hope that Novell can deliver something like this within a few releases. Their first release seems to pretty much be a bundling of everything similar to Red Hat and mandrake, with little focus on the final product. This is not a winning combination for desktop Linux IMO.
How is it intuitive to a new user of Linux. There is a very strong possiblity that that new user has used MS Windows. MS dropped the open each folder in a new windows by default behaviour a long time ago. I would guess that the reason MS dropped that behaviour by default is from spending money on focus groups with the end result being that users do not like it.
I have been using Gnome since 1.4 and really liked it until around 2.2 or whenever they made that stupid choice to switch the OK/Cancel buttons around. Howver, I got used to that and was looking forward to 2.6. After using 2.6, I don't see any big changes to write home about. I think the Nautilus change is brain dead. I have tens of thousands of personal files, music, home video, source code, etc, etc. all organized in a nice tree structure to help me find things. With the new Nautilus, getting to some files could cause 15 stinken windows to open. It is a royal pain to have to do extra steps to go and close each old window when a new one opens. I use the command line extensively, however, for moving files around, a GUI can be much better. CTL+Click to select non-sequential files instead of typing out all the names. Oh, and I have a good computer, so I want to see thumbnails of images and video, you can't get that on a command line. For the first time in a few years, I am giving KDE a serious try to see if I can switch from Gnome.
I find that normal business users don't know how to do much of the fancy stuff in Excel or any of their other programs. The most common usage of Excel I've found is glorified forms
I have to agree 100% here. I have worked for 3 fortune 500 companies and 99.5% of all MS Office docs that go around are very simple and OpenOffice.org can handle those needs with no problems. The only users I have seen using more complex features are financial analyst who had to be trained in Excel, so they can be trained in OOo. OOo is ready now to displace MS Office in the workplace. However, it needs to be a corporate wide choice. A single user cannot start asking everyone to send them docs in OOo, they will be laughed at. Now if the whole company converts, then there is a lot of weigth to go with that choice. Any other company that wants to do business will have to send in an open format such as OOo, CSV, HTML, PDF, etc. The hardest part of the switch is not OOo, but getting upper management to become "un-brain-washed" by the MS Sales guys of how MS Office will "save them money", make them "more productive" and help them to one day achieve the dream of a "paperless office".
MS's XML format is more of a PR stunt then really being open. MS has barked a million times about "IP" and MS Office is one of their biggest cash cows. Basically they made a schema that will let you read the MS Office docs, but they still keep tons of closed proprietary stuff in those XML files. What is the purpose of being able to read the file if the important content is a binary blob in some proprietary format? The plain text is readable, so a simple Word doc is easy to read (though competing office apps have been able to do that for a long time). MS Office will truly be open when MS release full specs of the file format and all that could possibly be in them. I can give you an XML file with a Base64 encoded blob of proprietary data. Just because it is XML does not make it Open. OpenOffice's format is _really_ open. You can get docs that explain the format and how to read or write OOo's file formats. This is not the case for MS. If it is, please provide a link to the MS Office document _specs_ and not just some silly schema.
As a little test, create a new Excel file and on Sheet 2 put the following data:
1 1 1 2 1 4 1 4
Now on Sheet1, insert a chart using the data on Sheet2. Now try to save it as "XML SpreadSheet (*.xml)". You will get a warning that all "AutoShapes, other objects and Charts" will be removed. What is the point of this "open" XML format if it cannot save complex spreadsheets? MS will never let their MS Office format go. End of story.
Did Sun only release a version of Java for Solaris? When Java first came around, Unix was the #1 server OS with Sun's Solaris having the #1 spot out of the differnt versions of Unix. Sun could have only released Java for Solaris to try to gain an edge in the server OS market, yet they did not and made it Open, while not Free (as in speech), Open is still a very good thing.
Also, there is far more to.Net then the C# language and CLI. MS has not submitted the framework wich is the "bread-n-butter" that.Net developers will be using. Without the.Net framework for another platform,.Net is a single platform technology and locks you into an MS only solution.
One improvement I (totally subjective) noticed with.net was speed - ASP.net apps seemed a lot "snappier" than JSP/servlet apps.
We ran into this same issue. However, we were comparing Servlet/JSP/J2EE running on Slowaris with dog slow Sparc processors, while.Net ran on dual 3GHz Xeons with HT, 2GB Ram, SCSI, etc using Windows 2003. We switched Servlet/JSP/J2EE to the same piece of hardware using Red Hat Enterprise Linux 3.0 and Oracle 9iAS as the J2EE server and the numbers were about equal across the board, with the exception of Linux/Oracle 9iAS scaling to more concurrent users then.Net. Oracle 9iAS was a little bit slower then.Net until we turned on Oracle 9iAS's Web Cache which made a nice difference and then the two were about equal.
That will only handle ASP.Net web apps and not any GUI, console, etc type applications. As other have pointed out there is a GPLed app called SharpDevelop for the GUI side, though you now need two IDE's to work with the same framework unless you want to pay MS for VS.Net.
For people with *nix development experience like myself, an environment like the.Net SDK are fine. However, MS built a huge industry of low quality VB coders that hit a wall when they cannot point-n-click. They don't care about understanding the code, abstraction, OO, etc. The.Net SDK is not much use to your average MS developer.
Do you have a link for the Shockwave plugin? Or are you thinking of the Flash plugin? I know that Macromedia has a _Flash_ plugin, but I never found a _Shockwave_ plugin. You can use cross over plugin which will use the MS Windows version of Shockwave, though that is not the same as having a native Linux version.
Some of the problems with IE "only" sites are they they are created by dummies that cannot handle making a simple HTML page with a little standards comliant JavaScrit and use JavaScript like document.all which is non-standard crap MS uses. A quick change of document.all to document.getElementById is often all that is needed.
I doubt desktop Linux will happen that way. Linux on the server yes, it is growing extremely fast. However on the desktop, I don't think there will ever be the "Year of desktop Linux". It will be a gradual adoption. More users will try it and for some it will be ready and for other it will not. A few months later the same thing will happen. Lather, rinse, repeat. The main issue for desktop Linux is commercial software support. That will be a slow change as more and more companies dip their feet in the water and try it out. Again, I think Linux on the desktop will see a slow gradual acceptance while Linux on the server is already strong now and growing very fast.
Slashdot Mirror
I don't see why one could not do what you are suggesting. It may be a lot of work though. It was built to around a proprietary ATL, so you would be basically be starting from scratch and just keeping the same interface.
Not really. Anyone can download Java freely, even though it may not be Free as in speech. Try to search google for "ATL download" and see what you get. I couldn't find anywhere to get the developer files needed to _develop_ with ATL. To the best of my knowledge, you need to buy MS Visual Studio to get the developer files needed to program with ATL. That is not the case with Tomcat, JBoss, etc, both of which I use and never had to buy a license from Sun.
and went with Linux on Apple Xserve's. I know I would fee much safer in a submarine running on more robust PPC hardware then x86. And just about any OS is more stable then NT 4. What were these people thinking? It had to be a big political stunt. Some MS goons lobby some Swedish politicians, and wham, you have yourself a new sub "powered by NT"(tm).
What? What closed source underpininings are you talking about? What part of Linux is closed source? And what closed source code/API's/etc does Abiword depend on?
I do agree with your statement, however, as I stated, WTL still needs the non-OSS ATL to function. As far as I know, Quake does not have any limitations like this.
Also, just like the last OSS release from MS, this is not what I would call an important piece of software. In MS's new world .Net is king. So the WTL can go away. However, it is pretty clever to get a little PR out of this release instead of just letting the WTL fade away.
All MS needs to do to make a big positive change of their image, is release ONE important piece of their software as OSS and actaully let the community contribute. I am not talking about a bread-n-butter app like MS Office. Some non-revenue generating app like IE or even explorer.exe. Those two apps could use some serious fixin.
The better approach would have been to make this new feature an _option_ and not the default. I think it is pretty dumb to make a major change to how one browses the file system and then make it the new default. Also, to get back to the old method is not very intuitive or easy to find, unless one spends their time following Gnome development. If the Gnome developers were adamant about making the new spatial mode the default, then the first time the user runs the new Nautilus, they should be prompted with detailed instructions on how to turn it off or an option to switch to the old method. A friend of mine uses Dreamweaver and he was showing it to me. The first time he ran it, he was prompted to pick the layout he wanted. One for programmers or one for designers. To me this is very user friendly and should have been the approach used by the Nautilus team.
Do you have any experience with complex enterprise class apps? Here are a few paths from Oracle9iAs
Basically a few Nautilus developers considered the spatial method to be more "intuitive" and forced it to end users of Gnome, dispite a lot of controversy on Gnome mailing lists. The smart thing would have been to leave Nautilus the way it was and make the new spatial method an _option_.
You can call it "spatial" or whatever you like. The fact is, is that when I double click on a folder I get a new window. When I need to get down to a subfolder that is about 15 levels deep, I have 15 windows open that clutter my desktop. I then need to move around/close 15 windows. I think this is brain dead. The Gnome guys should have left nautilus the way it was and made this new spatial option just that, an option. You don't go and make a major change like this without extremely positive user feedback, wich the new spatial mode did not have. Gnome needs to focus on leaving options in and not taking them out. Instead of stripping out options, the Gnome people should just have sane defaults.
Lets face reality, commercial software will never go away since it would be impossible to have all the software that the world needs to be developed in the spare time of OSS developers. So we have a problem of commercial companies not even looking at Linux on the desktop because it is such a moving target. What toolkit should they use? QT, GTK+, lesstif, motif, xlib? If they pick QT, will they lose sales to Gnoem users? None of these toolkits work well together. Drag-N-Drop, copy-n-paste, etc are all a pain when working with apps from different toolkits. This all leads to a non-integrated end-user experience. Integration is very important in a desktop OS. Yes, IMO you can over-integrate such as MS has done, however there needs to be _some_ integration. For example, I don't like Konq as a web browser and use Firebird. Firebird uses GTK+ and because of that, I lose functionality if I use it with KDE with simple tasks like copy-n-paste. I like K3B, but it just doesn't fit under my Gnome desktop. This is similar to how Mac OS X users don't care for running X11 apps under Mac OS X and prefer native OS X apps.
I personally think a community approach is the best way to develop applications. However, I think a corporate/organization/commercial approach is the best way to deliver an OS. A community approach for a full OS is too fragmented, where as a corporate/organization/commercial approach could make sound choices as the final product from a great pool of community developed applications. I hope that Novell can deliver something like this within a few releases. Their first release seems to pretty much be a bundling of everything similar to Red Hat and mandrake, with little focus on the final product. This is not a winning combination for desktop Linux IMO.
I have been using Gnome since 1.4 and really liked it until around 2.2 or whenever they made that stupid choice to switch the OK/Cancel buttons around. Howver, I got used to that and was looking forward to 2.6. After using 2.6, I don't see any big changes to write home about. I think the Nautilus change is brain dead. I have tens of thousands of personal files, music, home video, source code, etc, etc. all organized in a nice tree structure to help me find things. With the new Nautilus, getting to some files could cause 15 stinken windows to open. It is a royal pain to have to do extra steps to go and close each old window when a new one opens. I use the command line extensively, however, for moving files around, a GUI can be much better. CTL+Click to select non-sequential files instead of typing out all the names. Oh, and I have a good computer, so I want to see thumbnails of images and video, you can't get that on a command line. For the first time in a few years, I am giving KDE a serious try to see if I can switch from Gnome.
As a little test, create a new Excel file and on Sheet 2 put the following data:
Now on Sheet1, insert a chart using the data on Sheet2. Now try to save it as "XML SpreadSheet (*.xml)". You will get a warning that all "AutoShapes, other objects and Charts" will be removed. What is the point of this "open" XML format if it cannot save complex spreadsheets? MS will never let their MS Office format go. End of story.Also, there is far more to .Net then the C# language and CLI. MS has not submitted the framework wich is the "bread-n-butter" that .Net developers will be using. Without the .Net framework for another platform, .Net is a single platform technology and locks you into an MS only solution.
One other note: This is actaully a port of the GPLed SharpDevelop which works with MS .Net. So for .Net use SharpDevelop and for Mono use MonoDevelop
There is MonoDevelop, while it is still in early development, it is coming along very well.
That will only handle ASP.Net web apps and not any GUI, console, etc type applications. As other have pointed out there is a GPLed app called SharpDevelop for the GUI side, though you now need two IDE's to work with the same framework unless you want to pay MS for VS.Net.
There is a free ASP.NET Web Development Tool. Though it is only good for ASP.Net and not GUI development with .Net.
Some of the problems with IE "only" sites are they they are created by dummies that cannot handle making a simple HTML page with a little standards comliant JavaScrit and use JavaScript like document.all which is non-standard crap MS uses. A quick change of document.all to document.getElementById is often all that is needed.
I doubt desktop Linux will happen that way. Linux on the server yes, it is growing extremely fast. However on the desktop, I don't think there will ever be the "Year of desktop Linux". It will be a gradual adoption. More users will try it and for some it will be ready and for other it will not. A few months later the same thing will happen. Lather, rinse, repeat. The main issue for desktop Linux is commercial software support. That will be a slow change as more and more companies dip their feet in the water and try it out. Again, I think Linux on the desktop will see a slow gradual acceptance while Linux on the server is already strong now and growing very fast.