I'm sorry, but I don't see your point. Are you saying that, as Gnome does what you need it to do, choice ain't better? But then you say "it comes to what you need it for"?
How is it having choice worse than no choice at all? That defies common sense, buddy. Choice is good, and Gnome has all the choices that KDE has. The difference is presentment of choices. KDE presents to you every choice you could ever want, while Gnome presents you with the choices they think you want, and hide the rest in gconf. It's really a matter of opinion which you like better, some people want to tell the waiter to cook their burger medium/medium-well with ketchup but not mustard, a single leaf of lettuce, no tomato, two pickles and a sesame seed bun. Others just want a damned burger.
In the end, I want a desktop environment that I don't have to use. I have applications that I use, my desktop is just a means of accessing them. After that, I don't want to be bothered by it. That's why I use Gnome.
The Register you will note that the winner said that he would have been able to pull this off on any of the operating systems. Ubuntu (nor OS/X) is in no way immune to this attack. Now, how did he pull it off? Because Adobe/Macromedia in their wisdom decided they needed escalated privileges (I really don't know for what reason) for some tasks. Because the plugin cannot break out by itself they designed a "broker process" which runs as the currently logged on user. This process talks to the browser plugin and performs privileged tasks on behalf of the plugin. Does the linux version of Flash require such a "broker process"? Since, as you mentioned earlier, Firefox on Linux already runs with full user permissions, perhaps there is no broker process to exploit. Also, it may be that the exploit used the broker process to pass unsanitized commands to the Win32 layer, where the actual vulnerability existed, in which case the hacker would have to find a comparable vulnerability somewhere else in Linux where a user process can gain extra privileges.
Shouldn't it be default that something written in Java runs on ALL platforms which got a JRE? I'm pretty sure that AIR is C/C++. They're probably talking about the Java Plugin being supported inside an AIR application. I think AIR uses WebKit, so it's probably just that Sun has a WebKit plugin, and GNU doesn't.
From what I recall, most of the memory savings are coming from preventing fragmentation and being more conservative in their caching. While there were memory leaks that got fixed, they don't make up any large share of the new memory savings.
Some of the 2.x plugins have been upgraded to work with 3.0, others work fine they just need the version compatibility bumped up to 3.0. Currently I'm running:
AdBlock Plus Aging Tabs Download Statusbar Firebug (disabled when not in use, otherwise it is a memory hog) Fission Flashblock IE Tab (Stupid intranet sites) Java Quickstarter (huh, I didn't install that) Web Developer Toolbar
The only one not working is LiveClick for more usable RSS bookmarks.
It's been running all day, with between 1 and ~20 tabs open, and is currently sitting at 65MB memory with 5 tabs open. Really, the memory improvements in FF3 are really very impressive. It'd have been running longer, but I had to reboot Windows, it started forgetting to drawn windows when they would popup, or I restored minimized ones, go figure.
Personally I'll wait to see if Firefox3 doesn't consume 700MB of memory after being open for three days before I think about buying into the idea that the Mozilla foundation should be pushing any envelopes. I've been using Beta 3 and now 4 on windows at work(*) for several weeks now, and after a leaving it open for a week it will usually consume ~75MB of memory. I have seen it peak at about 130MB when in heavy use. But nowhere near the 300+MB I would get doing the same with FF2.
(*) I have beta 3 on Ubuntu at home, but I've never checked the memory usage because even though I run the same apps at home, I've never had any memory problems on Ubuntu.
Gimp doesn't group windows in a way that is familiar to windows users. If I'm using photoshop, there's a main container window then the smaller windows. If I alt+tab between applications, all photoshop windows (pallet, history, open images, etc) are shown or hidden as a group. Your window manager can probably do the same thing for you, as tracking window state is really it's responsibility, not the application's. Heck, you can probably even attach non-Gimp windows to the same window group, and it will be minimized/restored, even moved, when any of the other windows are. I know I can do this with Compiz anyway.
I use Evolution every day, it works fine for me. I've even used it with Exchange which worked surprisingly well, but having to use the OutlookWebAccess was a bit slow. I hear Novell has done some work to improve on that, but I haven't that the misfortune of having to use Exchange since then.
I am by no means a GIT expert, but my understanding is that in GIT, every "revision" calculated based on every revision that came before it (GIT versions look more like an MD5 hash than a 1.1.1 type version number), so I would assume that merging your local repository with someone else's would show each individual commit you made to yours in the other person's history after the merge.
Use something like GIT or Mercurial, where users can edit/save in a disconnected environment, then merge their changes back with the centralized repository when they get back, or merge their changes with another user's local repository (You can't do that with Word's change tracking as far as I know).
A GIT/Mercurial OO.o plugin could save the revision data directly in the META-INF folder of the ODF file itself every time you save, giving you the best of both worlds. Doing it this way would even give you separate revision histories for content, styles and meta-data. Suppose you only want to accept style changes but not content changes (or vice-versa), this could do that. The majority of the work would be developing an XML-based diff and displaying it inline with the document.
Not using repositories makes life much easier. Software can be wrapped with all the dependencies in a container and be deployed, installed, removed, upgraded as needed. More or less like a sandbox environment. I've installed many programs on Ubuntu as downloaded.DEB files, which seems more inline with what you want. Doing so still lets you maintain the software in your local database, and allows you to download and install dependencies from your repositories if required, and adds menu entries in your Gnome/KDE/XFCE desktop menu. I prefer to use a repository so I get automatic updates and the like, but it's not strictly necessary if you don't like it.
As a counter-example, the official way to get Firefox 2 on Ubuntu 6.0 LTS is to upgrade the WHOLE OS to the next release. I have IE6 on Windows2000, how do I get IE7? However, unlike IE on Windows, if you really want FF2 on Ubuntu 6.06, you can just download it and install it directly, nothing stops you from doing that.
Ubuntu is not even backwards compatible with itself, as most packages can't run on the previous release, and the cycle is only 6 months long! You call that stability? Um, when a package takes advantage of the new libraries and such that come with a new release, then of course you can't run that on an older release. That's not a Linux thing, it's true of any OS. If a program relies on new features/libraries in Vista, it won't run on WindowsXP. Your complaint is that new versions of a program take advantage of new features in the OS, but the alternative is for your programs to _not_ take advantage of new features when they become available, and that would make me (and many others) very unhappy.
and if the release is stable, the repos are frozen - good luck with those new features you need. The release is stable _because_ the repos are frozen. If you want the latest software, then upgrade to the latest software, but then you won't be guaranteed a stable system, you can't eat your cake and have it too. Nothing prevents you from upgrading, either your entire OS or just the apps/libraries that you want to upgrade, it's just that Canonical isn't going to put bleeding-edge code in their stable branch.
By using third party repositories and risking a breakage of the system Again, this is true for any OS, when you introduce unknown variables you risk breakage. Heck, it's not even an OS thing, if you add non-standard parts to your car, you risk breakage.
Repositories are, IMHO, just a lame excuse lazy developers have found for not doing their jobs till the end and compile their own fucking software. You are, of course, entitled to your opinions. But they are wrong. A developer's job is to make a program, not to ensure that the user's environment has all the necessary libraries for it to run, or to ensure that it gets added to the desktop's menu system, or to provide a UI for installation/deletion of the program. Programmers should not be responsible for enforcing the Operating System's conventions, that is the OS developer's job.
Hmm. I thikn I see what you mean, but I disagree. If I look there one minute and see a series "B I U" buttons, and the next minute I see a "add merge del" - even if I couldn't make use of "B I U" in that context - it's still going to be confusing. The reason is partly because you're putting different things in the same place; but also because it would not be internally consistent. For example if I selected the table, I would see just the table buttons; if I was editing text, I would see the formatting buttons What would be even better would be to add background-highlights on the toolbars that OO.o thinks you'd be most interested in depending on your cursor location. In your example, editing text in a table, the "Text" toolbar would be most highlighted, the "Table" toolbar slightly less so, the "Page" toolbar less agian, and the "Picture" toolbar would not be highlighted at all (or perhaps anti-highlighted/dimmed to indicate it serves no purpose in your current context). Visually it would be a lot like the "Aging Tabs" extension to Firefox.
That way it makes it easier to find the toolbar buttons you'd be interested in without them getting lost in all the rest, and at the same time keeping them all in their original locations so that once you've memorized where a button is you don't have to hunt for it again when in a different context. If anyone else thinks this is a good idea, I'll post it to OO.o.
Gimp handles GIF/JPEG/PNG files just fine, and can open most PSD files.
You can run IE 5.5, 6 and 7 simultaneously on Linux+Wine. As a fellow professional web developer, I test in IE6, IE7, FF2 (and now FF3b4), Opera and KHTML/Webkit.
* if there is a way to install without booting to the live environment, it didn't jump out at me when I put the CD in. The Alternative CD will do that.
The first issue I encountered with Ubuntu was the fact that it wouldn't allow me to select a sensible resolution without editing xorg.conf - despite the fact that my monitor and video card was identified correctly by name. The next issue I came across happened minutes later and was even more serious - the title bars and borders of all windows disappeared so you could not move or resize them. With my nVidia GeForce3, Ubuntu was able to auto-detect a fairly extensive list of available resolutions without my having to edit xorg.conf. I suspect that your second issue is somehow related to your first, in that after manually editing your xorg.conf, you were missing some parameters for your video card that caused compiz to crash (such as "AddARGBGLXVisuals", leaving you without a window manager. Compiz can use metacity is a failsafe window manager if it crashes, I'm not sure if that's default or not. If this should happen to you again, go to a terminal and run "metacity --replace &" to get them back.
I need to use to a lot of Windows software and WINE just doesn't cut it, not to mention (as I said elsewhere in this thread) the fact that hacks like WINE are the exact type of unnecessary tinkering people doing real work don't need. If you want to run Windows software without issue*, use Windows. Not all of us need Windows software, I get along just fine without it. I use Ubuntu at home because it works for me, and yes it is fun. I have used Ubuntu at work in the past because it worked far better than Windows for the things I was doing, and I still had access to all the company resources that the Windows users did, only the client programs I used ran _better_ than theirs.
(* By "without issue" of course I mean "like it runs in windows", because we all know that windows apps don't always run "without issue", even on Windows.)
After some more digging, it appears that you must implement "at least" 2 of the 5 "features". I put "features" in quotes because it includes displaying the AIM Start Page, installing the AIM Toolbar, or including AOL advertisements. So it's really more of a "You can use OpenAIM if you do something to make us money".
To be fair to AOL, your 2 features can just be showing Buddy Info and Buddy Icons, which presumably don't make them money, and most AIM clients would want to have anyway. Still, this whole requirements seems odd to me.
From the OpenAIM website:
Development of AIM-Enabled, Multi-IM Protocol Clients
* AOL now allows multiheaded clients to access the AIM network But here is the part that gets me:
Flexible Value-added Application Feature Requirements
* Developers can pick 2 of 5 features to implement from a list of the most popular and valuable AIM features Does this mean that applications using OpenAIM can only use 2 of the 5 features offered by OpenAIM?
Slashdot Mirror
How is it having choice worse than no choice at all? That defies common sense, buddy. Choice is good, and Gnome has all the choices that KDE has. The difference is presentment of choices. KDE presents to you every choice you could ever want, while Gnome presents you with the choices they think you want, and hide the rest in gconf. It's really a matter of opinion which you like better, some people want to tell the waiter to cook their burger medium/medium-well with ketchup but not mustard, a single leaf of lettuce, no tomato, two pickles and a sesame seed bun. Others just want a damned burger.
In the end, I want a desktop environment that I don't have to use. I have applications that I use, my desktop is just a means of accessing them. After that, I don't want to be bothered by it. That's why I use Gnome.
It looks like the standard XFCE panel.
From what I recall, most of the memory savings are coming from preventing fragmentation and being more conservative in their caching. While there were memory leaks that got fixed, they don't make up any large share of the new memory savings.
Some of the 2.x plugins have been upgraded to work with 3.0, others work fine they just need the version compatibility bumped up to 3.0. Currently I'm running:
AdBlock Plus
Aging Tabs
Download Statusbar
Firebug (disabled when not in use, otherwise it is a memory hog)
Fission
Flashblock
IE Tab (Stupid intranet sites)
Java Quickstarter (huh, I didn't install that)
Web Developer Toolbar
The only one not working is LiveClick for more usable RSS bookmarks.
It's been running all day, with between 1 and ~20 tabs open, and is currently sitting at 65MB memory with 5 tabs open. Really, the memory improvements in FF3 are really very impressive. It'd have been running longer, but I had to reboot Windows, it started forgetting to drawn windows when they would popup, or I restored minimized ones, go figure.
(*) I have beta 3 on Ubuntu at home, but I've never checked the memory usage because even though I run the same apps at home, I've never had any memory problems on Ubuntu.
I use Evolution every day, it works fine for me. I've even used it with Exchange which worked surprisingly well, but having to use the OutlookWebAccess was a bit slow. I hear Novell has done some work to improve on that, but I haven't that the misfortune of having to use Exchange since then.
Which is essentially what the GP was saying, you can't edit a PDF as a text document in Gimp, you can only edit it as an image.
I am by no means a GIT expert, but my understanding is that in GIT, every "revision" calculated based on every revision that came before it (GIT versions look more like an MD5 hash than a 1.1.1 type version number), so I would assume that merging your local repository with someone else's would show each individual commit you made to yours in the other person's history after the merge.
Unless of course you're referring to a metric fuckton.
Use something like GIT or Mercurial, where users can edit/save in a disconnected environment, then merge their changes back with the centralized repository when they get back, or merge their changes with another user's local repository (You can't do that with Word's change tracking as far as I know).
A GIT/Mercurial OO.o plugin could save the revision data directly in the META-INF folder of the ODF file itself every time you save, giving you the best of both worlds. Doing it this way would even give you separate revision histories for content, styles and meta-data. Suppose you only want to accept style changes but not content changes (or vice-versa), this could do that. The majority of the work would be developing an XML-based diff and displaying it inline with the document.
Now, about that job....
That way it makes it easier to find the toolbar buttons you'd be interested in without them getting lost in all the rest, and at the same time keeping them all in their original locations so that once you've memorized where a button is you don't have to hunt for it again when in a different context. If anyone else thinks this is a good idea, I'll post it to OO.o.
Gimp handles GIF/JPEG/PNG files just fine, and can open most PSD files.
You can run IE 5.5, 6 and 7 simultaneously on Linux+Wine. As a fellow professional web developer, I test in IE6, IE7, FF2 (and now FF3b4), Opera and KHTML/Webkit.
Linux is ideal for web development.
(* By "without issue" of course I mean "like it runs in windows", because we all know that windows apps don't always run "without issue", even on Windows.)
That and the fact that you have trains.
So it all finally comes back to Java then?
The MP3 format is patented, the LAME encoder is copyrighted. Even after the MP3 patent expires, LAME will still be copyrighted.
If I still had mod points today, they would be yours.
After some more digging, it appears that you must implement "at least" 2 of the 5 "features". I put "features" in quotes because it includes displaying the AIM Start Page, installing the AIM Toolbar, or including AOL advertisements. So it's really more of a "You can use OpenAIM if you do something to make us money".
To be fair to AOL, your 2 features can just be showing Buddy Info and Buddy Icons, which presumably don't make them money, and most AIM clients would want to have anyway. Still, this whole requirements seems odd to me.
* AOL now allows multiheaded clients to access the AIM network But here is the part that gets me: Flexible Value-added Application Feature Requirements
* Developers can pick 2 of 5 features to implement from a list of the most popular and valuable AIM features Does this mean that applications using OpenAIM can only use 2 of the 5 features offered by OpenAIM?