Red Hat plans to ship the next version of its premium Linux product on February 28, debuting major virtualization technology but missing an earlier deadline by about two months. Red Hat Enterprise Linux 5 had been scheduled to ship by the end of 2006. However, the company gave itself scheduling wiggle room in September, when Red Hat released the first RHEL 5 beta; a second beta arrived in November.
Better to take your time and do the job right the first time than do it like Microsoft does and release half baked code.
Dave
As much as I dislike MS in general, I have to admit that they’ve been very good about having solid releases of their Server platform. Server 2003 was arguably the single best operating system they’ve ever produced.
Actually, that’s an exceptionally good point. That deserves a good mod point.
Dave
Too bad MS’s best OS costs so much. Last I checked it started at over $700 per license and went up from there depending on what your doing.
Better to take your time and do the job right the first time
Agree per 100%. Too me, it is much more important that a product is stable, mature and well thought out than that it comes on time.
than do it like Microsoft does and release half baked code.
I am not a MS fan at all, neither a mac fan, nor a radical FLOSS advocat (:)), but my impression is, that Microsoft are working hard to continously improve the quality and stability of their products and more and more give stability and realiabilty precedence to a short release cycle.
Examples:
1) Vista is more than two years late and Microsoft always proposed “Vista comes when it is finished”
2) XP and 2000 are much better OSes than 98, 95, DOS when it comes to stability, realiability and security. XP SP2 has a very good firewall included which is much better than most of the third party firewalls you can get on the market.
3) They do research on the fascinating Singularity microkernel project. As Prof. Andrew S. Tanenbaum of the Vrije Universiteit, Amsterdam states in his article “Can We Make Operating Systems Reliable and Secure?” (http://www.cs.biu.ac.il/~wiseman/2os/microkernels/tanenbaum.pdf):
The most radical approach comes from an unexpected source Microsoft Research. In effect, the Microsoft
approach discards the concept of an operating system as a single program running in kernel mode plus
some collection of user processes running in user mode, and replaces it with a system written in new typesafe languages that do not have all the pointer and other problems associated with C and C++.
4) There are more.
FYI.. shots are available at LinuxQuestions http://shots.linuxquestions.org/index.php?linux_distribution_sm=Red…
Looks like FC6.
The desktop will most likely not be the primary target for this release but it’s still a pity that they didn’t wait for the next Gnome version that will contain some of the usability enhancements that was introduced by Novell. I’m thinking of things like the new “slab” gnome menu. Not that the current Gnome menu is that bad, but nothing beats things that have been tested on real non geek users.
Incorporating something that isn’t even tested on Fedora would be a great risk for them.
RedHat needs to support their major server releases for at least seven years. Such a move could bring unnecessary support issues, which they smartly avoid.
True, RHEL5 will have a long lifespan, that makes it even more urgent to make use of the latest research in usability. There are too many pointy haired bosses out there that will judge even a server OS from its looks.
Besides, a server needs to be managed, and the people doing that would benefit from a better interface. You are right that, not testing things like the slab in Fedora poses a risk, but then again, who’s fault is it that. Novell managed to include it in their distro long before FC6 was out.
As for the “slab” menu, we are talking about a Gnome applet, i.e. it is a very limited amount of code. It should be quite possible to test it in a relatively short time. If they can’t do that, I would say that there are something seriously wrong with their testing procedures.
Well, my experience is a little bit different. I’ve seen several small firms and mid range system rooms.
Normally servers are not operated directly. They usually sit in an isolated room, and probably tens of them are connected to a single console switch.
The server operators, and administrators access via SSH, and any X11 application is also tunneled in such a connection.
The only time you use the actual console is for installing a new operating system or doing low level maintenance.
Even in those situations, there are remote access cards on servers that allow full remote operation (access to BIOS, RAID cards, OS install, and full remote VGA cloning).
This is also true Windows 2003 servers. They are usually operated through Terminal services.
So a new “start menu” which will probably cause destabilization is not on potential customers’ wish list. (Any organization willing to spend thousands of dollars on RedHat will probably employ competent system administrators too).
Edited 2006-12-29 11:44
… a pity that they didn’t wait for the next Gnome version that will contain some of the usability enhancements that was introduced by Novell. I’m thinking of things like the new “slab” gnome menu.
Novell’s Gnome menu is not more usable than the present default Gnome menu.
The former actually needs more mouse clicks.
Plus, why sacrifice one window thumbnail/virtual desktop bar? Because virtual desktops are confusing; however it’s nice once you know it. Why not explain the nice ‘secret’ features of Gnome in a KDE style “tip of the day” way?
I may be ostracised for saying this, but if anything, Novell’s Gnome menu is an attempt to make Gnome look more like Windows, which is not by definition more usable. It may be good for people that get scared by anything remotely different from Windows, though, if you want to have them use another OS.
Not that the current Gnome menu is that bad, but nothing beats things that have been tested on real non geek users.
Ergonomists know, for starters, that a menu on top of the screen is better. For your neck, for example. So here Novell is making a first, untested I assume, mistake. If these ‘tests’ on ‘non geek users’ didn’t take efficiency (number of mouse clicks) and ergonomics (menu bar down by default, not smart) into account, and presumed that users should not be told of anything they could not discover for themselves, then this ‘test’ is already compromised by the fact that most people sort of know how the Windows interface works.
Couldn’t agree more. I’m not sure who they tested on because:
1) The “Computer” menu doesn’t mean anything to me. Why would I look for applications in the computer section? An “Applications” or “Programs” menu is more intuitive. Even the much maligned “Start” button is more obvious (just make sure you place the logout/shutdown button outside of this menu;-]).
2) It doesn’t obey Fitt’s Law. It’s easy to get to the “Computer” menu, but once there, you have to hunt, in two dimensions for what you need to look for. Two dimensional mouse hunting is something frowned upon in the accessibility community.
3) Novell KDE’s SLAB solves the “two dimensional hunting” issue a bit better:
http://www.desktoplinux.com/files/article106/sled102-new_menu.jpg
But it still suffers from a number of problems (shared with the GNOME version in some form). For instance, if we’re going to have tabs, why not just make them separate menu items (a la GNOME Applications/Places/System) to conform with Fitt’s Law and expose the options. And it’s not at all obvious how to add something to your favourite’s list (I assume it’s possible). If it’s not possible and it’s autogenerated, them the “Favourite’s List” can confuse people since items will pop in to or out of that menu and move around in that menu as you use it. Anyone who has had to support windows users who turn on “Personalized menus” has heard the common complaint “It was there yesterday, I swear, but now it’s gone.” Anyone who has used it know you can’t rely on “muscle memory” to quickly find items because they keep moving around.
This isn’t to say that the GNOME menu couldn’t do with some improvement. The (KDE) SLAB, does have some good ideas, and so does Gimmie ( http://beatnik.infogami.com/Gimmie ). But blindly following what Windows won’t help GNOME (or KDE) become more usable or intuitive or easily learned.
Edited 2006-12-29 15:47
Novell’s Gnome menu is not more usable than the present default Gnome menu.
The former actually needs more mouse clicks.
That would be true only every program was used at the same frequency by all users. In reality most users only use a few programs on a daily basis. By having these on shortcut, lots and lots of mouse clicks are saved.
Plus, why sacrifice one window thumbnail/virtual desktop bar? Because virtual desktops are confusing;
I agree with you, that the Novell finding that one panel should be better than two is somewhat odd. If we apply Fitt’s law, two panels should present a much larger target to interact with.
However, there are also advantages. For one thing most documents we create have portrait layout. Removing the top menu would create more screen space for the document you are working on. This would be especially important on small screens such as laptops.
The main reason for users to prefer one panel at the bottom layout is probably, as you suspect, that they are used to having it that way from windows.
Allowing to users make use of knowledge gained elsewhere is part of being usable. E.g. we could probably make a car easier to control by using a joystick, but any benefits from doing so would be undone by the fact that people already are used to having a steering wheel, a clutch, a brake and an accelerator. This way of thinking also applies to desktop environments.
I also doubt that having the panel on top or at the bottom makes much real difference with respect to muscle tension related problems. The reason is that you use it too seldom. How you arrange your monitor, and what chair you use will have much bigger impact on this.
BTW,the menus you use the most will be the menus of your applications, and they are likely to be closer to the top than the bottom.
Novell is making a first, untested I assume, mistake. If these ‘tests’ on ‘non geek users’ didn’t take efficiency (number of mouse clicks) and ergonomics (menu bar down by default, not smart) into account, and presumed that users should not be told of anything they could not discover for themselves, then this ‘test’ is already compromised by the fact that most people sort of know how the Windows interface works.
Usability is more than counting number of clicks. Theory is good, but when theory and reality differs, it is usually the reality that is right. That 90% of potential Linux users have experience with windows is part of that reality, and whatever tests we do will be influenced by this. What good would it be to design systems that would be usable for users that are not likely to use the system.
Some truths to the above, especially:
That would be true only every program was used at the same frequency by all users. In reality most users only use a few programs on a daily basis. By having these on shortcut, lots and lots of mouse clicks are saved.
But the following is short sighted:
That 90% of potential Linux users have experience with windows is part of that reality, and whatever tests we do will be influenced by this.
From what I’ve read of Novell’s usability testing, it seems that most of it is done on people in the USA or Europe (maybe all USA, I don’t know for sure) who have experience with computers. But 90% of potential future linux users are probably in Africa, China, South America, South Asia, etc. In the next 20 years, school districts, community centers, hospitals, and offices in these countries will begin to adopt linux more and more.
Sure, maybe this is not where Novell’s revenue will come from primarily, which is why they do testing in areas where they think they can sell licenses and support contracts. But don’t think that new users to commercial linux and new users to linux in general are the same group of people.
I personally don’t care whether Corporate USA wants to give their money to MS, Novell or IBM, but when it comes to public school systems in Bangladesh making decisions about software, I hope they choose freedom and free-as-in-beer products rather than giving money to MS.
FYI, those are screenshots of the first beta. The second one is much more polished, and can be downloaded freely as well.
Even if they wait for the next version. RHEL is about being stable and not bleeding-edge. If anything, it will mean that they’ll be using 2.10.
Actually, looking at the screen shot myself they’re using 2.15.x, so maybe 2.16?
I wish they’d hurry up and release so the CentOS guys can get to work.
I blame in my blog the RedHat contributions to OpenSource, but who knows, probably they will return back to help the community.
http://blog.360.yahoo.com/blog-hfBUWyYycrbdJvnnCIOPpqEC1Za5UCLp?bid…
Edit: the title I should be change: RedHat must help Fedora as well.
Edited 2006-12-29 14:37