That's actually a good thing, and RH has done it before in *.0 versions. I would prefer, however, that 7.1 greet my machine with a rock-solid gcc 3.0 and recompile my kernel in the process.
So, in other words, you expect a problem in a Red Hat.0 release to be fixed by a.0 release of the compiler?
With regards to xinetd, I have long ago replaced the original inetd with xinetd on my machines. Enhanced security features (like binding to specific interfaces) and easier configuration are two of the reasons.
Furthermore, I do feel that every once in a while you should throw away something and write it from scratch. Experience shows that this cleans up the design and makes the program better.
The whole business of unsupported, unfinished, unknown, unstable ingredients that forces a soup-to-nuts upgrade cycle is made even worse when the non-compatible nature of that upgrade is not (at the very least) advertised clearly beforehand.
In my view, increasing the major version number is a very clear signal indeed. How much stronger a signal do you want?
Do you really feel that a kernel is usable as reference code for a compiler? An operating system kernel must be one of the "messiest" things to actually write in a not-so-low level language as what gcc compiles.
One might compare Tarantella to VNC, Citrix and other products in the same genre. However, Tarantella is more "Enterprise-oriented" than these:
In addition to X, it also supports Winframe and dumb terminals.
X (and Winframe and terminal) protocol messages are converted to a proprietary protocol, AIP in the Tarantella server. The X applications regard the Tarantella server as the X server.
With the Tarantella server being an X server, X font problems are eliminated by configuring Tarantella to us a X font server or adding font paths
The AIP protocol compresses X messages (up to 1:10), and implements these more efficiently. Bad X applications with stupid refresh algorithms actually looks better within Tarantella than on a regular X server, because redrawings are handled more efficiently.
SCO claims the compression makes X usable on a 9.6k connection. (It definetely works very well on 64k).
The X connection (between the application and the Tarantella server) and the AIP connection (between the Tarantella server and the user's browser) are semi-independent. If the user's browser crashes or the user logs out, the X connection is kept alive (this is configurable), and resumed when the user logs in again, possibly at a different location.
Tarantella servers can be arranged in clusters, making the server with the lowest load serve new users.
Equivalent application servers can be arranged in host lists, ensuring even distribution of users among the various servers.
What applications a given user is able to access is configurable and depends on the given user's location in an object hierarchy
All configuration and application and user management is done via an applet made available from within Tarantella for Tarantella administrators.
On the client side, all which is necessary is a (newish) Java-able browser. A native client is available for the big M.
The licensing policy is rather friendly, with licenses based on the number of active users using a certain server or cluster.
In my view, Tarantella is as much an enterprise and management solution as a technical solution. In a number of companies, this counts as much as the technology part.
Have you ever gotten a software for Linux that for which Free Source was not provided? Here's what you do: run ldd on it. Look for what shared libraries it uses. If not shared, check then with nm. What you are looking for is GPL'd libraries that they have used to make their slaveware. If you find any, then we win, the you may free the slave from his chains.
You forget that most libraries does not use the GPL license, but the LGPL license. This license allows further restrictions, and I don't think we kan make a case against software using such libraries.
Well, it's actually 48x98 mm. Just look at http://www.gaulanett.no/site/Storen_Trelast/articl e.php?id=16 . They don't even mention the inch dimensions or the unit of measurement.
I like Python for its clean and simple, while still somewhat string, syntax. The OO is a bit clumsy, however.
Java has good OO concepts, and a widely spread and known syntax.
Perhaps Jython, the Python interpreter running inside a JVM, would be a good combination for leveraging the best of the two environments?
It is kinda unavailable right now, for some reason. The link seems up, so it might just be that the server kneeled.
I tried both http://yahoogle.com/ and http://goohoo.com/. They were both already registered, however.
You could always try arguing that BitTorrent saves campus bandwidth for popular downloads like Fedora Core, Knoppix, and Firefox.
Two of these trackers are actually run by universitites..!
You might want to try this, or if that's to puny you might try this.
This might void your warranty, though...
You really should have submitted the screenshot link using Freecache ..
Only now it's too late, ofcourse..
So, in other words, you expect a problem in a Red Hat .0 release to be fixed by a .0 release of the compiler?
With regards to xinetd, I have long ago replaced the original inetd with xinetd on my machines. Enhanced security features (like binding to specific interfaces) and easier configuration are two of the reasons.
Furthermore, I do feel that every once in a while you should throw away something and write it from scratch. Experience shows that this cleans up the design and makes the program better.
In my view, increasing the major version number is a very clear signal indeed. How much stronger a signal do you want?
Do you really feel that a kernel is usable as reference code for a compiler? An operating system kernel must be one of the "messiest" things to actually write in a not-so-low level language as what gcc compiles.
One might compare Tarantella to VNC, Citrix and other products in the same genre. However, Tarantella is more "Enterprise-oriented" than these:
In my view, Tarantella is as much an enterprise and management solution as a technical solution. In a number of companies, this counts as much as the technology part.
cat. If you need more editing power, you have a design problem, not an editor problem.
You forget that most libraries does not use the GPL license, but the LGPL license. This license allows further restrictions, and I don't think we kan make a case against software using such libraries.