Darren New wrote:

Given that MS is, to a large extent, working to maintain backward compatibility in many things,

.Net and the CLR kinda broke the *entire freakin' Windows API*. Visual Basic .Net didn't even maintain compatibility with Visual Basic let alone the API. OpenOffice is better at reading old Word documents than new versions of Word.

Microsoft is a *lousy* example of backward compatibility on almost every front.

and given that open source projects like Ruby nevertheless needlessly break backward compatibility with every minor release, I don't see where having it "open" is that big a win.

And other projects like Apache, Python and Tcl *don't* don't break backward compatibility on every minor release. In addition, they also declare certain "business versions" that are then supported for much longer than the intermediate ones.

Finally, being "open" means that I can choose a specific version of, say, Ruby, an *not get forced to upgrade* because I can backport changes that I need. It's not easy, but RedHat has certainly demonstrated that backporting can be made to work.

How many Java AWT-based applets still run?

Ummmm, almost all of them.

Computational Geometry Applet from 1998:
http://maven.smith.edu/~orourke/books/CompGeom/CompGeom.html

This has been my actual experience with Java. An abandoned Java project is *far* easier to resurrect that an abandoned C/C++ project. Not only will the Java project still compile with the correct flags, the actual bytecode and application probably still runs *without* recompiling. In addition, any library dependencies probably exist as JAR's and even *they* still work.

Just try to find a 1998 C/C++ program that runs without recompiling on a 2007 OS. Games are your best shot, and even most of them don't.

-a

--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Reply via email to