Archive of UserLand's first discussion group, started October 5, 1998.

Re: Java vs MS Common Runtime

Author:Paul Snively
Posted:7/24/2000; 4:47:26 PM
Topic:Java vs MS Common Runtime
Msg #:18980 (In response to 18978)
Prev/Next:18979 / 18981

Timothy O'Hear: They have IE on the Mac though and I imagine that future versions *could* include the common runtime.

Sure, they could, but MS used to have Mac cross-platform development tools for their development environments and eventually gave up on them; I don't see this "common runtime" being any different. They'd either have to become MacOS experts or go back to the old approach of layering some Win32 subset atop MacOS. Those are the parts I find unlikely enough to rank their probability at 0.

Timothy: Is it the case that you can currently develop a reasonably complex Java (client-side) application on one system (Windows) and expect it to run on the others with good performance and reliability (Mac, Linux)? I've been very unimpressed with Java apps so far (the Forte IDE geing the most recent example). I also remember that at JavaOne Jobs & McNealy made a big thing about apologizing for how poor the current Apple Java VM is and promised things would get better "soon"...

Subjectively speaking, I think MRJ 2.2.2 is quite good for a non-Hotspot-client implementation. Java 1.1 itself is rather naive about the threading model and garbage collection (Java's GC relative to modern Lisp or Smalltalk implementations has always been embarrassing). I think it's true that the Symantec JIT has lagged the development on the more popular platforms, and this is only more true now that everyone else is doing dynamic optimization a la HotSpot. I've encountered no particular reliability problems in moving among VMs and platforms apart from buggy integration into browsers.

I can wish all I want that Apple had put more effort into MRJ prior to MacOS X, but by all accounts Java on MacOS X is just amazing, so perhaps it will have been worth Apple putting the overwhelming majority of their effort there. God knows it's easier to port the Java VM to a UNIX-derived OS than to MacOS.

Timothy: Probably the aspect that got me the most excited about the whole MS Common Runtime thing : based on previous MS frameworks, virtual machines and compilers I feel that they have a good chance of providing a great solution for dynamically loading code onto *Windows* machines. If their solution is reliable and performant then it could succeed in providing client apps where Java (as far as I can tell) has so far failed.

Sure, but it's not hard to succeed in creating client apps for a single platform! And dynamically loading code onto a single platform is no great shakes either. The real win will come from ubiquitous, secure, massively-distributed computation across the entire panopoly of computing devices both extant and forthcoming: everything from my Java ring to my iMac to the university's Cray to my boss' Windows box to my colleague's Palm Pilot to all the devices that haven't even been designed yet. The processor(s) and the OS(es) don't matter. Computation and communication matter.

There are responses to this message:

This page was archived on 6/13/2001; 4:55:48 PM.

© Copyright 1998-2001 UserLand Software, Inc.