Messages in this thread | | | Date | Thu, 12 Nov 1998 18:38:35 -0800 (PST) | From | Alex Belits <> | Subject | Re: ActiveWhy (was Re: Comments on Microsoft Open Source document) |
| |
On Thu, 12 Nov 1998, Kevin Grey wrote:
> I thought the problem with creating crossplatform apps was with the actual > porting of the program... Isn't activeX only PC win? Plus compiling such > an app would still be an enourmously long amount of time relative to java > loading and execution.
Compiling will be necessary only once when applet/control/whatever is loaded. And I suspect that most of such things will be near trivial -- note that standardized library that should be used to make the thing cross-platform, will be already there, and most likely the transfer time for code and especially data will be longer than compile time on any reasonably fast processor.
Java however in all of its forms runs slow because: 1. Java VM is slow and buggy, 2. JIT compiler does not exist for a lot of platforms (I have never seen reliable JIT compiler for Linux). I can't see any benefit in Java unless someone needs a solution to do something that works on a lot of platforms and refuses to send the source code for compilation on platforms other than his own. While Java's original idea was different, it seems now that it all degenerated into "how to keep our programmers from using MFC in Windows, so things will kinda work on Solaris with Motif, too, and then avoid sending the source to the nearest guy with Linux box to type make".
> The best solution (I think) is to have Java > compiled machine code when speed is an absolute necessity. Then selecting > the program specific to your platform would be nice, although there is > absolutely no security features with such a situation because all the > security features of java are imbedded in the interpreter. Plus you would > never know whether the app was truely coded in Java and not C or some > other language. Still it would be nice to compile java byte-code into > machine code. I remember seeing some GNU util that dumps ASM... but not a > fullscale compiler yet.
It may be a good idea to allow binary download/execution if fuch thing is available, however it must be optional -- without source it will leave a lot of platforms unsupported, and if one needs an example, please look what happened with Quake when Quake II switched the support for its modifications from interpreter to compiled shared libraries (and, as opposed to ActiveX, it's possible to compile them for Linux -- just people didn't do that, and I suspect that few of them kept the code portable).
-- Alex
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |