[Date Index] [Thread Index] [Author Index]

Re: Why use Java in Mathematica?

• To: mathgroup at smc.vnet.net
• Subject: [mg78211] Re: Why use Java in Mathematica?
• From: David Bailey <dave at Remove_Thisdbailey.co.uk>
• Date: Tue, 26 Jun 2007 04:09:16 -0400 (EDT)
• References: <a6e65e8d0706220205m63dc9817id956db71bd8457c3@mail.gmail.com> <f5o7iq$4tk$1@smc.vnet.net>

??? wrote:
> Dear all,
> 
> I have tried V6.0, and find that whenever kernel starts, JavaLink is
> loaded. By inspecting javaw.exe's running status, it can be found that
> Java is started for PacletManager.
> 
> Even disallow Mathematica to access Internet, PacletManager is loaded
> at start up as usual. I find that the loading of PacletManager is coded
> into mainload.mx, so it seems difficult to get rid of PacletManager.
> Since I do not need PacletManager currently, so I modified
> PacletManager.m to create the PacletManager` contex only and do
> nothing else. Then Java is no longer loaded at start up of kernel.
> 
> Personally, I hate Java:
> 1. JRE uses two much memory, and >50MB for PacletManager
> 2. for personal taste.
> 
> Personally, I think that JavaLink is great feature to show that
> Mathematica is open, and can interact with other programming
> languages. But it is not a good thing to make JavaLink a required
> component to Mathematica. (Some functions need Java)
> 
> The reason for using Java in Mathematica is only for Internet access.
> And remind that some other scientific computing software (such as
> ..... as you know) even use Java for building their user
> interface! Mathematica is immaculate in most part of the kernel and
> user interface. But still, I wonder why using Java, as a required
> component?
> 
> I suppose that WRI just want to make the codes related to Internet
> access cross-platform. WRI has already made both the frontend and
> kernel cross-platform, why not implement Internet accessing functions
> on different platform NATIVELY?
> 
> All in all, I am afraid that one day, Mathematica will use Java to
> build their interface, for cross-platform. If this happened, I would
> say that Java had done great harm to the whole software industry.
> 
> BTW, Borland CLX library, WxWidges (WxWindows) are all cross-platform
> libraries on source code level and we can build native executable for
> different platforms from the same source code.
> 
> --
> Li Zhengji
> -------------------------------------------------------------
> If all you have is a hammer, everything is a nail.
> -------------------------------------------------------------
> 
I think it is important to remember that modern operating systems page 
their memory. In other words, if JAVA.EXE is not often active, it will 
not be using much if any memory - but simply occupying a little disk 
space - of which most machines have plenty.

Java is obviously convenient because it provides so much cross-platform 
functionality, and I can't quite see how one can justify excluding it 
for reasons of personal taste!

Having said all that, my experience is that applications that are 
constructed out of multiple processes, become less stable in proportion 
to the number of such processes (because of unresolved timing issues), 
despite the best endeavors of their developers!

I would also like to know just how far the packlet concept is intended 
to go. For example, if actual code improvements (bug fixes) were to be 
delivered in this way, I think it would be a very unfortunate step.

David Bailey
http://www.dbaileyconsultancy.co.uk

• Follow-Ups:
  • Re: Re: Why use Java in Mathematica?
    • From: Murray Eisenberg <murray@math.umass.edu>