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

Re: V.6.0.2 gripes...

• To: mathgroup at smc.vnet.net
• Subject: [mg86679] Re: V.6.0.2 gripes...
• From: David Bailey <dave at Remove_Thisdbailey.co.uk>
• Date: Tue, 18 Mar 2008 04:49:31 -0500 (EST)
• References: <fr8b56$gu8$1@smc.vnet.net> <frash5$1no$1@smc.vnet.net> <frcluo$rrv$1@smc.vnet.net> <frdft4$b0d$1@smc.vnet.net> <frkv56$vf$1@smc.vnet.net>

Helen Read wrote:
> David Bailey wrote:
>> Helen Read wrote:
>>> We have been seeing the Java timeout error a lot, on machines on campus 
>>> and at home. It started a few months ago, and the upgrade from 6.0.1 to 
>>> 6.0.2 has not helped. If anything, I'm seeing the Java timeout more 
>>> frequently than before, even on two very speedy new machines where I was 
>>> rarely getting the Java error before. What happens much/most of the time 
>>> now is that when you start up Mathematica and go to execute a command, 
>>> Mathematica just hangs for a while. Eventually, after quite a long wait, 
>>> you get the first output, together with the Java timeout error. 
>>> *Usually* it's OK if you quit Mathematica and restart, or if you just 
>>> close up the error window and go on.
>>>
>> In the absence of a definitive response from WRI, I think it would be 
>> worth experimenting with a machine disconnected from the internet. I 
>> don't get this problem myself, but I have written code that depended on 
>> paclet information. This worked well if the machine could not access the 
>> internet (once the relevant data had been downloaded once) and well if 
>> the machine had a good internet connection. However with a weak 
>> connection, the system would seem to start the check for more up to date 
>> information, and then get stuck/crash.
> 
> It seems unrelated to internet access. We have quite good connections at 
> the university, and we get the Java error there all the time. I have 
> dial-up at home, and I get the Java error regardless of whether I am 
> connected.
> 
>> I wonder if Mathematica is trying to update some component on startup. 
>> Any internet access would probably occur via Java. In particular, have 
>> you configured your Mathematica to pre-load any software?
> 
> Nope. But apparently Mathematica itself tries to load some Java stuff on 
> startup, hence the timeout errors.
> 
Clearly Mathematica must only wait for so long for Java to come live, so 
a sufficient delay in starting Java will cause the problem. I am trying 
to think of ways in which your machine (is it Windows XP, which I use, 
or the dreaded Vista) might differ from mine, because so far I have 
never seen the timeout. One obvious possibility raised by the original 
post is that your virus scanner introduces some delay - maybe only the 
first time a command executes... Disconnecting both the internet and 
your virus scanner would be an interesting experiment. Failing that, I 
would try disabling any other software that you have running - again as 
an experiment. Of course, a very old and slow machine might fall foul of 
a timeout.

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