Re: Quit Kernel when using Manipulate crashes
- To: mathgroup at smc.vnet.net
- Subject: [mg103082] Re: Quit Kernel when using Manipulate crashes
- From: Vince <blueschi at gmail.com>
- Date: Mon, 7 Sep 2009 02:36:21 -0400 (EDT)
- References: <h7qs00$8nk$1@smc.vnet.net>
On Sep 4, 7:00 am, Bruno Silva <bfasi... at gmail.com> wrote: > Hi, > > Whenever I use Manipulate and quit local kernel, Mathematica quits (and > often I lost unsaved inputs). > Does anyone have an idea of what may be wrong in my procedure? > (I'm working with Ubuntu) > > Thanks > Bruno Bruno, That is only one of the failure modes of the Front End (though a new one to me). There are others, that have significantly disrupted my workflow; and recently, threatened an imminent customer presentation. For example, I had a computation underway, with an expected runtime of ~ 1.5 hrs. Simultaneously, I was writing a set of charts in the Front End. I enabled 'Slideshow' mode (stylesheet) for a preview, and discovered I needed a slide break. So I took the fatal step (in Slideshow mode, I suppose) of inserting a 'Navigation Bar'. The Front End froze. After a long time, it displayed that dialog about a dynamic evaluation not completing in time (to that effect). It prompted me to wait or abort. I trusted neither and so took the fifth, as I needed the background computation to complete (and save to disk) for the above-mentioned presentation. Unfortunately, the Front End eventually decided to quit on its own, and promptly disappeared. I restarted the computation, but for a smaller data set, since time was now short. This is what I mean by 'disrupted my workflow'. Sure, you can attach caveats and prohibitioins to the tool, to avoid such close-calls. But that's also a reminder that you're not working with something that's industrial grade. The Front End is certainly clever and competent, but it's also fragile. I believe that WRI has (further?) destabilized the Front End with the introduction of dynamics in version 6+. I hope they spend some time cleaning house and tying up loose ends before releasing the next major version, to improve its reliability, especially for larger problems. Vince Virgilio