s-news
[Top] [All Lists]

Re: blind computation

To: Prof Brian Ripley <ripley@stats.ox.ac.uk>
Subject: Re: blind computation
From: Spencer Graves <spencer.graves@pdf.com>
Date: Fri, 24 Dec 2004 12:34:44 -0800
Cc: Xao Ping <xao_ping@yahoo.com>, s-plus user list <s-news@lists.biostat.wustl.edu>
In-reply-to: <Pine.LNX.4.61.0412241638530.2711@gannet.stats>
References: <20041224155643.64615.qmail@web54001.mail.yahoo.com> <41CC45D0.7050200@pdf.com> <Pine.LNX.4.61.0412241638530.2711@gannet.stats>
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax)
Dear Prof. Ripley: Thanks very much for the additions. For me, "Options -> 'Save ... as default'" was only part of the story. Under Windows 2000, when I want to store work in a new folder, I copy an old *.SSC file into the new folder and double-click on it. S-Plus then creates new ".Data" and ".Prefs" subfolders, copying from some place like "C:\Documents and Settings\spencerg\SPLUS_Home". To make this change apply by default to all new project folders, I carefully replaced the master ".Prefs" subfolder after turning off buffering. That's what I meant about "standing on one's head in a very special way". Of course, I can't do that when I'm using a computer that's not primarily mine. Then "Ctrl-W" can be quite helpful.

Xao Ping: Regarding "the meaning of the parameter 'flush timeout'", I when Held -> "Available Help" -> "Search S-Plus Help" for "flush timeout". This search produced 2 hits, both of which explained that output would be buffered for the selected number of milliseconds. It will be a great feature whenever Insightful gets it to work as designed. hope this helps. spencer graves
Prof Brian Ripley wrote:

On Fri, 24 Dec 2004, Spencer Graves wrote:

You've just described a known bug with buffering S-Plus 6.1 as well. Fortunately, you can turn this off via Options -> "General Settings" -> Buffering: Turn it off or reduce the "Flush timeout" to something like 20.


You can toggle it on and off with Ctrl-W.

I turn it off routinely. This can be turned off permanently, but it requires standing on one's head in a very special way; I did it once under coaching from Insightful Tech Support and promptly forgot how.


All you have to do is to set it off and save the preferences with the last line of the Options menu -- hardly complicated!

Perhaps someone else can tell us how to access a file as "read only" while S-Plus is still writing to it. That should be possible, but I don't know how. hope this helps. spencer graves


Any decent editor can do it, e.g. Emacs, or running a Windows version of less or tail -f in a terminal window.

Xao Ping wrote:

Dear All:
I am working with S+6.2 under Win XP. With this version I am having difficulties which I did not have with the previous ones. I am performing a long simulation and trying to keep track of the current status of my computation. I previous versions, I used to just send a one-line info to screen after a certain number of loops. My understanding is that S+6.2 does not allow to print anything to the screen until the loop is ended. What I see is just a grey screen without ability even to minimaze it. Hence, I use ' write.table ' to send the info to an external file. However, S+6.2 does not allow me an access to this status file either until the loop is ended. Is there a way around this problem? I appreciate everybody's advise. But even more I would ask the Insightful people,


This is not the address to `ask the Insightful people', but I have seen it in the 6.x documentation. If you list a 10,000 x 100 matrix to the console you will discover why it is a Good Idea.

what is the motivation behind this seemingly impractical modification and what is their idea how a user should keep track of his computations.


Insightful Technical Support can point you at where exactly this is documented.


--
Spencer Graves, PhD, Senior Development Engineer
O:  (408)938-4420;  mobile:  (408)655-4567



<Prev in Thread] Current Thread [Next in Thread>