RE: [S] Warning: GUI Spreadsheet editor full of bugs

Olsen, Norm (
Mon, 28 Sep 1998 12:32:23 -0700

I followed Bert Gunter's procedure exactly as he stated but encountered no
errors. The dataframes showed up in the object browser and on the command
line, complete with the columns containing NAs. I'm using S-PLUS 4.5
professional, release 1, running on Windows NT 4.0.

Having said that, I also think that the new GUI stuff is terrible. Far too
complicated and unintuitive. I'm gradually switching over to R. Not only
is R free but it offers greater performance (I've seen some of my code run
approximately 10x faster under R than under S-PLUS).


Norm Olsen
Stock Assessment Division
Department of Fisheries and Oceans
Pacific Biological Station
Nanaimo, B.C.
Canada. V9R 5K6
Phone: (250) 756-7328
FAX: (250) 756-7053

> -----Original Message-----
> From: Gunter, Bert []
> Sent: Friday, September 25, 1998 6:24 AM
> To: 's-news'
> Subject: [S] Warning: GUI Spreadsheet editor full of bugs
> To all:
> Just had an unpleasant experience trying to use S-Plus's (4.5,
> professional,
> under Nt4.0) gui spreadsheet editor to enter data. My experience was: it's
> a
> mess! Don't use it. Here's what happened:
> I tried to first name 4 columns in the GUI data frame editor. Unless I hit
> <enter> after entering the name at the top of the column, I got all kinds
> of
> error messages. However, the spreadsheet showed the names entered. I
> wanted
> to enter the data in only one column and then fill the others with regular
> sequences of indices (the usual simple flat file format). I did this and
> the
> other unentered columns filled with NA's. I then closed the spreadsheet
> and
> tried to print it from the command line. This showed a spreadsheet with
> only
> one column instead of 4 (The NA columns were lost). Opening the
> spreadsheet
> in the browser, however, the NA's showed. I then filled in the column
> sequences MANUALLY (ugh -- but not much data). The command line printing
> showed the data but did not have the names for the columns. Moreover,
> there
> were NA's in the first row even the the browser showed non-NA entries, so
> apparently the browser and object storage representations get
> unsynchronized
> somewhere along the line.
> Then I tried to open the spreadsheet again. This time the data that i
> entered was gone! Back to the command line. The data were gone there, too!
> ?$$%%! It's obvious that S-Plus has not figured out how to synchronize
> the
> stored data frame with whatever the spreadsheet is trying to do.
> Anyway, this is far too simple a task to put up with this nonsense, and I
> did nothing that was unusual. This appears to be yet another example of
> the
> difficulty S-Plus has had trying to integrate "standard" gui functionality
> into S-Plus. I endorse the effort, but, for the moment, it ain't ready for
> prime time so far as I'm concerned. I'll stick with the command line, at
> least when it comes to data entry. Of course, like most list readers I
> suspect), the gui would be for me only occasional useful (though it is
> essential for the scientists with whom I collaborate if they are to use
> the
> software), and I'd abandon S-Plus like a shot if they ever gave up the
> command interface.
> Hope this saves someone else some pain.
> -- Bert Gunter
> Merck & Co.
> The views expressed here are personal and not endorsed by Merck & Co. in
> any
> way.
> -----------------------------------------------------------------------
> This message was distributed by To unsubscribe
> send e-mail to with the BODY of the
> message: unsubscribe s-news
This message was distributed by To unsubscribe
send e-mail to with the BODY of the
message: unsubscribe s-news