🚀 go-pugleaf

In article <8jj1H3ihoa@atlantis.psu.edu> mek@guinan.psu.edu (Mark E. Kotanchek) writes:
>
>NeXT has lost a hell of a lot more than 200 sales because they have  
>inadequate numerical tools. I got my NeXTstation a year-and-a-half ago and  

There is always DaDisp. (1/2 smilie)  I have played with the demo version for
the NeXT.  Unfortunately, it is little more than a port of their PC version
with a NeXTStep (almost) front end slapped on.  The graphics look pretty awful
IMO, (just like the PC version).  The fonts aren't PostScript.  They even
made a menu with function keys labeled (F1, F2, ...)!  

I will say that DaDisp on the PC is usable.  It provides a somewhat intuitive
environment for manipulating signals.  A comparison of MATLAB and DaDisp that
I did a few years ago showed that DaDisp was sorely lacking in crunch power.
I believe that this stems from their implementation of virtual memory.  DaDisp
was built on a DOS extender originally.  The size of the buffer where virtual 
memory swapping takes place is about 4K or 8K data points.  That's a real
pain when you want to work with Mega point data sets.  Secondly, if you have
tried importing data into DaDisp (specifically binary data), the process is
convoluted and not scriptable (at least for the 2.0 version it wasn't).  I
found that when trying to isolate a segment of data more than 999999 points
into the data set, I couldn't get it imported.  (I was treating the previous
million points or so as a header.  The length of a header was limited to 
999999bytes.)

Well, in my best Linda Ellerbee impression,
And so it goes,

-- 
Bill Murphy               | The Call of the Norwegian Whipperwil:
			  | "I tink so, I tink so, probably."
murphy@physics.purdue.edu | "I dunno, I dunno, who can tell."