Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!ucbvax!ucsfcgl!babar.mmwb.ucsf.edu!srp From: srp@babar.mmwb.ucsf.edu (Scott R. Presnell) Newsgroups: comp.sys.sgi Subject: Re: Searching for good vis. software Message-ID: Date: 11 Apr 91 14:59:14 GMT References: <1991Apr10.145812.1@happy.colorado.edu> <1991Apr11.134841.23776@zoo.toronto.edu> Sender: daemon@cgl.ucsf.edu Lines: 75 martin@zoo.toronto.edu (Martin Hofmann) writes: >In article <1991Apr10.145812.1@happy.colorado.edu> kmsahr@happy.colorado.edu writes: >> I'm new to the net so I'll bet this topic has been >>addressed before, but here goes... We just got an SGI >>4D/340VGX and we need a program for graphing and data >>visualization. Ideally, we'd like a program with these >>features: >> >> 1. Ability to do a wide variety of plots, from simple 2D >>histograms up to large 3D contour and surface plots. >> >> 2. Easy to use; our previous experience with plotting >>programs is on Macintoshes. Plotting programs there are very >>user-friendly but not very powerful... >> >> 3. Able to run from the console, X-terminals, and >>Tektronix terminals. >> >> 4. Able to create contour and surface plots from random >>data (i.e., data that doesn't lie on a regular X-Y grid). >> >> 5. Allows us to freely add labels, text, etc. to the >>finished plot using Mac-like drawing tools. >I would like something like this, and with one more feature: > 6. Ability to produce high quality PostScript output of >the graphs. >Surely, there must something that does this out there in net-land. >Doesn't everbody need to produce graphs of various sorts? Oh, and I'd like a Ferrari 512 BB (well, I'll settle for a Porsche 959), and a house on the Monterey coast... But seriously. Our group has spent some time looking into this problem. And that's what it is - a problem. About the best we've come up with is "S" from Bell Labs. It will do 1), 3), 4), 5) and 6) listed above if you remove the phrases containing the word "Mac." XImage (from NCSA) has a more "point and click" interface, but not nearly as powerful (as of 6 months ago). The advantages are that it has graphical drivers for NeWS, X11, gl, and postscript. You can place labels, change fonts, line styles, marker symbols etc. There is a crude tool for viewing N dimensional data (but under the current IRIX implementation of the X server, it is verrrrry sloooow). And semi-complete on-line documentation. The bad news is its a memory hog (not unusual to have 8M processes), the code is difficult to read (and therefore trouble shoot), core dumps occasionally (I've tried all possible permutations of malloc support), and it's driven by a functional language - not a point and click interface. "S" is really first and foremost for data analysis, not for drawing graphs - but it is the only thing I can think of that matches the requirements you have put forth. The way we have it set up here - I have it running on the Iri and people with Sparc I's use their machines as X11 windowing clients to the process running on the Iris. The best thing I can say about it is that our most heavy data analysis people like it very much. Check out the book "The New S Language" by R. A. Becker, J. M. Chambers, and A. R. Wilks (ISBN 0-534-09192-X). Or write: AT&T Software Sales P.O. Box. 25000 Greensboro, NC 27420 1-800-828-UNIX - Scott Presnell -- Scott Presnell +1 (415) 476-9890 Pharm. Chem., S-926 Internet: srp@cgl.ucsf.edu University of California UUCP: ...ucbvax!ucsfcgl!srp San Francisco, CA. 94143-0446 Bitnet: srp@ucsfcgl.bitnet