[comp.sys.mac.apps] Summary

chang hsu liu <cliu@copper.ucs.indiana.edu> (05/18/91)

	A couple of days ago I asked the net about the graphing program to 
substitute Cricket Graph. 14 people responded my request. KaleidaGraph get
the most votes, Igor is the next, DeltaGraph is the third. Many people addressed
that which program to go is depend upon your needs. KaleidaGraph and Igor are more 
suitable for scientific graphs; DeltaGraph is more suitable for business use. 

	My sincere thanks to all the people replied the mail. With your
suggestions, our office has more options to go. Below are the messages I
received.

From: eto@seti.jpl.nasa.gov (Edward T. Olsen)
 
Which program is best depends upon what is required and who the user is.
I have had experience with all you mentioned except WingZ.  My preference
is KaleidaGraph.
 
KaleidaGraph is best for scientific work, especially when looking at data
in different ways and making presentations.  It does not support 3-D.
 
DeltaGraph appears to be best for management types.
 
Excel and (I assume) WingZ would be best choice when you have dynamic data
(i.e., you keep changing parameters and checking result).  I do this often,
but still import the results to KaleidaGraph for my work.
 
CricketGraph was wonderful in its day, but it has not been supported by its
makers and is now eclipsed by KaleidaGraph (and likely will not run on the
newer machines or operating systems).
 
Other scientific graphics you might investigate are Igor and the SpyGlass
data visualization set (dicer, etc).  But they are for special scientific
use.
 
   Good Luck,
 
  Ed
 
---------------------------------------------------------------------------
 
From: <owen@raven.phys.washington.edu>
 
Excel stinks for scientific graphing. Even getting an x axis is a chore,
and I always have to look it up in the manual.
 
I have only used Cricket Graph, Wingz, and DeltaGraph for short periods. None
were very good (Cricket being the worst), but I may have missed some things.
 
We use KaleidaGraph, and it's good. It behaves like Cricket Graph done right --
similar paradigm, but better output, better number crunching, faster...
 
But if you are only doing 2-D graphs (e.g. not contour plots), and can live
without radial graphs (alas, we can't live without them), then buy Igor! It is
far better than even KaleidaGraph. Igor offers:
- gorgeous output, both of individual graphs and of pages of graphs
- flexible data input (e.g. many different format text files can be read)
- awesome macros -- you can automate nearly anything if you like, including
  data input, analysis, and page layout...
- incredible speed -- for large graphs nothing I've used comes close
- very sophisticated data handling (but only of arrays -- hence the 2-D
  graph limitation). KaleidaGraph and the spreadsheets are pretty good,
  but Igor beats them hands down.
- good on-line help
- extensible (but you need a compiler to write add-on modules, of course)
 
Get the Igor demo. Work with it a bit; the paradigm is a bit odd at first,
especially if you're used to CricketGraph, but once you catch on, it's really
easy to use and remember. And it's different for good reasons (mostly -- why
they call arrays "waves" is beyond me).
 
Igor is sold by Wavemetrics:
wavemetrics@applelink.apple.com
(503) 620-3001
 
I have no relation to them except I WISH I was a customer (until they support
r-theta graphs, we can't justify it).
 
-- Russell
 
 
---------------------------------------------------------------------------
 
From: Alan R. Fry <afry@uhura.cc.rochester.edu>
 
I have used all of the programs that you mentioned.  The one I find myself
using most is Kaleidagraph.  It is very powerful, very flexible, and very
fast.
 
Excel and Wingz are both multi-purpose programs which are quite good at a
lot of things.  I have both, and I mostly use Excel because I haven't taken
the time to vigorously learn Wingz, which seems to be a more modern
program.  I wouldn't recommend Excel for graphics, at least not version
2.2.  Version 3.0 is supposed to be a lot better, but I guess we'll have to
wait and see.  Anyway, my opinion is that spreadsheet programs are a bit of
overkill for scientific graphic, unless you do a lot of your calculations in
your graphing program.  Kaleidagraph, by the way, has great column-based
calculation abilities.
 
I have used Deltagraph, and I'm not all that impressed.  It has some nice
3-D graphing capablities (Wingz's are better in my opinion), but it's 2-D
graphing falls way behind Kaleidagraph.  You should probably play with it a
little to see if you like it.
 
If you are doing a lot of data *analysis*, then I recommend a program called
Igor.  It is not as easy to use as Kaleidagraph, and the graphs don't turn
out as pretty, but it is a very powerful data-manipulation and analysis
program.  It has a command line interface, full programability, and a lot of
nifty built in features.
 
It sounds like, for your purposes, Kaleidagraph is your best bet.
 
Good luck,
 
Alan
 
--

chang hsu liu <cliu@copper.ucs.indiana.edu> (05/18/91)

---------------------------------------------------------------------------
 
From kchang@ncsa.uiuc.edu Fri May  3 18:54:23 1991
 
I've used both Deltagraph and Kaleidagraph and right off, I
can safely say both are marked improvements over Cricket Graph.
(which hasn't been meaningfully upgraded since 1987 or something).
Which one is right for you depends on your needs:
 
1) Do you need 3-D graphs? Only Deltagraph offers those.
 
2) Do you have *lots* (thousands) of points? Deltagraph is
flakier on large data sets (as in those taking up 500K+ of
disk space). Also, it doesn't give you control over the number
of tick marks (except turning them off completely,
so if you have two many categories, you end
up with a solid mass of tick marks along the x-axis.
   In general, Kaleidagraph offers much finer control over the
appearance of the graphs. Another annoying thing is that
Deltagraph files blow up in size pretty quickly. And it's
pretty stupid in updating the screen (updates *everything*
even if you just change the font of the label, for instance)
which again can be annoying if you're working with large data
sets. (Actually, Kaleidagraph can be a bit slow too. And both
are still much better than CricketGraph.)
 
3) Deltagraph can export the files EPS files, which is mighty
nice if you're using TeX and psfig. Importing Kaleigraph files
into other programs is sometimes an adventure. (Works in WriteNow
and Word, but not Canvas, for instance.)
 
4) Kaleigraph has extensive macro/formula capabilities that
Deltagraph doesn't.
 
  The reviews have generally favored Deltagraph, but it's also
  obvious that the reviewers aren't scientists and didn't try
  to do anything with large data files. Or need to refine the
  graph appearance too much. Deltagraph seems to have been aimed
  at business graphing applications (where you have tens not
  thousands of points); that's not to say it can't produce good
  scientific output (which, for some reason, I assume you want)
  but it does have its limitations. Kaleidagraph also has its
  weaknesses (it's kind of ugly and one can get lost in its
  myriad of dialog boxes), but it does seem amalgam of people's
  wish lists in a 2-D graphing program. My recommendation is
  to get both if you can afford them both. If you can only afford
  one, then you'll have to base your choice on your needs. They're
  both good.
--
  Kenneth Chang           | National Center for Supercomputing Applications
  kchang@ncsa.uiuc.edu    |                 Consulting Office/(217)244-1144
 
 
---------------------------------------------------------------------------
 
From: nvi@mace.cc.purdue.edu (Charles C. Allen)
 
 
I used to own DeltaGraph and was disenchanted at how "business"
oriented it was.  It also had problems getting markers and error bars
centered correctly.  The interface is also somewhat clunky, with every
little change being done via its own modal dialog box.
 
 
For scientific graphs, the choices are Igor (WaveMetrics) and
Kaleidagraph (Synergy Software).  Neither is perfect, but they both
can make good, clean plots.  Neither can make z=f(x,y) 3d plots yet.
 
Charles Allen                           Internet: cca@physics.purdue.edu
Department of Physics                   HEPnet:   purdnu::allen, fnal::cca
Purdue University                       Bitnet:   cca@fnal.bitnet
West Lafayette, IN  47907-1396          talknet:  317/494-9776
 
 
---------------------------------------------------------------------------
 
From: INM501@csc1.anu.edu.au
 
        I've started using Cricket Graph some 5 years ago.  It was a great
program then.  Subsequently, I've used KaleidaGraph(KG), DeltaGraph(DG) and
a bit of WINGZ.
        If you come from a CG background, you don't need a manual to use most
feature in KG.  It is fast, reliable and neat.  I like it a lot but my
school decide to buy DG and I am now stuck with it.
        You definitely need a manual for DG.  The program is very flexiable
and you can define your own graph type.  The drawing tools are much nicer
than that of KG and CG.  The worst thing is most operation is not intuitive.
eg. double click on an axis does not allow you change the scale of the axis.
It does a lot of redrawing as in CG but is significantly faster.
        I've used a little bit of WINGZ that I don't claim to be an expert.
DG feels like a cut down version of WINGZ to me and is defintely much cheaper.
 
        I guess the bottom line is: what sort of graph do you want to plot.
For simple scatter, column, graph, KG is the best.  All operations are faster
compare to the other packages.  The best thing I like is the flexibilties of
the input file.  If you want flexibilties in graphs, DG may be a better choice.
I would buy WINGZ is I have money to spare.
 
Ida
 
 
---------------------------------------------------------------------------
 
From: hoepfner@heawk1 (Patrick Hoepfner)
 
   We use KaleidaGraph and Igor.  These are best for scientific uses.
They both will let you set up multiple graphs on the same page.  Igor
uses scripts and seems to be more powerful but it isn't as slick as
KaleidaGraph.  Neigher of these do 3D graphs.
 
   I just read an article some time back about meaningless graphs done
for the sake of that 'gee whiz' look.  The author questioned if a 3D
pie chart gave the audiance any more information than the 2D variety.
The author suggested that clean and simple is better than cluttered
and overdone.
 
      +--------------------------+---------------------------------------+
     /    Patrick Hoepfner       |    NASA/Goddard Space Flight Center    \
    / America Online: PatrickH9  | Internet: hoepfner@heasfs.gsfc.nasa.gov \
   +-----------------------------+------------------------------------------+
 

cliu@copper.ucs.indiana.edu (chang hsu liu) (05/18/91)

---------------------------------------------------------------------------
 
From: Scott Barvian <barvian@dart.ece.cmu.edu>
 
Check MacWorld 7/90 for a basic discussion.  It convinced me to buy DeltaGraph.
I have basically been happy with it, but:
 
1) it's really a "color" Mac program- to use it on B&W screens (I use both)
   you have to keep straight the difference between "colors", "grayscales"
   and "fill patterns".  Not tough but annoying at first.
2) The "formula evaluation" capabilities are limited.  Suppose I have two
   columns, of x data and y data.  I plot them, and then I use their "curve
   fitting" feature to get a least-squares-fit line formula.  I can't directly
   plot that line on my graph, but I can go back into the "spreadsheet" and
   calculate it. And it's not even that easy, because all the spreadsheet
   can do is process one operation at a time. That is, if the formula I have
   is "y=3x+5", the "3x" and "+5" are two separate operations.  And it's not
   like Excel either, if I decide to look at "y=2.9x+5" I have to repeat the
    whole process, I can't just change the coefficient and hit "recalculate".
   For scientific/engineering use this is DG's biggest weakness.  They do
   offer an Excel macro which is supposed to "hot link" the 2 programs, but
   I found that it doesn't always work.
 
Hope this helps.
 
                                       
---------------------------------------------------------------------------
 
From: abosse@reed.reed.edu
 
Here at Reed College, we use three different graphing programs. Each
has its own strengths and weaknesses, so I'll try and summarize:
 
CricketGraph:
 
This was one of the first packages to come out on the mac, and it
shows. Data entry is primitive, as are import and analysis.
Presentation is not that great either. It will let you manipulate the
numerucal data fairly easily. Very easy to learn, does not require a
Mac II.
 
DeltaGraph 1.3
 
A good choice for business. Presentation is its strong suite, with a
host of different options for attaching labels, using user-defined
symbols as data points, and so forth. Good import facilty. Reasonable
analysis tools, but still not good enough for serious scientific
work. Fairly easy to use, and fast. A good step forward after CG.
 
Igor 1.24 (?)
 
Without a doubt, the best graphing package available on the Mac.
Amamzing. Incredible. magnificent, what more can I say? Insane number
of highly flexible analysis options, built in automatic macroeditor.
Very flexible presentation options. Fast. Geared for the working
scientist. Downside: be prepared to spend a goodly amount of time
learning how to use it. The manual is well written, but dense.
Support and response to user-suggestions is excellent.
 
If your business has much of its data stored in spreadsheet files
already (Wingz, Excell) I suggest you stick with the graphing tools
attached to those packages. If you need a general duty, stand-alone
graphing program, get DeltaGraph. If you need to do scientific
analyses of data sets, and have the extra time, get Igor. By the way,
none of the three packages mentioned so far is particularly strong in
econometrics. If you need those kind of cpapabilities, stick with the
spreadsheets, or get a standalone statistics program.
 
Hope this helps,
 
Arno Bosse
Reed College
abosse@reed.edu
 
ps. The views expressed above are entirely my own, and not that of
Reed College.
 
 
---------------------------------------------------------------------------
 
From hedstrom@sirius.uvic.ca Sun May  5 23:44:54 1991
 
I've found found Igor very good for engineering and scientific type
graphs.
                
 
 
 

cliu@copper.ucs.indiana.edu (chang hsu liu) (05/18/91)

From macq@miguel.llnl.gov Mon May  6 16:37:09 1991
 
I've used Kaleidagraph, it is good. Interface probably similar to cricketgraph..
Unless excel has improved drastically in the last couple of upgrades, avoid it..
I'd worry about the hardcopy quality of wingz, as described by the following qu.
How easy is it to make a printed plot fill a whole page, and do you get fine li.
I've seen somebody else's Deltagraph plots on-sreen and they looked good.
Also investigate Igor. There are quite a few scientists who swear by it.
--
--------------------
Don MacQueen
macq@miguel.llnl.gov                           
 
 
 
---------------------------------------------------------------------------
 
From: john@micor.ocunix.on.ca (John Kane)
 
I was interested in your question about graphing packages.  
I must admit that we know nothing about any of them except 
Excel which I have used a bit. However, because you are talking 
about handling scientific data I thought I wod suggest that you 
have a look at the graphing capacities of the statistics progm
SYSTAT.  It is a package designed specifically for scientific 
graphical analyss l not will do a tremendous amount of things 
that something like Excel either will do or has a lot of trouble
doing.  For example it will jitter data to allow ber resolution in 
areas where there are a lot of data points, it can handle error easily, 
it can handle influence plots well when you are interested in the effef 
outlyers, etc.  It can easily handle all (or damn near all) of the
graphical as techniques that Cleveland in his book 'The elements of 
graphing data', (Clevel, Wm. S. (1985) "The elements of graphing data". 
Monterey Ca.:Wadsworth Advanced Books and Software.) describes (such as 
sploms) as well as a couple others thae way too esoteric for me to ever 
want to use.  If you ever need Tukey's Exploray Data Analysis types of 
graphs they are all there.  The package also allows almy conceivable
kind of data transformation plus the ability to graph about any two 
and three dimension equations you might want.
I currently am using V3.2 which has a less than pretty interface - 
a legacy ofrport from the DOS world.  There is reportedly a new
V5.0 out with a  real Mac interface. I've been too lazy and cheap 
to get the upgrade. I have seen the pl stats/graphic MYSTAT which 
reportedly is the pilot of V5.0 and it does follow e Mac interface 
quite well.  SYSTAT also exists in the DOS world and I think it n
VAXen and by now probably more platforms.
 
 For real data analysis rather than just pretty pictures I don't think 
you can o better.  Mind you with a good laser printer you can make 
some really pretty ps as well.  Oh yes - in my limited experience 
the technical support is good also You may be able to find the 
package at your university bookstore.  I bought my original copy 
(of V3.1) in my university's bookstore.  SYSTAT gives very substantial educational discounts.  If not, the address is
SYSTAT Inc.
2902 Central St.
Evanston, Il.
60201
Tel: (312) 864-5670
 
 No I don't work for the company but I am a satisfied user. If you really need y duty scientific graphing capacity then I really think you should take a good lt this.
 
 John Kane                                      john@micor.ucunix.on.ca
 Ottawa/Hull Canada                       (819) 770-5468
 Canada
                                           
-------------------------------------------------------------
 
From: gt7686b@prism.gatech.edu (BINDEMANN)
 
I have been a satisfied user of KaleidaGraph.  I use it extensively
for scientific line graphs.  It is much faster than the older versions
of CricketGraph is used years ago.  I do not know if Cricket has
made their program any faster.  You may also want to download the
demo version of IGOR from sumex-aim.stanford.edu (I think it is in
the info-mac/demo directory and is available via anonymous ftp).
I looked at this demo and found that IGOR seems to be more powerful,
however, KaleidaGraph is more suited for most of the work I do.
 
If you have any questions regarding KaleidaGraph I would be happy to
try and answer them for you.  Good Luck!
--
BINDEMANN,ALAN CHARLES
Georgia Institute of Technology, Atlanta Georgia, 30332
uucp:     ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt7686b
Internet: gt7686b@prism.gatech.edu
I'm a Nervous Wreck from Georgia Tech and a Helluva Engineer .
 
-------------------------------------------------------------
 
From MIKAEL_B@maja.teknikum.uu.se Wed May 15 07:01:26 1991
Date: Wed, 15 May 91 13:59 +0100
From: MIKAEL_B@maja.teknikum.uu.se
Subject: Graph
To: cliu@copper.ucs.indiana.edu
X-Vms-To: BOOT::IN%"cliu@copper.ucs.indiana.edu"
 
KaleidaGraph is in my opinion the best 2-dimensional graphing program around.
As easy to use as Cricket Graph but with several important extra features,
larger data, sets, faster, plotting of several data sets in the same diagram,
coordinate readout from the graph, zoom, scripting curve fitting to user
defined functions, good mathematics capabilities, excellent data import feature
etc etc etc etc. Deltagraph mostly sits on my bookshelf, together with Cricket
Graph, nice for 3-d colour graphs if thats what you like, also possible to link
to Excel under multifinder, so the limited set of mathematical functions might
not be to large a drawback.It is slow, does not place the plot symbols properly
so you can't read data from the diagram with high accuracy, 3d scatterplots ,
which I've used most, incorrectly mixes the x and y axis, very confusing. In
all a program I don't like to much but better than Cricket Graph which hasn't
hade a upgrade since ???. Besides Cricket Graph can even do a reliable linear
curve fit, you have to use polynom grade 1. Feel free to ask more.
Mikael Bergkvist
Uppsala University
Dept. of Tech.
Uppsala Sweden