Using DATAPLOT with the Web
|
Web Applications Increasingly Important
|
Incorporating applications into the web is becoming
increasingly important. There are several approaches to
doing this, which will be discussed in the context of
Dataplot. The approaches discussed are:
- Generating Dataplot
output for web applications
- Integration with the
NIST/SEMATECH e-Handbook of Statistical Methods
- Using Dataplot with CGI
scripts
- Robert Lipman's Web
Dataplot software
- Downloading files for
using Dataplot with the web
Dataplot is not currently available as a browser plug-in
and there are no current plans to do this.
|
|
1. Generating Dataplot Output for Web Applications
|
Output Can Be Either Text or Graphics
|
The most basic use of Dataplot in web applications is
the insertion of Dataplot output into web pages. The
Dataplot output can consist of
- Dataplot text output
- Dataplot graphics
|
Graphics Output
|
The most common file formats for graphics in web browsers
are:
- GIF (graphics interchange format)
- JPEG
- PNG (Portable Network Graphics)
- SVG (Scalable Vector Graphics)
The GIF and JPEG formats are almost universally supported
in web browsers. However, the GIF format uses a patented
compression algorithm that requires the payment of a royalty
by programs that generate GIF output. For this reason, free
programs, such as Dataplot, will not support GIF directly. JPEG
is an image format that was developed for video applications and
is able to generate highly detailed graphics with small size
through the use of "lossy" compression algorithms. Although
JPEG is most useful for photographic type graphics, it can be
used for statistical graphics as well. The PNG format was
developed in response to frustration with the royalty
issues of the GIF format. PNG was adopted by PERL, a
popular utility for web development, several years ago and
many other popular packages are now supporting PNG graphics.
For this reason, most of the popular browsers now support
PNG plug-ins. However, support for PNG in browsers is not
quite as universl as GIF and JPEG. SVG is an XML based
vector format. As XML gains acceptance in web development,
we anticipate the increasing importance of SVG graphics
for web pages.
Given the available options for incorporating graphics into
web pages, Dataplot provides the following options.
- On Unix, Dataplot can generate PNG and JPEG format
graphics directly. The JPEG driver is not
available on all Unix implementations. Also, these
drivers use an external library (specifically, the
GD library), so if your local implementation was
built from the source (rather than being a downloaded
executable), then check with your local implementor
to see if the PNG/JPEG driver is available. We
plan to port this driver to the Windows version
of Dataplot. More information for the
PNG/JPEG
driver is available in the online Reference Manual.
- Dataplot supports a SVG driver on all platforms.
More information for the
SVG
driver is available in the online Reference Manual.
- Another alternative is to generate a Postscript graph and
then use one of the free or commercial image conversion
programs to convert the Postscript graphic to GIF or
JPEG. One advantage of this approach is that you can
utilize the high quality typeset Postscript fonts. The
disadvantage is that you have to go through an
extra step in order to generate a web compatible
format.
|
Text Output
|
The simplest way to add Dataplot text output to a
web page is to do the following:
- Use the CAPTURE
command to an ASCII file.
- Within a web page, enclose the output in the
CAPTURE file between a <PRE> tag and a
</PRE> tag.
The 8/2002 version of Dataplot added an option to generate
Dataplot output in HTML format:
The HTML option specifies that the CAPTURE output should be
in HTML format. Note that most commands simply use a
<PRE> ... <PRE> syntax. Curently, the exceptions
are the TABULATE and CROSS TABULATE commands which write the
output using HTML table syntax.
Note that an additional feature of the HTML option is that
if DEVICE 2 is set to either PNG/JPEG or SVG, the graphics
will be incorporated into the web page using the
IMG tag.
You can use the WEB
command to view the generated HTML output from within the
Dataplot session. For example,
SKIP 25
READ RIPKEN.DAT Y X1 X2
ECHO ON
CAPTURE HTML C:\TABLE.HTM
TABULATE MEAN Y X1
CROSS TABULATE MEAN Y X1 X2
END OF CAPTURE
WEB file://C:\TABLE.HTM
In addition, if DEVICE 2 is set to PNG, JPEG, or SVG, Dataplot
will incorporate the graphics into the web page using the
IMG tag. For example,
device 1 x11
.
skip 25
read berger1.dat y x
.
line blank solid
character x blank
echo on
capture html fit.htm
set ipl1na data.png
device 2 gd png
title original data
plot y x
device 2 close
fit y x
set ipl1na pred.png
device 2 gd png
title predicted line
plot y pred vs x
device 2 close
end of capture
.
web file:///home/heckert/dataplot/solaris/fit.htm
NOTE: The following enhancements were made to the CAPTURE HTML
command in the 1/2003 version of Dataplot.
- If DEVICE 2 is set to Postscript, Dataplot will
automatically invoke Ghostscript to convert the Postscript
file to JPEG format. It will also add the appropriate
<IMG> tag to the HTML file. This capability is
supported under Windows and Unix. Enter
HELP CAPTURE for details.
- The SET POSTSCRIPT CONVERT command can be used to specify
one of several devices. When a Postscript device is closed,
Ghostscript will automatically be invoked to convert the
Postscript output to the specified format.
PDF (Portable Document Format) is one of the supported
devices. For PDF, the CAPTURE HTML will automatically
put a link in the HTML page to the PDF file. PDF may
be preferable to JPEG if you need to preserve higher
image quality.
Enter HELP POSTSCRIPT CONVERT for details.
- The SET HTML HEADER FILE and SET HTML FOOTER FILE comamnds
can be used to specify files that will be used for the
header and footer, respectively, of the generated HTML
file. Enter HELP HTML HEADER FILE for details.
|
|
2. Integration of Dataplot with the NIST/SEMATECH
Engineering Statistics Handbook
|
Engineering Statistics Handbook
|
Dataplot has been integrated with the
NIST/SEMATECH Engineering Statistics Handbook.
The Handbook is a web based statistics handbook that
has a number of case studies which can be run
using Dataplot directly from the web pages.
|
Steps in Executing Dataplot Macro When Invoked from Web
|
When a Dataplot macro is invoked from the case study
web page, the following steps occur.
- A check is made to see if Dataplot is currently
running.
- If a Dataplot session is currently running, the
macro is executed in the currently open session.
The check for a currently open Dataplot is
accomplished by the Tcl/Tk scripts that implement
the Dataplot graphical interface.
- If no Dataplot session is currently running, the
Dataplot software is invoked and then the macro is
executed.
- In either case, the Dataplot session remains
remains open (unless the macro explicitly closes
Dataplot).
This approach makes it possible to execute macros
successively. This is quite useful in the context of
the Handbook case studies. That is, the user can step
through the case study in small, easily understood steps.
|
Advantages and Disadvantages of Approach
|
The advantages of this method are:
- Dataplot and the web page are in separate windows.
This can be an advantage if you want the user to
be able to view explanatory text while also viewing
the Dataplot output.
- You can implement fairly involved macros in
small, manageable steps.
- If the Dataplot session is left open, the user
can execute Dataplot commands independently of the
web page.
- Dataplot is run on the users local platform, not
the web server. This can help minimize the load
on the server.
The primary disadvantage of this method is that Dataplot
needs to be installed on the user's local platform.
|
Two Basic Steps
|
There are two basic steps in implementing this method.
- A link to the macro needs to be incorporated in
the web page. You can view the HTML source for
the Dataplot case studies in the Handbook. For
example, check the HTML source for
one of the eda examples.
The Dataplot macros are contained in files with a
".dp" extension. These macro files are stored with
the other HTML files on the server. Any needed
data files can be added to the data sets directory
where the built-in Dataplot data sets are stored.
- Both the web server and your browser need to be
configured to know how to handle a ".dp"
file. The link contains instructions for how to
do this for the most common browsers and the most
common platforms.
|
Approach Useful for Tutorial Applications
|
If you only need to run a single macro, you can simplify
the above steps to run the command line version of
Dataplot as well.
In general, this approach is useful for tutorial type
web pages. That is, you want to teach your audience how
to perform a particular analysis using Dataplot.
|
|
3. Using Dataplot with CGI Scripts
|
Basic Approach
|
You can install Dataplot on a server computer. An HTML form
page will typically prompt the user for certain inputs.
Once the form is submitted, a CGI script will run Dataplot
on the server computer and then return either a graph or
Dataplot output to the web page.
CGI scripts are most commonly written in Perl, although
they can also be written in C or other scripting or
programming languages.
We do not document the details of implementing this
since there are many books on writing CGI scripts available.
This approach has been used by at least one project
at NIST for a production system.
The CGI approach can be used with many different
programs. There are no particular issues that you need to
be aware of when using Dataplot this way. The fact that
you can write command driven macros in ASCII files means
that Dataplot is well suited for this approach.
|
Advantages and Disadvantages of the CGI Approach
|
The advantage of this method is:
- The software does not need to be installed on the
user's local system.
The disadvantage of this method is:
- For a heavily used site, running a number of
Dataplot applications simultaneously can put
a heavy load on the server.
This approach is useful if you simply want to use Dataplot
as a computation engine or to generate a specific type of
plot dynamically. In this case, the end user may or may
not care whether or not Dataplot was used for the
computation.
|
Active Server Pages (ASP)
|
In a "web based computing" project in the NIST Statistical
Engineering Division, we have used Active Server Pages to invoke
Dataplot.
This is a similar idea to CGI scripts. Basically, you have a
web form that accepts certain parameters. Then ASP and VBSCRIPT
are used to generate a Dataplot macro on the fly and Dataplot
is invoked to run the macro and return the results to a web
page.
If you would like more information about this approach, contact
Alan Heckert.
|
|
4. Robert Lipman's Web Dataplot
|
Web Dataplot can be Adapted for Local Applications
|
Robert Lipman, currently of the NIST Building and Fire
Research Laboratory, has written "Web Dataplot". This is
a CGI script written in Tcl/CGI. We used to provide this
to run a live Dataplot session from these web pages. We
have removed it from the public web pages for firewall
considerations. Internal users can still run
web Dataplot. External users who would like to adapt
this for use on their own sites should contact alan.heckert@nist.gov
for more information.
One possible application of the web Dataplot software
would be as an online "problem solver". That is, if your
site has a specific set of problems with a standard
solution, you could modify the web Dataplot scripts
to create forms and run macros that solve these particular
problems.
|
|
5. Downloading Files for Using Dataplot with the Web
|
Downloadable Scripts Provide a Framework, Not A Finished
Solution
|
If you are interested in using Dataplot in a web application,
we have included links to appropriate software to help.
Note that you will need to modify the source provided to
fit your own application. These sources are provided as
a basic framework for developing your own web applications using
Dataplot, not as ready to use scripts. This source is provided
"asis" and we do not offer any additional support for them.
If you want to adapt these to your own applications, you
should have some basic web development experience (you should
not need a lot, but should have a handle on the basic concepts).
- If you are interested in an approach similar to the
NIST/SEMATECH Engineering Statistics Handbook, then
you don't really need any additional software other than
the user installing Dataplot on their local machine. The
discussion above provided links to pages that discuss how
to configure your browser on the most popular platforms
(other platforms should be similar) and a link to a page
that discusses configuring your web server for Dataplot
macros.
- Probably the most common approach will be to write a CGI
script. You can download a C
program called "step1.c" and the corresponding
HTML Form page called
"step1.html" that initiates the call. The step1.c
program also uses util1.c.
These sources were provided by Dale Bentz of the NIST
Building and Fire Research Laboratory. They are from a
web project that uses Dataplot as an underlying
computational engine. In this case, the end users do
not know or care that the computations and graphs are
being generated by Dataplot.
You compile "step1.c" and "util1.c" to create an
executable called "step1" that is stored in the
appropriate "cgi-bin" directory. You will need to make
changes to the C code and HTML form to adapt it to your
specific application, but these sources provide the
basic framework. Note that "step1.c" makes a call to
the "convert" program. The convert
program is a program that converts Dataplot graphs in
Postscript format to "gif" format for web viewing. You
will need to replace convert with whatever conversion
utility you have available on your site (or have
Dataplot generate PNG format graphs directly).
The C program could also be easily replaced by a Perl
script. Although I intend to provide an example of this,
I haven't had time to put this together yet.
NOTE 3/8/2001: I now have a Perl based example available for
downloading. This is just a simple example that accepts
some data from the user and generates a simple run sequence
plot and returns a PNG file to the browser. You will need
to modify it for your own applications, but it shows the
basic steps involved. Dataplot has added support for
PNG and JPEG graphics. This is based on the GD library
of Mark Bouttel (this is the library used by Perl and
numerous other software programs for PNG/JPEG graphics).
I will be updating the Unix executables on the ftp site
the week of March 13, 2001. Support for PNG and GIF on the
PC executables will be added in a few weeks (hopefully).
For Dataplot executables that do not support the PNG or
JPEG graphics, you will need to add an extra step to
convert Dataplot Postscript output to either GIF, JPEG,
or PNG. The example includes a
HTML form page,
a Perl script to parse the
form arguments, run Dataplot, and return a PNG format
graph, and the
sample Dataplot macro.
- Bob Lipman's web dataplot code requires the following:
- You need Tcl/Tk and Expect (these are typically
already downloaded and installed if you are running
the GUI version of Dataplot).
- In addition, it requires the
Tcl/CGI
library from
Don
Libes (the author of Expect).
- A tar file of
cgi scripts is available. This file should be
untarred in your "cgi-bin" directory. The main
script is "wdpi.cgi". The file "wdp.cgi" is the
HTML file that initiates "wdpi.cgi". This can
be installed as part of your web pages.
This approach provides a bit slicker approach than the
C program or Perl script provided above. The downside is
that it requires more web knowledge and additional software
installation. I would generally recommend starting with the
C program (or Perl script). Then if you desire a more
sophisticated approach you can try the web dataplot scripts.
|
Date created: 6/5/2001
Last updated: 9/10/2002
Please email comments on this WWW page to
sedwww@cam.nist.gov.
|