The tar file contains five files:
Distributed in the tar file is README_jgofs-V2.2_INSTALL_UPGRADE. It provides instructions for upgrading a JGOFS data server from V2.00 to V2.2. (V2.1 was released but should not have been used.)
If what you want to do is create a new JGOFS data server, rather than upgrade an existing one, use the instructions in file README_jgofs-V2.2_INSTALL_NEW, also distributed in this tar file.
A note about the "readme.*" files in this directory and in the src tree: Those files contain detailed documentation about the sub-modules assembled into this release. For the upgrade procedure detailed in the present file, you will not need to consult them. They are available in case you ever do need detailed information about the sub-modules.
The installation/upgrade tar file should be unpacked in either an otherwise empty directory (for a new server) or in the root directory of an existing V2.00 data server (for an upgrade). If you are using a version of the software below V2.0, you must first upgrade to V2.00. However, if this is a new installation, you can install V2.2 immediately.
The procedure up to the point of doing site customization consists of a single step, for creating a new server, or two steps, for upgrading. The second step in upgrading is to run a second perl script, which will have been unpacked by the first step.
The reason this is a separate step is that it will wipe out the existing V2.00 server, and we want to encourage verifying (from the log file) that all is well before beginning the demolition and reconstruction. (The first script will have made a backup copy of the existing server, though, in case anything does go wrong.)
Extensive information, in English, about the progress of the procedure is reported to the screen, and additional detail is provided in the log files.
The degree of automation supporting carrying forward the site customizations, and recovery if something goes wrong during the installing/upgrading, is the same as for V2.00. The details are in the README* files.
The tar file contains five files:
The first two are the instructions for installing a new data server and for upgrading an existing V2.00 one. They are separate because the description of how to customize is necessarily more complicated for an upgrade. The third file is the script that either does a fresh installation or identifies customizations in an existing server. The last two files contain both the actual V2.1 distribution and additional scripts and other files for use during upgrading; they aren't for direct use by the user; the script handles them.
The installation/upgrade tar file should be unpacked in either an otherwise empty directory (for a new server) or in the root directory of an existing V2.00 data server (for an upgrade). Step (1) of either process is to run the perl script, jgofs-V2.1_1install_script.pl. The script will determine whether it's an upgrade or a new installation. If the directory is neither new nor a server root, the script will complain and quit, so running the script won't damage anything if it's started in an inappropriate directory.
The procedure up to the point of doing site customization consists of a single step, for creating a new server, or two steps, for upgrading. The second step in upgrading is to run a second perl script, which will have been unpacked by the first step:
The reason this is a separate step is that it will wipe out the existing V2.00 server, and we want to encourage verifying (from the log file) that all is well before beginning the demolition and reconstruction. (The first script will have made a backup copy of the existing server, though, in case anything does go wrong.)
Extensive information, in English, about the progress of the procedure is reported to the screen, and additional detail is provided in the log files.
The degree of automation supporting carrying forward the site customizations, and recovery if something goes wrong during the installing/upgrading, is the same as for V2.00. The details are in the README* files.
The first two are the instructions for installing a new data server and for upgrading an existing one. They are separate because the description of how to customize is necessarily more complicated for an upgrade. The third file is the script that either does a fresh installation or identifies customizations in an existing server. The last two files contain both the actual V2.00 distribution and additional scripts and other files for use during upgrading; they aren't for direct use by the user; the script handles them.
New features of the release require Perl: build.pl replaces build, build-env.pl replaces build-env.
New outer.c paradigm, to support possible per-method customization.
New handling of makefile.jgofs (the makefile to build the library), to protect against accidentally clobbering a locally customized makefile.jgofs; requires an extra step in the installation.
A new subdirectory called src/images is added to include the gif images used by the methods for the different buttons.
Changed default.h buffer sizes for TOKEN, DATUMSIZE, and OUTBUFSIZE.
New installation instructions; separate instructions for fresh installation and for upgrade.
Reorganize the defgb distribution by placing common subroutines in the lib/ directory and library file. Move the "brev" version of outer to the standard location so all methods use the same one.
Add a preliminary release of the Matlab utility, loadjg, to allow reading of JGOFS data directly by Matlab. [Implimentation not yet complete. The makefile needs additional work.]
Improve the installation instructions.
Edit the parse_path_info doc file.
Update default.h and outer.h with larger ATTRSIZE. Change from 40 to 80.
Replace jdb.c version 1.1c with version 1.1e. Improved error handling with add conditional compilations for debug options.
Replace listvar.c with Version 1.3a 16 Oct 2003. It explicitly exists with status of 0 after a normal execution.
Change DATUMSIZE from 80 to 240 in outer.h
Remove the older, unused, and archived files out of the distribution.
There is a new release of list.c, V1.6b. Several new switches were added and some *.h files are expected to be in the include subdirectory.
Change MIN_DATUMSIZE 120 to 240 in list.h.
HTTCP.c V1.4 dated 3 December 2003 replaces sys_errlist and sys_nerr with strerro. This fixes a problem which occured in Linux compilations. Add version definition to the code.
Prior to Releaase V1.5a, the JGOFS software was only available from MIT. The release history for this version of the software follows.
The U.S. GLOBEC Georges Bank Program uses the JGOFS software to serve our diverse data sets via the World Wide Web. This is the same software used by the U.S. JGOFS Program to serve their data. Chris Hammond, my counterpart at the JGOFS program; Glenn Flierl, one of the original developers of the JGOFS software; Warren Sass, a programmer at WHOI; and myself have been working on enhancements to the JGOFS software for many months. The results of these efforts are available in the form of a new release, Version 1.5, of the JGOFS software.
The enhancements made in Version 1.5 include:
Alternatively, the width of each column can be extended by adding an appropriate number of underscores to the field names. These are stripped off by the software, after they are counted to determine the field width attribute. For example:
field name: cast can be extended to a field width of 6 by either of the following: cast__ cast[width=6]
Several programming bugs are also being addressed in this new release, including better sensing of non-numerical data within a data value and improved handling to data buffers to minimize the risk of "buffer overflows."
While everyone who browses and accesses the data will see the improvements in the look of the directory, it is only those sites who are serving data that need to "do anything" about this release. You can access the new software at the install software page at URL
and follow the instructions on this page. You should know that all of your existing methods will continue to work as is with this new release, but before you recompile any of your methods with the new software, you need to make some changes to your code. Contact the Data Management Office (DMO) for instructions and help when you are ready to recompile and relink your own methods.
A few data serving sites that take advantage of the local "info" feature, will also need to run a utility to convert the contents of their descript.html files into .info files, one for each object name. If you need this script, please contact the DMO.
We plan to prepare new releases of the JGOFS software, approximately every four months. In addition to futher improvements in the methods, we also plan to enhance the data analysis and display functionality in upcoming releases, based on your feedback and suggestions and available funding.
So, if you use data served by the JGOFS software you should see several improvements on how it looks and performs. If you serve data to others via this software, you need to get the new software release and install it on your system. As always, I am available to help you with the installation process. Please contact the Data Management Office when you are ready to install the new software.