single
user installation |
multiuser
or server installation |
If you will be the sole user on a single computer, create a folder for BIRCH within your $HOME directory Example: /home/susan/birch |
If two or more users will use BIRCH, create a separate user account solely for BIRCH, and install BIRCH in the $HOME directory for that account. Make sure that $HOME directory is world-readable, world-searchable. Examples: /home/birch /usr/local/birch |
Notes: Do not simply install BIRCH in your $HOME directory. BIRCH consists of more than a dozen subdirectories that could potentially over-write already-existing directories of the same name. Most importantly, BIRCH has its own public_html directory, which is where web documentation is located. Upon installing, BIRCH would overwrite your existing $HOME/public_html directory. As well, the public_html directory is deleted on each BIRCH update. All of these problems are avoided by simply creating a separate folder for BIRCH. |
Notes: Actually, even for multiple users, you could just create a birch subdirectory within a user's $HOME directory, as done in a single-user install. However, there are many advantages, regard to organizational issues such as system administration and backups, to create an account solely for BIRCH. BIRCH is deliberately designed so that the BIRCH Administrator does not need root privileges. Even if the person who administers BIRCH has root access, there should be a separate user account for the BIRCH Administrator. This limits the severity of potential mistakes made by the BIRCH administrator. As well, if BIRCH files and directories belonged to root, the severity of potential security holes would be far greater than if they are owned by a regular user account. Ask your system administrator to create a separate account for administering the BIRCH system. For example, the userid might be 'birch' and the home directory might be '/home/birch'. This account is simply a regular user account with no special privileges. If administration of BIRCH is to be shared among several people, it may be good to also create a group called 'birch', to which all BIRCH administrators belong. |
Python 2.4.3 or later Required |
standard in most Linux and Solaris distributions http://www.python.org/ The 'python' command must be in the users' $path. Test python by typing 'python -V' at the command line. If Python is installed, the -V option will print out the version of python eg. |
Java 1.6 or
later Required |
standard on most Unix and Linux systems http://java.sun.com/j2se/ Test java by typing 'java -version' at the command line. If Java is installed and present in the system $PATH, you should get a message like java version "1.6.0_03" If you get this message, no further steps are necessary. On the other hand, if you get a 'Command not found' message, or a message like this java version "1.4.2" either Java is not on your system, or is not in the search path. gij is NOT Java! For full instructions on setting up Java on your system, inst.java.html. |
/bin/csh Required |
Surprisingly, some
systems (eg. Fedora) do NOT have a /bin/csh, by
default. This is
odd, since it greatly limits compatability other Unix systems. In any
case, at least some systems do not have csh, or even tcsh, by
default. To find out whether or not you have csh, type 'which csh'. If
you don't have csh, tcsh is a good substitute (which tcsh). If you have
neither of the two, I suggest installing tcsh. In Fedora Linux, it can
be added in the GNOME "Add or Install Software" tool. Once tcsh is
installed in /bin, do the following: cd /bin ln -s tcsh csh This command creates a symbolic link called 'csh', which points to tcsh. Now, when a csh script executes, it will be run by tcsh. tcsh has replaced csh on most recent Unix and Linux systems. |
Java3D Optional |
The TIGR Microarray programs, MIDAS and MEV will run without Java3D, but certain analytical methods such as Principal Component Analysis require Java3D. |