Getting Started

Documentation

Glish

Learn More

Programming

Contact Us

Version 1.9 Build 1367

News	FAQ
Search	Home

Installation

Many aspects of the installation process have been discussed in passing in the preceding sections. However, this topic is of sufficient importance that it deserves separate attention.

In the first instance an end user site will obtain a copy of aips++ from their local aips++ consortium member, via anonymous ftp or tape. Many sites will be able to copy the aips++ binaries directly, thereby saving them the trouble of recompiling the system. These binaries must therefore be ``exportable'' in the sense that they contain no hard-coded site-specific information. The installation would thus be completed by encoding the site-specific information into ASCII databases such as the $AIPS/hosts and $AIPS/$ARCH/$VERS/$SITE/aipsrc files.

Inevitably, many sites will not be able to take advantage of the binary distribution, either because they have incompatible hardware, or incompatible releases of their operating system or utilities. These sites will have to rebuild aips++ from its source code, and we must ensure that the aips++ system is completely constructible from the $AIPS/$VERS/code subtree. This includes architecture-specific procedure scripts (if any). Apart from defining site-specific parameters in databases like those mentioned above, the installation would first involve editing a file used by the aips++ makefiles, $AIPS/$ARCH/$VERS/$SITE/sysconf, containing a more detailed description of the system. This might include the location of various public domain libraries or utilities required by aips++ and which the site had already installed, but in a place different to where the makefiles expect to find them by default. Template versions of this ASCII descriptor file should contain defaults which the installer may change as required by using a text editor. After editing the sysconf file, a single recursive make should rebuild the whole of aips++ .