CD-Rom Problems

Read the instructions which are on the CD.

The first thing to say is that this sort of application is not really that for which web browsers were designed. Operation of this search facility should be regarded as rather experimental, though we have successfully run this software on a wide variety of UNIX machines (Solaris, Tru64, IRIX, Linux (Redhat & SuSE)).

System requirements

A UNIX/Linux operating system
write permission on /tmp
perl at a location in the user's $PATH. (versions 5.002 or later have been tested, though earlier versions probably work.)
netscape web browser (Only netscape has been tested, other web browsers could be tried, however this would require you to copy the scripts which run the code to you local disk and edit them).
The environment variable $PWD to be set to the current working directory (should be the default on most/all UNIX/Linux systems).

Known potential problems

You may only run one instance of 2QZ.sh per machine. This is because log files are written to /tmp and multiple versions of 2QZ.sh would attempt to overwrite each other's search results. Secondly, if there is already a version of the perl CGI servlet running on your machine the port number 8008 (the default port) will already be in use. You could possibly get around this restriction by editting the scripts to change the file names and port numbers, but it is a bit complicated and we do not provide detailed instructions.

Netscape gives the error:

(111) Connection refused

When you run the ./2QZ.sh script, the command shell prompt should not immediately return. One possible cause of this happening is that you already have a netscape running. Close the existing netscape and try again.

If the command prompt has not returned, Netscape may be attempting to connect to the search script via an internet proxy. Obviously there is no point in doing this, since it would then be connecting to itself via another server. This is likely to happen if your computer operates behind an internet firewall or is using a webcache. We have so far only found this to cause a problem on Linux installations.

Only plain text is presented to the screen starting with
This problem is known to occur on newer versions of Mozilla or Netscape (version > 7.0). It is caused by a MIME type having been set wrong in one of the scripts on the CD. Most browsers will ignore this mistake and work fine, but a couple have been seen to be very strict. Depending on your broswer, you may be able to edit the MIME type settings to force MIME "text/plain" to be rendered as HTML, but this is unlikely.

If netscape starts, but either loads no page or your usual home page instead of the 2QZ data, try simply typing iin the following URL:
http://127.0.0.1:8008/mnt/cdrom/html/index.html
This is assuming your CD is mounted at /mnt/cdrom. If now, change the path in the URL accordingly.

The following is drawn directly from the documentation for the perl HTTP server which we are using (CGIservlet-1.2.pl). We have not observed this problem on any systems we have tested.

An odd server side network error is reported by Netscape when a Post is initiated from a Javascript Submit of a <FORM>. This was found on Red Hat 6.1 Linux with perl 5.00503, 5.00503 and 5.6.0. But not on IRIX or Red Hat 5.0.

Installing to Hard Disc

If none of the above helps, it is most likely that you are not going to get the search engine running directly from the CD-Rom. You could try installing onto the hard disk, but it is unclear that this is likely to solve any real problems. Running from the hard disc should also result in modestly better performance.

Method 1:

If you have sufficient disc space, you could copy the entire CD to a directory on your hard disc and run from there. e.g.,
cp -r /path/to/cdrom/* /path/to/new/location/on/harddisc
As long as you continue to invoke ./2QZ.sh from the directory in which it resides, everything should work as on the CD.

Method 2:

If you wish to retain the data on the CD, or run the 2QZ.sh script from anywhere other than its default location, you need to specify some path names in the scripts. This is considered an option for `expert users' only, but it should be straightforward for anyone familar with the UNIX enviromment. The paths that are likely to need editing are:

in 2QZ.sh

location of CGIservlet-1.2.pl. This may be absolute or relative with respect to the location of 2QZ.sh. (Currently ./html/CGIservlet-1.2.pl)
location of index.html. This must be an absolute, not relative path. (Currently $PWD/html/index.html)
If you wish, you could change /tmp/2qz.log in order to personalise its location.
If you wish, you could also change /tmp/2QZ.log and /tmp/2QZ.out.html both here and in cat_search_www2.pl in order to personalise their locations.

in html/searcher.cgi

The location of cat_search_www2.pl must be given relative to the path whence CGIservlet-1.2.pl was run. (Currently html/cat_search_www2.pl, since it is in the directory html/ on the CD.)

in cat_search_www2.pl

If you wish, you could also change /tmp/2QZ.out and /tmp/2QZ.out.html both here and in 2QZ.sh in order to personalise their locations. These two paths are specified by the OUT and HTML_OUT filehandles. Each files location is also hard-coded in at once. Sorry about that!
The location of all the data needs to be specified either absolutely or relative to the location of cat_search_www2.pl. These variables appear in cat_search_www2.pl as fitsname, gifname, psname, fitsnameb, gifnameb, psnameb and dssname.
Set location of full catalogue relative to location of 2QZ.sh. (Currently set by the line $catfile="cat/2QZ_10k.txt"; )