### 3 The difficult way

Given the lack of detailed documentation on ORAC-DR it is difficult to not see it as a black-box. Most people who use it is as their preferred method of data reduction are likely to have spent some time with somebody who knows a lot about it, or to have dug down into the primitive code to see what is actually done on a step by step basis. The new xoracdr interface will to some extent make the reduction less opaque, it can step you through the recipes. However most astronomers, being by nature suspicious and distrustful, are likely to want to reduce their data at least once in a step by step manner. The rest of the cookbook outlines how to do this. It is by no means a waste of space for the ORAC-DR user, ORAC-DR calls on the Surf routines to reduce the data.

First log in on a unix workstation and create a directory where you want to store your reduced data. Next type

% surf
% kappa
% figaro
% convert

This starts up the SCUBA software and the main Starlink packages needed for the data reduction.

The next thing you will need to do is to find the data and create logs of your observing run, so that you know what scan numbers to reduce.

#### 3.1 How to find and access SCUBA data?

All SCUBA data are stored on disk at JAC and also for a shorter period at JCMT. The data are initially stored on the summit VAX, and copied to a unix disk (mounted as /jcmtarchive) after each observation is completed. Each night a new subdirectory is created using the convention of year, month and UT date, i.e. 19980215, would contain data for the night between the 14th and 15th of February 1998. As of now data obtained on for example the evening shift of February 15 are stored on the UNIX disk in

/jcmtarchive/19980215

and copied to Hilo the following day, where they reside in a protected UNIX directory. If you are a visiting astronomer, it may be wise to copy all your files into your local directory, so that you later can transfer them to your home institution. Another way to obtain SCUBA data is to search the JCMT archive. If you find what you are looking for, please mark and retrieve the files and within an hour you can start reducing data. Since the source files are not tagged to calibration information, you will have to do an additional search for the date or dates that the files originate from and retrieve skydips, pointing and calibration data, so that you can calibrate your data sets.

In the following we assume that you have copied the files into your local directory called /home/user/scubadata. To start the data reduction you will therefore need to set up some environmental variables and get a listing of your data.

Here yyyy stands for the year, mm is the month (as a numeral) and dd is the UT date for the night you want to reduce, i.e. 19980401 would be the night between March 31st and April 1st in 1998. For the best performance you should try to ensure that your working directory is on a local disk. At the very least, you should make sure that HDS_SCRATCH is local since this is where temporary files will be written by Starlink applications.
Here the summary information about all the maps in our directory is piped into a text file, which we called Dec8.maps. You can choose any name you want. This is a simple ASCII-file, which you can edit and print. If you want to be systematic in your data reduction, you can edit the file to include your reduction notes. The same way one can make summaries of photometry (photsum) and pointsum, to get a log of the pointing, which you will need if you plan to correct your maps for pointing drifts. You may find it useful to include a complete log of the night with sculog (or obssum), if you want to make sure that you know when the telescope was focused and when noise measurements were done, although the latter you can easily find with scunoise.