[Date Prev][Date Next][Thread Prev][Thread Next][Search] [Main Index] [Thread Index] [HEASARC Mailing List Archives]

Re: OGIP/93-013 - A list of standard strings for HE missions,instruments & filters.

Jeff Bloch asks:

> We are in the process of thinking about FITS data file formats for data
> from our ALEXIS satellite. We would like to propose a set of keywords for
> the six EUV/ultrasoft x-ray telescopes on our satellite. The detectors and
> filters are fixed for each of the telesopes. The questions we have are:

> 1. What is the maximum number of characters to be considered for each keyword?

It is not clear from the context whether the question is about the
keyword or the value.  The *keyword* is eight characters, with
trailing blanks to fill if necessary, and may be composed only of
upper case letters, numbers, hyphen, and underscore.  The rules for 
the *value* were basically covered by Ian George's comments. Some 
additional considerations are related to the next question.  

> 2. How should we weight considerations of description vs. itemized labeling?
>    For example:  four telescopes have Al/Si/C filters. Should they all
>    have a keyword like "Al/Si/C", or should they have their filter ID number
>    as the keyword, such as "03479 #1", or some hybrid of the two, which would
>    make a longer keyword? ... 

The first FITS paper specified, for the benefit of primitive computers
and FITS readers, that "Originators of FITS tapes may generate longer
strings, but should not expect recipients to decode more than the
first eight characters."  This rule has been interpreted since to mean
that information required (by the software) to read the data must be
available in the first eight characters of a string value.  This
restriction does not apply to values used only for the scientific
interpretation of the data.  In addition, FITS readers with the
eight-character limitation are not as prevalent as they may once have
been.  Nevertheless it is a good practice, though not required, to
provide meaningful and unique information in the first eight
characters of a string value. 

Also, one consideration that is not a FITS issue, and is not covered
by the FITS standard, needs to be noted. The ANSI standard for ASCII 
recommends against the use of the "#" character (hexadecimal 23) for 
international exchange unless there is agreement between sender and
recipient as to its meaning, as this character, along with nine
others, has been set aside as a national character and the meaning may
change from country to country.  It is probably wise to avoid such
characters in the values of keywords. 

> 4. Should we stick to upper case only?

I would agree with a number of the points made by Ian George and 
Jonathan McDowell.  I think that it is usually better not to mix cases 
in a string, but that it is more important that the case structure of
a standard name, such as a unit or chemical element, be properly
reproduced.  Specifically, I would concur with Jonathan's points about
not having separate detector names AF02 and Af02, and about describing
a chemical name as 'MgFl   '. 

				Barry Schlesinger
				NSSDC/NOST FITS Support Office