AML, a new format for astronomy based on SGML

[Peter Bunclark]

You can't replace FITS unless you do handle large binary images. 
Perhaps you could use FITS as your image format, which would be
referenced from AML.

Peter.

Damien wrote:
> 
> The main goal of FITS was to provide a standard exchange format
> for astronomy. With some advantages, especially its flexibility,
> it really became used by everybody in astronomy.
> However, it is still limited to some data, and is
> completly ignoring (as far as I know) astronomical articles or
> data on single astronomical objects (amongst many other).
> Also, it has become a bit outdated as an exchange format (strange
> 80 characters lines in the header, no structure, ...)
> 
> This is why I started the creation of a new exchange format
> for astronomy, within my thesis on information retrieval in
> astronomy. It is based on SGML and as such is much more flexible,
> readable by any SGML/XML browser/parser, and is appropriate
> to the web.
> AML is now supporting 4 astronomical data types, but is in
> no way restrictive: astronomical objects, articles, tables
> and images. Because only characters may be used in SGML,
> and because it is easier to manage metadata this way, the big binary
> files (such as tables or images) are external to the AML files,
> and are referenced in AML as images are referenced in HTML.
> I began the creation of an AML browser on the web (a Java applet),
> so that anyone can easily read these files.
> 
> There are of course some interactions between FITS and AML,
> especially for astronomical images: as everybody is now using
> FITS, the translation in the two directions must be easy.
> To enable an easy translation, I just copied the FITS keywords
> and put them into AML in an SGML-way. However, some improvements
> may be easy to do on that basis, such as classifying FITS
> header information (on the image, on the observations, and so on).
> 
> As I hope AML could become a standard format, I cannot
> define it alone, and I'm now looking for people interested in
> its definition. Maybe it could be seen not as a rival, but as
> a new version of FITS...
> 
> More information on AML is available at:
> http://www.infm.ulst.ac.uk/~damien/these/
> 
> Damien Guillaume