Re: [xcbf] Vendor Specific Formats

[Bancroft Scott <baos@oss.com>]

On Fri, 10 May 2002, Ed Day wrote:

> If the goal is to get those in the BioAPI group to "buy-in" to this new
> standard, then I think the goal should not only be 100% backward
> compatibility, but ease-of-use as well.  My initial impression of the BioAPI
> document was that it was an easy-to-follow document and I had a relatively
> good understanding of the structure of the data.  When I then went to the
> ASN.1 document, I had a hard time relating what was in the BioAPI C
> structures to the ASN.1 specification.
>
> So it would seem to me like the suggestion to use record types like the
> following makes the most sense:
>
> > > >      BioAPI-BIR-BIOMETRIC-DATA-FORMAT ::= SEQUENCE {
> > > >         formatOwner     INTEGER (0..65535),
> > > >         formatID        INTEGER (0..65535)
>
> It should be possible to create almost a one-to-one mapping of fields in the
> BioAPI structure to fields in an ASN.1 type.  This could be done at the
> beginning of the specification with all of the more complex stuff concerning
> information objects and the like pushed down to where it would of be out of
> the way of the casual reader.
>
> The other advantage is ease-of-use.  Using an ASN.1 compiler, one could
> create the BIOAPI structure format directly from the ASN.1 source, making it
> possible to read a header directly into a variable of the generated header
> structure.  With a few more lines of code, one could then create the XML
> document corresponding to the BIR.

I entirely agree with all of the above.

Bancroft