experimental schema: format.sourcestatus

[Steven Bird]

Last week Baden Hughes presented a new encoding scheme called source
status.  Here are some initial comments:

> Pre-Compiled Binary

or just "binary"?

> Requires Compilation
> Requires Make
> Wrapped Installation

These three are closely related - a build is required, and the
difference is in how much work the person has to do.

> Script

So a simple starting point here would be to have a three-way
distinction between binary, interpreted and compiled.

[Aside: In all three cases, other packages may need to be downloaded,
built and installed before the software can be run, and these will
need to be documented using the relation.requires element/refinement.
Presumably we won't bother specifying that a C compiler is required
for a resource that is specified as being in the C language, unless a
particular compiler/version is required.]

Notice that the distinction between interpreted and compiled is
largely predictable from the source language, and that the source code
might not actually be provided.  Therefore, we want to focus not on the
source code, but the nature of the distribution (format.distribution?).
Obviously, this now applies to data as well as software, since data can
come in binary or source forms, with our without wrapping.

The distribution methods include archives (tar, zip, rpm) which may be
compressed, and may be self-extracting or require other software.  The
self-extracting kind might actually manage the download and
registration process, as in the case of the CSLU toolkit.  To some
extent, the distribution method is predicable from the MIME type of
the file, which weakens the case for special treatment of distribution
types.

An orthogonal issue is size: can I download this over a modem line?

Anyway, to move things forward here, we may need to do some more study of
end-user needs.

-Steven

--
Steven Bird        Email: <sb at cs.mu.oz.au>  Web: http://www.cs.mu.oz.au/~sb/
A/Prof, Dept of Computer Science, University of Melbourne, Vic 3010, AUSTRALIA
Senior Research Assoc, Linguistic Data Consortium, University of Pennsylvania