Re: What is source code?
Date: Mon, 19 Apr 2004 12:42:23 -0500
"Laconic2" <laconic2_at_comcast.net> wrote in message
> In another topic, the question got raised whether source code is metadata.
> There is a way in which source code might be considered "metaprocess". It
> tells the compiler what to do, in order that the compiler will tell the
> target machine what to do.
One can have metadata about processes as well as about data attributes, right? The info that an IDE collects in order to generate classes - what do you call that? What do you call a constraint placed on a data field? What if that constraint is specified to a Java compiler instead of a proprietary RDBMS compiler or interpreter?
If metadata is only about data attributes for persistent data, then that seems like a very narrow definition, but I prefer a narrow definition to none. So, if we were to enumerate some metadata, it would include the name of a field, but not the name of a process? It would include specs about a field, but not specs related to a "virtual field"? or related to a virtual field and not to a subroutine to aggregate the information for the virtual field?
Defining metadata as "data about data" doesn't narrow that down. I store all code as "data" -- it isn't any less data than a set of digital films would be, right? Metadata is also data. That is why typically data about the data that describes data is called "metadata" too. A set of VIEWS -- are those metadata?
This is a matter of definition, but I think a definition worth pursuing to get some common use of it. If source code is not metadata, is there a super-type for source code or is it a root concept? --dawn Received on Mon Apr 19 2004 - 19:42:23 CEST