content, format, ctype, or xtype ?

[Mark Taylor]

On Tue, 12 May 2009, Rob Seaman wrote:

> - Or isn't it more appropriate to burden the programmers with the requirement
> to recognize a narrow range of formats?  After all, reasonable code will have
> to validate the inputs anyway - we can't just rely on somebody's metadata to
> be utterly truthful.  Does a string parse as a number - then it is floating
> point.  Does it parse as ISO-8601?

that approach can work OK if you're talking about a single value.
But if you're talking about characterising a column of a table, 
which is the case here, it's much more problematic, since you may need
to read some, much or all of the data before you know what type it has.
If it's not obvious to you why this is problematic I can elaborate,
but basically it doesn't just end up being a (programming) burden on
the programmer, it ends up being a performance burden on the user,
and possibly a resource burden on some server as well.

-- 
Mark Taylor   Astronomical Programmer   Physics, Bristol University, UK
m.b.taylor at bris.ac.uk +44-117-928-8776 http://www.star.bris.ac.uk/~mbt/