[fdo-internals] FDO RFC 7 - Add New Methods to FdoClassDefinition

[Frank Warmerdam]

Barbara Zoladek wrote:
> Hi,
> 
> FDO RFC 7 has been posted for review.
> Please let me know if additional information about this enhancement is
> needed.

Barbara,

I've skimmed http://trac.osgeo.org/fdo/wiki/FDORfc7 and have a few
questions.

The property being added to classes is an indication of whether the class
is view based or not.  The motivation given seems to revolve around the
fact that view based classes are likely to be read-only - at least in the
sense that the layout of the class (ie. field definitions) can't be
practically changed.  You also mention that class data (which I take to
mean record field values) may not be editable.

My question is why not address the specific issue of editability more
directly.  That is have a property indicating if the schema is editable,
and perhaps another indicating whether the data fields are editable
(possible on a per field basis?).

I think this avoids ambiguity (after all even a view could potentially
be editable) and it makes a lot more sense for non-RDBMS data sources.

I'd also appreciate it if the RFC could make it clear what the default
value of isViewBased is (presumably false?) and if any open source
providers will be updated during implementation of this RFC.

Incidentally, there are some RDBMS based OGR providers that support views
but the OGR API does not expose information on which layers are actually
views vs. tables.  OGR does attempt to provide information on what layers
are modifiable and which are not, though many drivers (aka providers) do
not implement these tests carefully.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org