On 12/05/18 17:44, Andy Seaborne wrote:
On 12/05/18 17:26, Dave Reynolds wrote:
On 11/05/18 20:35, ajs6f wrote:
On May 9, 2018, at 3:31 AM, Dave Reynolds
<[email protected]> wrote:
On 08/05/18 16:55, ajs6f wrote:
Forking a thread off to dev@:
Do we have a global policy about where null is accepted as a
wildcard? I know it works in at least some places...
I would love to (over an appropriate period of time and with lots
of warnings and deprecation and so forth) stop letting it be a
wildcard and require code to use the actual wildcard objects.
-1
We have a huge amount of code that assumes null-as-wildcard as
provided for in the public RDF API and I doubt we're alone. There
would no chance of porting all that through a normal deprecation cycle.
Fair enough, Dave. Just to be clear, are you saying as much as that
new methods in the public API should always accept null-as-wildcard,
or just that we oughtn't be changing extant methods now?
The latter. I've no particular attachment to null-as-wildcard as a
design pattern so new APIs (at least ones which aren't just tweaks on
the existing API) could go a different route.
I'm against having a mixed design in one API (e.g. new methods on Graph,
or elsewhere in related classes/interfaces). Design uniformity matters
and the current SPI is what it is.
Agreed, I should have clearer with my "at least ones which aren't just
tweaks".
Dave
