Hi,

I briefly applied the solution (addjars-maven-plugin) from the Simple DDR
implementation to Reza's quick draft. It compiles, but the unit test class
shows, it's nowhere near W3C compliance.
DeviceMapW3CSimple simple = new DeviceMapW3CSimple();
is a chimera, not a proper use of W3C Simple DDR. See the web based
examples something like
Service simple = ServiceFactory.newService(...);
would be the right way to do it.

Happy to help at some point, but until the data changes in ways that force
significant refactoring of all current client libraries/implementations
there is a valid and working W3C DDR implementation, so why not release
that working implementation now and look into improving a similar wrapper
then.

Werner

On Tue, Jan 6, 2015 at 8:51 PM, Werner Keil <[email protected]> wrote:

> After the latest commits (both to /classifier/test code and POMs on top)
> happy to see,
> mvn clean install on /devicemap/java works again.
>
> On Tue, Jan 6, 2015 at 8:32 PM, Werner Keil <[email protected]> wrote:
>
>> Please see how the "stable" usable W3C implementation did this for a long
>> time, the Jenkins builds also worked till our Jenkins instance faded:
>> https://svn.apache.org/repos/asf/devicemap/trunk/devicemap/java/simpleddr/
>>
>> The W3C may no longer maintain a code repo (if so it probably was CVS at
>> the time, SVN at most;-) but the original sources (not recompiled, JavaDoc
>> and headers with @author etc. are all intact) were rescued here:
>> https://github.com/fnk/w3c-ddr/tree/master/src/main/java/org/w3c/ddr/simple
>>
>> Bertrand also did a similar "rescue" to the OS state of WURFL, so if
>> someone wanted to see or recompile it from the source, pointing to this
>> repo or creating a SRC-JAR with Maven and putting it somewhere is no
>> problem. Since we discussed earlier, that we shall not build it from there
>> and the binary JAR is available, Maven builds of the so far working W3C
>> implementations use a special Maven plugin (see the POM) that automatically
>> creates a maven dependency from that lib JAR.
>>
>> You don't need to run any mvn install "magic" any more, that was an early
>> approach by OpenDDR I fixed after the implementation was contributed to
>> DeviceMap.
>> Please revisit again if necessary, but I believe that meets the Apache
>> requirements for 3rd party libraries and W3C is obviously used a lot also
>> by other projects.
>>
>> Werner
>>
>> On Tue, Jan 6, 2015 at 7:30 PM, Bertrand Delacretaz <
>> [email protected]> wrote:
>>
>>> On Tue, Jan 6, 2015 at 7:27 PM, Reza Naghibi
>>> <[email protected]> wrote:
>>> > ...This is a data project, so we should look to standardize our data
>>> format, not
>>> > an API ontop of the data...
>>>
>>> Totally agree with that, but having usable clients doesn't hurt - as
>>> long as we have volunteers to maintain those clients, and as long as
>>> they don't interfere with the main task of maintaining and releasing
>>> the data.
>>>
>>> -Bertrand
>>>
>>
>>
>

Reply via email to