Re: AW: Implementation report

See my responses in line.

Le 14/03/2012 13:02, Bailer, Werner a écrit :
> Hi Thierry,
>
>>> I have discussed with Werner about the general test cases and input data.
>> It is not possible to declare input data for these test cases that is valid for
>> both modes of implementation.
>>
>> Can we have test cases for each modes ?  We need to explain this.
>
> There are the following issues with these test cases:
> (1) some need no input
> (2) for those that need input, this mainly concerns the parameters to be passed, but not a RDF source file
> (3) for G1, the JSON output depends on the modes implemented
>
> Ad (1)
> This applies to
> - G1 (correct as is)
> - for G2 and G3 this could be arbitrary source metadata files, but the abnormality being tested (NULL, duplicates) is in the parameter list being passed to the call, not in the source file.
> - G5 (correct as is, multiple source files must be present, but they can be arbitrary)
> - G6 (correct as is, by definition of the test case there is no source)
> - G6 (correct as is, by definition of the test case the specified source is not there)


I have added a note about these remarks
see
http://www.w3.org/2008/WebVideo/Annotations/drafts/API/implementation-report.html

>
> A general comment to the sources files in the G section: why do they all have xmpmeta as root element, although just DC properties are used?

Good catch.
All files are updated.


> Ad (2)
> Maybe it would make sense to add a column specifying the relevant parameters of the method call for each of the cases?
> - G1 n/a
> - G2: propertyNames=["fakeIdentifier"]
> - G3: propertyNames=NULL
> - G4: propertyNames=["identifier","identifier"]
> - G5: code from example in 4.1.2 could be referenced, though the exact URLs of the metadata sources will depend on which metadata files are accessible by the implementation
> - G6: metadataSources = NULL;
> - G7: metadataSources = new MetadataSource[1]; metadataSources[0] = new MetadataSource("http://a.non.existing.url/file.xml","dc");

I have added a new column in the general testcases table with these 
parameters of the method call. Still need to discuss case G5.


> Ad (3)
> - G1: we do not require an implementation to cover both modes, thus we can only require that the JSON repsonse contains at least either 1 or 2


Let's discuss how we can solve this case.


> Best regards,
> Werner
>
>> -----Ursprüngliche Nachricht-----
>> Von: Thierry MICHEL [mailto:tmichel@w3.org]
>> Gesendet: Mittwoch, 14. März 2012 08:05
>> An: Florian Stegmaier
>> Cc: public-media-annotation@w3.org
>> Betreff: Re: Implementation report
>>
>>
>>
>> Le 13/03/2012 18:52, Florian Stegmaier a écrit :
>>> Thank you very much!
>>>
>>> Their full names are Thomas Kurz and Sebastian Schaffert both work at
>> Salzburg Research Forschungsgesellschaft m.b.H.
>>
>> I have added their name in the authors section.
>> http://www.w3.org/2008/WebVideo/Annotations/drafts/API/implementati
>> on-report.html
>>
>>
>>>
>>> I am not quiet sure if he is able to fix this issue. I will ask him tomorrow and
>> come back to you,
>>
>> Please do.
>>
>>
>> otherwise i would suggest to skip these to move on asap!
>>
>> I am afraid we can't move forward if we don't have two implementations
>> passes for each test.
>>
>>>
>>> I have discussed with Werner about the general test cases and input data.
>> It is not possible to declare input data for these test cases that is valid for
>> both modes of implementation.
>>
>> Can we have test cases for each modes ?  We need to explain this.
>>
>>
>>>
>>> Unfortunately there was not much attention today - there is also the issue
>> with the XPath expressions in DC that should be adressed.
>>
>> Could you let me know the issue with the XPath expressions in DC that
>> should be addressed ?
>>
>>
>> Best
>>
>> Thierry.
>>
>>
>>> _____________________________
>>>
>>> Am 13.03.2012 um 18:44 schrieb Thierry MICHEL:
>>>
>>>> Florian,
>>>>
>>>> Excellent work !
>>>>
>>>> And please thank Thomas and Sebastian for their active participation.
>>>> Please give me their full name and I will add their name in the ack section.
>>>>
>>>>
>>>> We are now close to the CR exit criteria.
>>>>
>>>> Remains a few Issues:
>>>> Dublin Core:
>>>> D8-b: implementation 1 fails
>>>>
>>>> YouTube
>>>> Y8-b: implementation 1 fails
>>>> Y12-b: implementation 1 fails
>>>>
>>>>
>>>>
>>>> Do you think the Salzburg Research implementation can be fixed to pass
>> these 3 tests ?
>>>>
>>>>
>>>> Could you please provide the G1, G5,G6, G7 General behaviour tests.
>> These are currently not linked.
>>>>
>>>> Again thanks for your great work, in these days when the Group is very
>> silent.
>>>> Daniel, there was no telecon setup this week, can you please make sure
>> to schedule one next week?
>>>>
>>>> Best,
>>>>
>>>> Thierry
>>>>
>>>>
>>>>
>>>> Le 13/03/2012 10:35, Florian Stegmaier a écrit :
>>>>> Hi Thierry, all,
>>>>>
>>>>> please find enclosed our implementation report. Thanks again to
>> Thomas and Sebastian from Salzburg for their efforts in supporting us.
>>>>>
>>>>> @Thierry: What are the next steps now?
>>>>>
>>>>> Cheers.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _____________________________
>>>>> Dipl. Inf. Florian Stegmaier
>>>>> Chair of Distributed Information Systems
>>>>> University of Passau
>>>>> Innstr. 43
>>>>> 94032 Passau
>>>>>
>>>>> Room 248 ITZ
>>>>>
>>>>> Tel.: +49 851 509 3063
>>>>> Fax: +49 851 509 3062
>>>>>
>>>>> stegmai@dimis.fim.uni-passau.de
>>>>> https://www.dimis.fim.uni-passau.de/iris/
>>>>> http://twitter.com/fstegmai
>>>>> _____________________________
>>>>>
>>>
>>>
>

Received on Tuesday, 20 March 2012 09:47:03 UTC