@Bertrand, I know you may be busy, but as you provided some (more license related) input here only a day ago or so, what is your or Adobe's opinion on the question of full or at least significant enough W3C compliance in the API(s)?
To this list http://www.w3.org/2005/MWI/DDWG/drafts/api/test-report.html you can of course add the MaDDR project and OpenDDR as of the GitHub project (plus a couple of derived projects like the Play/Scala one, etc.) I am not sure, if the Telefonica Open Source project is still alive? Certainly if somebody at Telefonica in a relevant position was interested to get involved in DeviceMap, it would also be good to offer them a familiar environment (or an API that if they still have stuff in production based on "MyMobileWeb" and I helped enough of the biggest telcos to tell you they DO, even if they may not admit to it[?]) or incentive to migrate to DeviceMap. So aside from XML files that don't matter so much if they can be provided in a different form in the future, do we want to offer a W3C compliant API for Java, .NET and maybe other languages, or shall we forget about it and do something completely different?[?] Cheers, Werner On Wed, Jul 9, 2014 at 6:48 PM, Werner Keil <[email protected]> wrote: > If you compare something like > > https://docs.deviceatlas.com/apis/enterprise/java/2.0.1/DeviceApiJavadoc/mobi/mtld/da/Property.html > with > http://www.w3.org/TR/DDR-Simple-API/#sec-PropertyValue > > or > > https://docs.deviceatlas.com/apis/enterprise/java/2.0.1/DeviceApiJavadoc/mobi/mtld/da/PropertyName.html > with > http://www.w3.org/TR/DDR-Simple-API/#sec-PropertyName > > Everyone who knows Java well enough will understand, the "pattern" layed > by W3C is respected by DeviceAtlas API. > > If you want to reinvent the wheel, or call things completely different, go > ahead, but you must not call it W3C compliant any more then. > > It's not even about directly implementing the W3C JAR or its interrfaces, > but a "behavioral" compatibility (W3C doesn't have a TCK like the JCP does, > but you see, APIs by DeviceAtlas and a few other vendors were at least > reviewed once and found compliant: > http://www.w3.org/2005/MWI/DDWG/drafts/api/test-report.html > > How does our "new" API look like in this matrix?[?] > > The OpenDDR contribution passes every single one because it directly > implements the "Spec" JAR by W3C. > > If we care about this compliance, the "new" API isn't complete or 1.x > ready until it can provide a similar compatibility (regardless of the name, > but at least certain methods should be comparable) > > Werner > > > > On Wed, Jul 9, 2014 at 6:36 PM, Werner Keil <[email protected]> wrote: > >> @Eberhard, is it possible to generate a JavaDoc like documentation for >> the .NET client somewhere? >> >> Briefly probing the code I could not see much similarity with the W3C >> recommendation: >> http://www.w3.org/TR/DDR-Simple-API/#sec-interfaces >> >> Note, these abstract interfaces are more or less like UML, you don't have >> to generate code from it but the overall behavior should match. >> DeviceAtlas may have named a few classes differently, e.g. "Property" >> instead of "PropertyRef" (in the literral W3C Java interface spec) but the >> overall behavior is the same. >> >> So can you explain (both of you, also Reza) how we can claim the Java >> and .NET "new" API does that now? >> >> As mentioned, if we chose to significantly deviate from it like WURFL, >> that could be OK, but we must no longer claim to be "W3C compliant" then if >> the API completely violates the recommendation or differs so much it has >> nothing in common any more. >> >> Cheers, >> Werner >> >> >> >> On Wed, Jul 9, 2014 at 6:26 PM, Werner Keil <[email protected]> >> wrote: >> >>> Then what you work on is not a 1.0 release but a 1.x or 2.x release, see >>> the DeviceAtlas API (and even others like WURFL probably got a bit of a >>> synergy between different platforms) >>> >>> So instead of a rushed, incomplete "new" API the 1.0 release should be >>> what has been ready and mature for the past 2 years and AFAIK would feel >>> very familiar to Eberhard's .NET API. >>> >>> Whether or not the .NET API also gets a rewrite, we shall see, ideally >>> (just take DeviceAtlas and other commercial APIs, not to mimick them but >>> see their consistent behavior across platforms) it would, but for 1.0 the >>> goal would be to support Java and .NET in a consistent way. >>> >>> Werner >>> >>> On Wed, Jul 9, 2014 at 6:21 PM, Werner Keil <[email protected]> >>> wrote: >>> >>>> All, >>>> >>>> While the exact names of particular classes in the recent DeviceAtlas >>>> API differ a bit from W3C simple, you can see here, the behavior feels >>>> almost identical in Java and .NET (and all other languages we may support >>>> in the future instead of creating too many different libraries to do the >>>> same thing that already works[?]) >>>> >>>> https://deviceatlas.com/resources/enterprise-api-documentation >>>> >>>> The JavaDoc and .NET equivalent feel almost identical. >>>> >>>> OpenDDR not only by the OpenDDR team has been adopted on GitHub, see >>>> https://github.com/search?q=openddr&ref=cmdform >>>> Including a port to the Play! framework and several other languages. >>>> >>>> So the Java API should feel along the lines of Eberhard's .NET port to >>>> C#. If it matches or even directly uses W3C artifacts, so should the 1.0 >>>> release for Java. >>>> >>>> Werner >>>> >>>> >>>> On Wed, Jul 9, 2014 at 6:00 PM, Werner Keil <[email protected]> >>>> wrote: >>>> >>>>> See DeviceAtlas, I don't agree that dropping W3C compliance on the >>>>> Java side is correct. >>>>> >>>>> If you want to create a "new WURFL", sure, go ahead let's drop it, but >>>>> if we prefer to stay compatible with the de facto market leader here (plus >>>>> one or two alternatives, semi-open like MaDDR or completely open like >>>>> OpenDDR) we should maintain W3C support like they do. >>>>> >>>>> Werner >>>>> >>>>> On Wed, Jul 9, 2014 at 5:51 PM, eberhard speer jr. <[email protected]> >>>>> wrote: >>>>> >>>>>> -----BEGIN PGP SIGNED MESSAGE----- >>>>>> Hash: SHA1 >>>>>> >>>>>> Hi, >>>>>> >>>>>> the .Net W3C implementation follows the >>>>>> http://www.w3.org/TR/DDR-Simple-API/ specs, to the letter. >>>>>> >>>>>> Like Reza, I think it's nice to have but not vital to this release. >>>>>> And some kind of integration with DeviceMap still needs to happen. >>>>>> >>>>>> >>>>>> esjr >>>>>> -----BEGIN PGP SIGNATURE----- >>>>>> Version: GnuPG v1.4.8 (MingW32) >>>>>> Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ >>>>>> >>>>>> iQEcBAEBAgAGBQJTvWUPAAoJEOxywXcFLKYc5WQH+we0ZTLfsxRhWzqpNhHj+p+k >>>>>> anT2nfXK0iwz6va1VInHKEEuULMQD/FFPP33GrAj/dV4KHAKNB4w67G9TB0RcIhz >>>>>> 3Y2YPtg5eLSnOyY1O6+2ncCR/PwU7Sn78V5XKrWpaxWVcLmVSt1uCvOxUXG3KZhJ >>>>>> 8Jn9K2N5XiYfH+KI3JWWZhFvPu1eE6m5aS2fyEJiq3B0XfqmmiTYMa/iY/+B/Cha >>>>>> SdLBBVGpIWVt/RpLqdZou0XjXwvzb/b74SGr9l0fBQA2zjDB7D9PXRAz8PbV6h96 >>>>>> 8NRE2DWPL4aBxBKr1XHp1NKG8ccrbK/hf5oSnwygO2vWSyK+c4He58PCwnxe2Ec= >>>>>> =GSP4 >>>>>> -----END PGP SIGNATURE----- >>>>>> >>>>> >>>>> >>>> >>> >> >
