On Wed, Jul 30, 2014 at 4:29 AM, Reza <[email protected]> wrote:
> I totally agree. So I want to summarize some of the points here because I > think we hit on some good ones: > > -To Eberhard's point, now is the time to discuss the direction of features > and this project as a whole. I think moving fast is good, but only if we > make the right design decisions. So im in agreement regarding discussing > and scrutinizing more. > I asked Bertrand if we might use the planning features by new JIRA versions, especially some "Agile Board" (currently there are two, without a strict notion of "sprints" the Kanban version seems more flexible and useful) but he feels more comfortable with the mailing list here to delegate or pick things. On a high level, the Board Wiki summary/call is a bit like "user stories" but they do not need to be in greater detail, at most a "component" like Java, Data, .NET, etc. not so much down to parts like "Loader", etc. which was also handled by JIRA before. I try to select tickets and also put them "in Progress", at least those assigned to my JIRA account(s). Anybody who is "committer" in JIRA should be able to assign a task to either themselves or another person, and the person it's assigned to can "start progress". Even without a real Kanban board, JIRA tells you who is working on what and ideally we should have the discipline not to just rush and do something, either close the ticket or even leave it open forever. > > -We can and should detect desktops to greater accuracy than "generic > desktop". We can easily do apple, windows, and linux flavors. Basically we > can combine device and OS detection for desktops. > > -We should remove the resolution attribute from desktops since we dont > know it and it can lead to confusion. > > +1 It is a "Core" feature, so you can't remove the attribute but the desktop does not have to provide a value. IMHO that is less confusing. P.s.: If you visit ScientiaMobile see above, I just tried it on a Windows 7 machine and guess what, the "simple" (not paying 50k € per month in a large company or popular website) result is no better or worse than the original OpenDDR device data. While OpenDDR has always been W3C DDR compatible, hence nothing was ever taken "directly from WURFL" the DDR standardization process at W3C was also observed from the sidelines by the WURFL guys, and they used the same sources as OpenDDR, MaDDR or DeviceAtlas, DetectRight, etc. did, hence you get Capabilities: - is_tablet: false - pointing_method: mouse - resolution_width: 800 - resolution_height: 600 Thus the quality of an average desktop is exactly the same, and you may even have to pay ScientiaMobile 200€ or more per month if you are a startup or make money with your blog[?] > -We can and should add browser detection. > > -If there is ever a question regarding what is possible for this project, > just look at our friends over at wurfl ( http://www.scientiamobile.com/ > ). They list out some great use cases and they have a decent client list. > dClass also has a decent user base. The main use of dClass is to basically > redirect users between mobile and desktop sites, so thats a pretty big use > case and a big consumer of DeviceMap data. > > Compared to the old established vendors like DetectRight or DeviceAtlas, ScientiaMobile seems like a mere imposter of what they did. Neither Google nor Facebook are likely to really use a subscription or license, the "clients" banner seems blown-up. I know especially the Open Source Advocate at Facebook, James Pearce, who is among the first OpenDDR followers on Twitter, we talked at JavaOne and he said he's been involved in W3C DDR, also WURFL when it was still Open Source, but the likes of Facebook or Google don't pay SM 1 Mio.$ or more per year, they use either OpenDDR or their own home-grown detection, in Facebook's case eihter PHP or C/C++, so you may see something not so different from what you did at WheatherChannel [?] Same for Google, as soon as WURFL changed its license you can be sure some of the big companies that may have used the Open Source equivalent or (given the size of these companies) probably still do maintain it internally, nobody can keep them from that or force them to pay unless they consume new SM services or content. Some of the banks or FraPort, well, that's more likely, I came by a German start-up once where an entire FraPort project was "made-up" by a secretary (at least they said so) and I just mention "BER", the Berlin Airport, so for some of these huge logistic projects nobody even notices 50k or 500k paid to ScientiaMobile for the mobile version of an airport site. > Any other takeaways I missed??? > > > Not really. We don't have to do anything even commercial vendors can't do better or don't want to unless you pay them hefty. A simple recognition tool visible from the main DeviceMap site is probably the first step someone might want to work on (there is a simple task of "Improve Website" in JIRA, like "Examples" an epic sounds good so different smaller tasks could gather there over time) Werner > ________________________________ > From: Werner Keil <[email protected]> > To: "[email protected]" < > [email protected]> > Sent: Tuesday, July 29, 2014 10:18 PM > Subject: Re: OS Detection > > > > P.s.: in Nürnberg I believe somebody asked, if "ARM" vs. "Intel" or "AMD´" > can be recognized. > I told them, technically in theory one could also make that distinction, > but see Apple or others even an OS is not always tied to just one processor > architecture, so it may not provide real value to an application or its > user. The DDR Spec would allow it, but not every property seems equally > important. > > > Werner > On Wed, Jul 30, 2014 at 4:14 AM, Werner Keil <[email protected]> > wrote: > > That's why something like "max_image_size" (which not just WURFL has as > comparisons show) could not be bad. > >ODDR vocabulary added new properties like "is_tablet", which work on > well-known tablets, I tried that myself on Galaxy or Nexus, but e.g. a > brand new Surface was not found (I don't think there's even a fallback for > Surface, AFAIK some Lumia is recognized as "Windows Phone", but the > fallback in that case is "genericNokia" bringing the device back to Symbian > if it's new or unknown > > > > > > > >On Wed, Jul 30, 2014 at 3:42 AM, eberhard speer jr. <[email protected]> > wrote: > > > >-----BEGIN PGP SIGNED MESSAGE----- > >>Hash: SHA1 > >> > >>Actually, I have more than one monitor and all set to a different > >>resolution -- there's a reason for that -- I'd expect any serious site > >>to 'know' how to deal with me moving a browser to another screen. > >> > >> > >>On 30/07/2014 04:40, eberhard speer jr. wrote: > >>>> Simply assuming every desktopDevice must now have 1600x900 > >>>> pixel... > >>> > >>> I do not think anyone suggested such a thing. There are many ways > >>> to either get the correct resolution or, preferable : maximize the > >>> use of desktop-real-estate using all the nice stuff HTML5 and CSS > >>> has to offer. > >>> > >>> esjr > >>> > >> > >>-----BEGIN PGP SIGNATURE----- > >>Version: GnuPG v2.0.22 (MingW32) > >>Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ > >> > >>iQEcBAEBAgAGBQJT2E2VAAoJEOxywXcFLKYcMQsIAJzDT8CGR2nrzE5IGcf/4Ci+ > >>xEfjzIGjFoJNLBuUxkbyXMhLC3zFmi46GehUFs1pGfTl4aZeDB+C+/tj7MLR5qHN > >>YhuxkuOg8P8SteN4QUumGpJPAFIoAXeez3Zl6todKQTDovs+wkWNqo2yNiIrcAH4 > >>OPmFbHdQRD9yW9sgwTNSdDruQz/ltpymN/BdTMC90jp+m/o7ci0Gukt6jERgUIme > >>9LG66CFO/RGFv6xVOKNNh0bh9B72Xll+qKPvYn1zujLAW96+w7XDbKN6IszryQI1 > >>2xbquocLcgGhRnYIYXQPK8yYsGmWvmCRxLNrhoEk358vLbyEOVMesUpiDuFYfS4= > >>=D/4i > >>-----END PGP SIGNATURE----- > >> > >
