Actually, BrowserMap looks like it's doing mostly that: http://devicemap-vm.apache.org/browsermap/
Please talk to Radu, the source is also in SVN (and that's the only one with a Git mirror so far) https://svn.apache.org/repos/asf/devicemap/trunk/browsermap Werner On Fri, Jan 2, 2015 at 2:39 PM, Werner Keil <[email protected]> wrote: > This is an Open Source JavaScript solution on GitHub: > https://github.com/faisalman/ua-parser-js > > It mainly catches the "obvious" stuff, but adding more where necessary > sounds doable. And looking at the StackOverflow tips again, > "I am currently using JavaScript for mobile device detection on my > website..." > sounds like they run JavaScript in a web app, not PhoneGap style on the > actual device, so the formulas mentioned there or something better would be > the way to get e.g. the pixel density, etc. > > HTH, > Werner > > On Fri, Jan 2, 2015 at 2:33 PM, Werner Keil <[email protected]> wrote: > >> Hello Volkan, >> >> Thanks for the update. >> There are some UA sites that require login. 51DegreesMobi doesn't, but I >> am sceptical the result I just ran from my own Windows box with properties >> like >> >> - SuggestedImageButtonHeightMms >> - 3.5 >> - SuggestedImageButtonHeightPixels >> - 11.8 >> - SuggestedLinkSizePixels >> - 11.8 >> - SuggestedLinkSizePoints >> - 10 >> >> (those with slightly different names many still defined by OMA & co. are >> what we have "pixel_density", etc.) are acceptable for reuse. >> >> The footer claims: >> >> Format: *Enterprise* >> >> And that AFAIR is not under a Free or Open Source license. 51Degrees has >> some sort of dual-licensind, with parts of the material that can be used >> (as long as ASF and its legal team doesn't see a conflict, just look at >> recent disputes between GPL and Apache License;-) but those attributes >> considered "Enterprise" are likely proprietary just like the software they >> provide, so you cannot use the information (except the bare UA string which >> is none of their own business) >> >> I am not sure, if the StackOverflow link (which was the first hit on the >> subject;-) is missing something or may refer to JavaScript running on the >> actual device (Cordova) rather than in your browser, but without license >> breach you cannot simply copy and paste it from a majority of vendor sites, >> you must refer to sources that allow you to use the information, that's >> where all the DDR data contributed by the likes of OMA and others came from >> in the past. Most of what OMA did in this area is fairly old, see >> http://technical.openmobilealliance.org/Technical/technical-information/release-program/current-releases/user-agent-profile-v2-0 >> >> I don't think they provide working solutions like a UA-detector now >> either. >> >> Werner >> >> On Fri, Jan 2, 2015 at 2:17 PM, Volkan Yazıcı <[email protected]> >> wrote: >> >>> Hello Werner, >>> >>> One commercial product in my mind was 51Degree User-Agent Tester >>> <https://51degrees.com/Resources/User-Agent-Tester>. I can test the UAs >>> against that web page and collect certain specs asked by Reza. >>> >>> I also get in touch with Wikimedia Traffic Analysis Report - Browsers >>> e.a. >>> < >>> https://stats.wikimedia.org/wikimedia/squids/SquidReportClients.htm#errata >>> > >>> page >>> maintainer Erik Zachte and asked him whether he can provide the UA >>> collection to us or not. >>> >>> If I am not mistaken, the PPI calculation via JS method you have >>> suggested >>> is not usable given just you have the UA string of the device. I just >>> have >>> the UA string collection and I need to come up with certain device specs. >>> >>> For the upcoming DWX conference on June, I can be there for Java+JS >>> presentations. >>> >>> Best. >>> >>> On Fri, Jan 2, 2015 at 1:47 PM, Werner Keil <[email protected]> >>> wrote: >>> >>> > Volkan, >>> > >>> > Glad, having at least one JIRA ticket you filed snubbed and closed >>> instead >>> > of (a very easy and working;-) applying the patch available did not >>> > discourage you from trying to help;-) That sounds like a good idea, >>> just >>> > what do you mean with "proprietary UA resolver"? >>> > If it's a commercial product, do you know its license or terms of use? >>> That >>> > depends on whether you may use the information at least without >>> troubles >>> > from that vendor (there's been some thoughts about UA data by e.g. the >>> > WURFL guys but they later had to admit, that this and other COMMON >>> > knowledge like the UA are now owned by anybody, at most the maker of >>> each >>> > device could claim some "ownership" since they also make the device >>> itself >>> > and software running on it;-) >>> > >>> > It's been a long time, but I recall Bertrand and others worked on some >>> > JavaScript solution, @Bertrand/Radu is there something you could help >>> > Volkan with? >>> > Third party contributors like Wikimedia also gathered such >>> information, but >>> > I am not sure if they ever did share the information. >>> > Pixel density and most other stuff seems best handled with JavaScript, >>> see >>> > >>> > >>> http://stackoverflow.com/questions/16541676/what-are-best-practices-for-detecting-pixel-ratio-density >>> > >>> > Again, maybe we have some building blocks, but as these were never >>> released >>> > (aside from Browsermap) I can't say for sure what is there and what >>> may be >>> > missing. >>> > If we can avoid tie-in of commercial products, it would be best, that >>> way >>> > (as soon as you could be voted on as PMC member, assuming paperwork is >>> > already processed by ASF?) you could also contribute to such scripted >>> > solutions. Either improving what's there or writing something new. >>> > >>> > If you're in Europe close enough to Germany and have no constraints to >>> > travel (in June) unlike those Eberhard mentioned, I'm happy to file >>> the DWX >>> > proposal for a DeviceMap topic in the next few days. They said the >>> "more >>> > the merrier", thus a 2 or 3 person talk is fine. And showing not just >>> the >>> > Java pieces of software but e.g. something with JavaScript would be a >>> > perfect match to the conference scope (it's more than just .NET or >>> Java;-) >>> > >>> > Thanks and Regards, >>> > Werner >>> > >>> > On Fri, Jan 2, 2015 at 10:23 AM, Volkan Yazıcı < >>> [email protected]> >>> > wrote: >>> > >>> > > Hi all, >>> > > >>> > > The entire UA collection tickets (DMAP-94, 95, 96, 97, 98, 100, 102) >>> I >>> > have >>> > > submitted to JIRA are being marked as NeedsSpec, which is fine. In >>> the >>> > > explanation, I am being told that I need to collect the following >>> > > properties for each UA: >>> > > >>> > > - id >>> > > - vendor >>> > > - model >>> > > - marketing name >>> > > - resolution-x >>> > > - resolution-y >>> > > - pixel-density-ppi >>> > > - release-year >>> > > - default-os >>> > > - hardware >>> > > >>> > > I have some questions regarding these properties: >>> > > >>> > > - How other people collect this sort of information? Is there a >>> > certain >>> > > set of steps that I can follow? I was considering writing a >>> crawler on >>> > > top >>> > > of a web-based proprietary UA resolver, would that be ok >>> considering >>> > the >>> > > licensing issues? >>> > > - Is there scheme am I supposed to follow for the *id* attribute? >>> Or >>> > > something descriptive would be just fine? >>> > > - *resolution-x/y* means the width and height of the screen in >>> pixels, >>> > > right? >>> > > - How do we calculate *pixel-density-ppi*? >>> > > >>> > > Best. >>> > > >>> > >>> >> >> >
