Loading ...
Sorry, an error occurred while loading the content.
 

Re: .NET API and "generic" device id?

Expand Messages
  • sam_dibart
    That makes sense, but I m still not sure why the Tera-Wurfl Explorer will give me generic web browser (and hence is mobile = false ) for the same input...
    Message 1 of 14 , Jul 5, 2011
      That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...

      Thanks,
      Sam

      --- In wmlprogramming@yahoogroups.com, Luca Passani <luca.passani@...> wrote:
      >
      >
      > You are basically asking WURFL: what mobile device has user agent 'foo'.
      > WURFL will say: I do not recognize this, so my answer is "the generic mobile device"
      >
      > You then ask "is the generic mobile device a mobile device". And WURFL says "yes"
      >
      > Luca
      >
      > On 7/5/2011 6:55 PM, sam_dibart wrote:
      > >
      > > Any suggestions on why I am seeing random text strings evaluate to "is
      > > wireless = true" (ie "foo")?
      > >
      > > Thanks,
      > > Sam
      > >
      > > --- In wmlprogramming@yahoogroups.com
      > > <mailto:wmlprogramming%40yahoogroups.com>, Luca Passani <luca.passani@> wrote:
      > > >
      > > > On 7/5/2011 3:48 PM, sam_dibart wrote:
      > > > > I've run it both ways (with/without the web browser patch file). While
      > > using the web browser patch file does reduce the number of failed tests, it
      > > doesn't eliminate them (ie, render the same results as the previous version).
      > > I'm most concerned about the fact that it appears any random string of letters
      > > (ie, "foo") come back with "is wireless" being true. This behavior is going to
      > > be hard to explain to my QA Analyst!
      > > >
      > > > Well, WURFL was born as a tool to detect mobile devices, with web browser
      > > > detected as an ad on.
      > > >
      > > > Having said this, the heuristics to detect web browsers are a lot improved and
      > > > the UAs which are failing is really old stuff. Going after that crap would mean
      > > > a slower and more complex API, for negligible gain in terms of detection
      > > accuracy
      > > >
      > > > Luca
      > > >
      > >
      > >
      >
    • Luca Passani
      ... this was a design decision made for Tera-WURFL. Anyway, you have a point that now that all the APIs are under the ScientiaMobile umbrella, those
      Message 2 of 14 , Jul 8, 2011
        On 7/5/2011 8:31 PM, sam_dibart wrote:
        > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...

        this was a design decision made for Tera-WURFL. Anyway, you have a point that
        now that all the APIs are under the ScientiaMobile umbrella, those differences
        should be reconciled.

        What does the community think?

        what should happen to unrecognized HTTP requests?
        should API err on the side of web browser or should the API err on the side of
        mobile device?

        (generc/generic_xhtml) vs generic_web_browser

        Comments?

        Luca
      • Luca Passani
        ... btw, WURFL APIs ship with the source-code. You still need to acquire a commercial license if AGPL is an issue, but nothing prevents you from changing the
        Message 3 of 14 , Jul 8, 2011
          On 7/5/2011 8:31 PM, sam_dibart wrote:
          > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...

          btw, WURFL APIs ship with the source-code. You still need to acquire a
          commercial license if AGPL is an issue, but nothing prevents you from changing
          the behavior for unrecognized devices to suite your needs

          Luca
        • sam_dibart
          I think we re going to end up checking the device s id when we get it back from the WURFL and overriding the wireless attribute if we see that it s the
          Message 4 of 14 , Jul 8, 2011
            I think we're going to end up checking the device's id when we get it back from the WURFL and overriding the "wireless" attribute if we see that it's the "generic" device.

            I'll be curious to see what others are thinking/doing about this.

            Thanks again for all your help.
            Sam

            --- In wmlprogramming@yahoogroups.com, Luca Passani <luca.passani@...> wrote:
            >
            > On 7/5/2011 8:31 PM, sam_dibart wrote:
            > > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...
            >
            > btw, WURFL APIs ship with the source-code. You still need to acquire a
            > commercial license if AGPL is an issue, but nothing prevents you from changing
            > the behavior for unrecognized devices to suite your needs
            >
            > Luca
            >
          • jwrosewell
            Sam, You could create a patch file that contains a new definition for the generic device, but with the is_wireless_device set to false. This would save you
            Message 5 of 14 , Jul 8, 2011
              Sam,

              You could create a patch file that contains a new definition for the "generic" device, but with the "is_wireless_device" set to false. This would save you having to write any code.

              Something like this should work as a patch file.

              <?xml version="1.0" encoding="UTF-8"?>
              <wurfl_patch>
              <devices>
              <device id="generic" user_agent="" fall_back="root">
              <group id="product_info">
              <capability name="is_wireless_device" value="false"/>
              </group>
              </device>
              </devices>
              </wurfl_patch>

              This is how we overcome the problem without needing seperate code.

              Regards,

              James Rosewell

              --- In wmlprogramming@yahoogroups.com, "sam_dibart" <sam_dibart@...> wrote:
              >
              > I think we're going to end up checking the device's id when we get it back from the WURFL and overriding the "wireless" attribute if we see that it's the "generic" device.
              >
              > I'll be curious to see what others are thinking/doing about this.
              >
              > Thanks again for all your help.
              > Sam
              >
              > --- In wmlprogramming@yahoogroups.com, Luca Passani <luca.passani@> wrote:
              > >
              > > On 7/5/2011 8:31 PM, sam_dibart wrote:
              > > > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...
              > >
              > > btw, WURFL APIs ship with the source-code. You still need to acquire a
              > > commercial license if AGPL is an issue, but nothing prevents you from changing
              > > the behavior for unrecognized devices to suite your needs
              > >
              > > Luca
              > >
              >
            • Luca Passani
              this way, all devices will inherit the false value and be marked as non-mobile. If one changes the value in generic, all the subfamilies should be fixed to
              Message 6 of 14 , Jul 8, 2011
                this way, all devices will inherit the false value and be marked as non-mobile.
                If one changes the value in generic, all the subfamilies should be fixed to
                revert to a sensible default.

                Luca

                On 7/8/2011 6:46 PM, jwrosewell wrote:
                > Sam,
                >
                > You could create a patch file that contains a new definition for the "generic" device, but with the "is_wireless_device" set to false. This would save you having to write any code.
                >
                > Something like this should work as a patch file.
                >
                > <?xml version="1.0" encoding="UTF-8"?>
                > <wurfl_patch>
                > <devices>
                > <device id="generic" user_agent="" fall_back="root">
                > <group id="product_info">
                > <capability name="is_wireless_device" value="false"/>
                > </group>
                > </device>
                > </devices>
                > </wurfl_patch>
                >
                > This is how we overcome the problem without needing seperate code.
              • Steve Kamerman
                To expand, Tera-WURFL works like this: If a device is not matched to any other known device it is checked for occurrences of keywords and HTTP headers that
                Message 7 of 14 , Jul 11, 2011
                  To expand, Tera-WURFL works like this:  If a device is not matched to any other known device it is checked for occurrences of keywords and HTTP headers that would indicate that it is a mobile or desktop device (for example, if you lookup "armv" the result is "generic_xhtml").  If no keywords are found to push the result in either direction, the device is assumed to be "generic_web_browser".  In practice, I've found this to be very effective, but like Luca said, we will be reconciling the differences between the APIs and this behavior is likely to change.

                  Thanks,

                  Steve Kamerman

                  On Fri, Jul 8, 2011 at 3:07 AM, Luca Passani <luca.passani@...> wrote:
                   

                  On 7/5/2011 8:31 PM, sam_dibart wrote:
                  > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer will give me "generic web browser" (and hence "is mobile = false") for the same input...

                  this was a design decision made for Tera-WURFL. Anyway, you have a point that
                  now that all the APIs are under the ScientiaMobile umbrella, those differences
                  should be reconciled.

                  What does the community think?

                  what should happen to unrecognized HTTP requests?
                  should API err on the side of web browser or should the API err on the side of
                  mobile device?

                  (generc/generic_xhtml) vs generic_web_browser

                  Comments?

                  Luca


                • Luca Passani
                  people, I am not opposed to the idea of making generic default to web browser, but the following observations apply: - Are you aware that this implies a rather
                  Message 8 of 14 , Jul 12, 2011
                    people,

                    I am not opposed to the idea of making generic default to web browser, but the
                    following observations apply:

                    - Are you aware that this implies a rather deep philosophical change in WURFL?
                    from mobile device recognition database to all UAs recognition database

                    - We will need to shuffle around quite a few things to avoid that changes in
                    generic do not impact existing data.

                    Are you all ok with this?

                    Luca

                    On 7/11/2011 5:24 PM, Steve Kamerman wrote:
                    >
                    >
                    > To expand, Tera-WURFL works like this: If a device is not matched to any
                    > other known device it is checked for occurrences of keywords and HTTP headers
                    > that would indicate that it is a mobile or desktop device (for example, if you
                    > lookup "armv" the result is "generic_xhtml"). If no keywords are found to
                    > push the result in either direction, the device is assumed to be
                    > "generic_web_browser". In practice, I've found this to be very effective, but
                    > like Luca said, we will be reconciling the differences between the APIs and
                    > this behavior is likely to change.
                    >
                    > Thanks,
                    >
                    > Steve Kamerman
                    >
                    > On Fri, Jul 8, 2011 at 3:07 AM, Luca Passani <luca.passani@...
                    > <mailto:luca.passani@...>> wrote:
                    >
                    >
                    > On 7/5/2011 8:31 PM, sam_dibart wrote:
                    > > That makes sense, but I'm still not sure why the Tera-Wurfl Explorer
                    > will give me "generic web browser" (and hence "is mobile = false") for the
                    > same input...
                    >
                    > this was a design decision made for Tera-WURFL. Anyway, you have a point that
                    > now that all the APIs are under the ScientiaMobile umbrella, those differences
                    > should be reconciled.
                    >
                    > What does the community think?
                    >
                    > what should happen to unrecognized HTTP requests?
                    > should API err on the side of web browser or should the API err on the side of
                    > mobile device?
                    >
                    > (generc/generic_xhtml) vs generic_web_browser
                    >
                    > Comments?
                    >
                    > Luca
                    >
                    >
                    >
                    >
                    >
                  Your message has been successfully submitted and would be delivered to recipients shortly.