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

Re: [wmlprogramming] I think we need to update GAP (accesskeys are now broken)

Expand Messages
  • Miha Valencic
    For what it s worth, my old N95 has those keybindings from the beginning, so it s not something new. SonyEricsson K800i has an option to switch between using
    Message 1 of 21 , Apr 1, 2009
    • 0 Attachment
      For what it's worth, my old N95 has those keybindings from the beginning, so
      it's not something new. SonyEricsson K800i has an option to switch between
      using keys for navigation (similar to nokia, also includes pgup, pgdwn (3
      and 7), and what was even more interesting for me, a refresh page (5)) or
      using the keypad for access keys. It defaults to access keys though, if I
      recall correctly.

      Having said that, I think that access keys don't matter much on modern
      terminals, but are handy on the old ones. We've never used them to a large
      extent though.

      Miha

      On Wed, Apr 1, 2009 at 12:06 AM, Luca Passani <passani@...> wrote:

      >
      > People, I think I need to go back and change/remove one of the GAP
      > practices:
      >
      > http://www.passani.it/gap/#ACCESS_KEYS
      >
      >


      [Non-text portions of this message have been removed]
    • Luca Passani
      ... Yes, but GAP is also about 1) old devices 2) best possible one-size-fits all. So, these changes have negative impact. I think some companies rely on access
      Message 2 of 21 , Apr 1, 2009
      • 0 Attachment
        Miha Valencic wrote:
        > Having said that, I think that access keys don't matter much on modern
        > terminals, but are handy on the old ones. We've never used them to a large
        > extent though.
        >

        Yes, but GAP is also about 1) old devices 2) best possible one-size-fits
        all. So, these changes have negative impact.
        I think some companies rely on access keys for WAP games.

        As always, I don't like the approach where everyone in the industry
        feels entitled to break standards and adopted conventions unilaterally
        from one day to the next. This should be true of everyone. Not only
        Novarra and Vodafone.

        Anyway, any comments about 1) modifying GAP and 2) adding a specific
        capability?

        Luca
      • Jose Manrique Lopez de la Fuente
        I filled a bug in Forum Nokia about it sometime ago: http://wiki.forum.nokia.com/index.php/KIC001108_-_Accesskey_attribute_not_working_in_Web_Browser_for_S60
        Message 3 of 21 , Apr 1, 2009
        • 0 Attachment
          I filled a bug in Forum Nokia about it sometime ago:
          http://wiki.forum.nokia.com/index.php/KIC001108_-_Accesskey_attribute_not_working_in_Web_Browser_for_S60

          Neither Nokia nor Opera (Mini, Mobile) support accesskey attribute!
          Very sad since I think it is one of the key features for navigation
          help in a mobile phone.

          2009/4/1 Miha Valencic <miha.valencic@...>:
          > For what it's worth, my old N95 has those keybindings from the beginning, so
          > it's not something new. SonyEricsson K800i has an option to switch between
          > using keys for navigation (similar to nokia, also includes pgup, pgdwn (3
          > and 7), and what was even more interesting for me, a refresh page (5)) or
          > using the keypad for access keys. It defaults to access keys though, if I
          > recall correctly.
          >
          > Having said that, I think that access keys don't matter much on modern
          > terminals, but are handy on the old ones. We've never used them to a large
          > extent though.
          >
          > Miha
          >
          > On Wed, Apr 1, 2009 at 12:06 AM, Luca Passani <passani@...> wrote:
          >
          >>
          >> People, I think I need to go back and change/remove one of the GAP
          >> practices:
          >>
          >> http://www.passani.it/gap/#ACCESS_KEYS
          >>
          >>
          >
          > [Non-text portions of this message have been removed]
          >
          >



          --
          J. Manrique López de la Fuente
          http://www.jsmanrique.es
        • Jose Manrique Lopez de la Fuente
          ... I think that [ACCESS_KEYS] GAP rule shouldn t be changed, specially if there is a wurfl capability that let you know if the device supports accesskey. ...
          Message 4 of 21 , Apr 1, 2009
          • 0 Attachment
            2009/4/1 Luca Passani <passani@...>:
            > Having said this, what is to be done? my suggestion is that we change
            > the [ACCESS_KEYS] GAP rule into [NO_ACCESS_KEYS] and explain why access
            > keys are a bad idea for a whole bunch of mobile users.

            I think that [ACCESS_KEYS] GAP rule shouldn't be changed, specially if
            there is a wurfl capability that let you know if the device supports
            accesskey.

            > BTW, should we have a capability: xhtml_support_accesskey to mark
            > whether it's OK to use accesskeys for a given device?

            I think this is a good idea.

            Best regards,

            --
            J. Manrique López de la Fuente
            http://www.jsmanrique.es
          • Miha Valencic
            Well, we could move *access_key_support * from wml_ui to ____ (markup/display/???) group (since it also applies to XHTML devices, right?). Apart from that, a
            Message 5 of 21 , Apr 1, 2009
            • 0 Attachment
              Well, we could move "*access_key_support"* from wml_ui to ____
              (markup/display/???) group (since it also applies to XHTML devices, right?).
              Apart from that, a short mention of the issue at hand in the GAP seems
              perfectly sufficient for me.

              Then, there are devices that support access keys, but only if so configured
              (like SE K800). I have no idea what number of devices do that, so I would
              just ignore this for now.

              On Wed, Apr 1, 2009 at 9:37 AM, Luca Passani <passani@...> wrote:

              >
              > Anyway, any comments about 1) modifying GAP and 2) adding a specific
              > capability?
              >
              >


              [Non-text portions of this message have been removed]
            • Luca Passani
              ... that WURFL capability referred to WML devices. The problem was real in WML, because devices which did not support accesskey, would throw an error in the
              Message 6 of 21 , Apr 1, 2009
              • 0 Attachment
                Jose Manrique Lopez de la Fuente wrote:
                > 2009/4/1 Luca Passani <passani@...>:
                >
                >> Having said this, what is to be done? my suggestion is that we change
                >> the [ACCESS_KEYS] GAP rule into [NO_ACCESS_KEYS] and explain why access
                >> keys are a bad idea for a whole bunch of mobile users.
                >>
                >
                > I think that [ACCESS_KEYS] GAP rule shouldn't be changed, specially if
                > there is a wurfl capability that let you know if the device supports
                > accesskey.
                >

                that WURFL capability referred to WML devices. The problem was real in
                WML, because devices which did not support accesskey, would throw an
                error in the presence of an accesskey attribute!
                In XHTML, until the today, the worst that could happen is that accesskey
                was ignored.

                Luca
              • Jose Manrique Lopez de la Fuente
                ... Yes, I know, and I am agree with Miha about populating the capability to XHTML devices... -- J. Manrique López de la Fuente http://www.jsmanrique.es
                Message 7 of 21 , Apr 1, 2009
                • 0 Attachment
                  2009/4/1 Luca Passani <passani@...>:
                  > Jose Manrique Lopez de la Fuente wrote:
                  >> 2009/4/1 Luca Passani <passani@...>:
                  > that WURFL capability referred to WML devices. The problem was real in
                  > WML, because devices which did not support accesskey, would throw an
                  > error in the presence of an accesskey attribute!
                  > In XHTML, until the today, the worst that could happen is that accesskey
                  > was ignored.

                  Yes, I know, and I am agree with Miha about populating the capability
                  to XHTML devices...

                  --
                  J. Manrique López de la Fuente
                  http://www.jsmanrique.es
                • Luca Passani
                  ... So, which XHTML browsers do not support accesskey: Safari (all manufacturers? including android?). OperaMini. Anyone else? Luca
                  Message 8 of 21 , Apr 1, 2009
                  • 0 Attachment
                    Jose Manrique Lopez de la Fuente wrote:
                    >
                    > Yes, I know, and I am agree with Miha about populating the capability
                    > to XHTML devices...
                    >

                    So, which XHTML browsers do not support accesskey: Safari (all
                    manufacturers? including android?). OperaMini.

                    Anyone else?

                    Luca
                  • Boris Folgmann
                    Hi, ... Was there any progress made on defining a new capability for XHTML-Output? Relying on access_key_support I m still producing those nice but not
                    Message 9 of 21 , May 13, 2009
                    • 0 Attachment
                      Hi,

                      Luca Passani schrieb/wrote:
                      > So, which XHTML browsers do not support accesskey: Safari (all
                      > manufacturers? including android?). OperaMini.

                      Was there any progress made on defining a new capability for XHTML-Output?
                      Relying on access_key_support I'm still producing those nice but not
                      functional accesskey numbers before every link on my Nokia E90.

                      cu,
                      boris
                    • Yes WAP
                      Other mobile browsers that don t support access keys include, the BlackBerry browser, Bolt, UCWEB, Ozone and Skyfire. That said, I think access keys are still
                      Message 10 of 21 , May 13, 2009
                      • 0 Attachment
                        Other mobile browsers that don't support access keys include, the
                        BlackBerry browser, Bolt, UCWEB, Ozone and Skyfire.

                        That said, I think access keys are still important. All the embedded
                        browsers on feature phones do support them. Those are the devices
                        where access keys are most useful and needed because of clumsy and
                        inefficient scrolling and navigation.

                        Dennis Bournique
                        http://wapreview.com

                        On Wed, Apr 1, 2009 at 2:36 AM, Luca Passani <passani@...> wrote:
                        > Jose Manrique Lopez de la Fuente wrote:
                        >>
                        >> Yes, I know, and I am agree with Miha about populating the capability
                        >> to XHTML devices...
                        >>
                        >
                        > So, which XHTML browsers do not support accesskey: Safari (all
                        > manufacturers? including android?). OperaMini.
                        >
                        > Anyone else?
                        >
                        > Luca
                        >
                        >
                      • Luca Passani
                        ... Boris, can you tell me a bit more about the problems you are seeing on Nokia E90? are you recommending people to use accesskey in the page? or will your
                        Message 11 of 21 , May 13, 2009
                        • 0 Attachment
                          Boris Folgmann wrote:
                          > Hi,
                          >
                          > Luca Passani schrieb/wrote:
                          >
                          >> So, which XHTML browsers do not support accesskey: Safari (all
                          >> manufacturers? including android?). OperaMini.
                          >>
                          >
                          > Was there any progress made on defining a new capability for XHTML-Output?
                          > Relying on access_key_support I'm still producing those nice but not
                          > functional accesskey numbers before every link on my Nokia E90.
                          >

                          Boris, can you tell me a bit more about the problems you are seeing on
                          Nokia E90?

                          are you recommending people to use accesskey in the page? or will your
                          users discover the problem by themselves?

                          also, what should the capability be called? xhtml_disable_accesskey ?

                          Luca
                        • Luca Passani
                          ... Dennis, what do you recommend? a new capability called xhtml_disable_accesskey ? Luca
                          Message 12 of 21 , May 13, 2009
                          • 0 Attachment
                            Yes WAP wrote:
                            > Other mobile browsers that don't support access keys include, the
                            > BlackBerry browser, Bolt, UCWEB, Ozone and Skyfire.
                            >
                            > That said, I think access keys are still important. All the embedded
                            > browsers on feature phones do support them. Those are the devices
                            > where access keys are most useful and needed because of clumsy and
                            > inefficient scrolling and navigation.
                            >

                            Dennis, what do you recommend? a new capability called
                            xhtml_disable_accesskey ?

                            Luca
                          • Yes WAP
                            That seems like a perfect name for the capability.
                            Message 13 of 21 , May 13, 2009
                            • 0 Attachment
                              That seems like a perfect name for the capability.

                              On Wed, May 13, 2009 at 11:01 AM, Luca Passani <passani@...> wrote:
                              >
                              >
                              > Yes WAP wrote:
                              >> Other mobile browsers that don't support access keys include, the
                              >> BlackBerry browser, Bolt, UCWEB, Ozone and Skyfire.
                              >>
                              >> That said, I think access keys are still important. All the embedded
                              >> browsers on feature phones do support them. Those are the devices
                              >> where access keys are most useful and needed because of clumsy and
                              >> inefficient scrolling and navigation.
                              >>
                              >
                              > Dennis, what do you recommend? a new capability called
                              > xhtml_disable_accesskey ?
                              >
                              > Luca
                              >
                              >
                            • Boris Folgmann
                              Hi Luca, ... We ve ported an application from our own i-mode database to WURFL to be able to support all phones that understand some HTML dialect. No WML
                              Message 14 of 21 , May 14, 2009
                              • 0 Attachment
                                Hi Luca,

                                Luca Passani schrieb:
                                > Boris, can you tell me a bit more about the problems you are seeing on

                                We've ported an application from our own i-mode database to WURFL to be
                                able to support all phones that understand some HTML dialect. No WML support.

                                Our output complies to XHTML but for now we decided to omit the DOCTYPE at
                                the top, so we can additionally use <marquee> and <blink> in quirks mode.
                                So far this seems to work well.

                                In the past we always had accesskeys since they usally worked on i-mode
                                handsets. Now we're checking access_key_support.

                                In fact I mixed it up. For the E90 access_key_support is false, so no
                                access keys are used. Sorry! An example is the Siemens M55.
                                access_key_support is true, but in fact the numbers call special bookmarks
                                of the browser. This means that you will end up on a totally wrong page
                                when you try to select a link using an access key.

                                > are you recommending people to use accesskey in the page? or will your
                                > users discover the problem by themselves?

                                We have hints like [1] or an emoji (if supported) automatically rendered in
                                front of the link, input field or button. This encourages the user to try
                                the access key. If they don't work or even call a totally different special
                                function of the browser, this is of course confusing to the user.

                                > also, what should the capability be called? xhtml_disable_accesskey ?

                                The question is: are there devices that support access keys for WML pages
                                but not for XHTML pages or the reverse? Or is the value of
                                access_key_support simply wrong for the M55 and other devices?


                                cu,
                                boris
                              • Luca Passani
                                Boris, the access_key_support you see in WURFL dates back to 2002 and models the behavior of WML browsers. I wouldn t count on it. Have a look at the
                                Message 15 of 21 , May 14, 2009
                                • 0 Attachment
                                  Boris, the access_key_support you see in WURFL dates back to 2002 and
                                  models the behavior of WML browsers. I wouldn't count on it.

                                  Have a look at the capability xhtml_avoid_accesskeys I just created. I
                                  can do an extraction of the wurfldb for you if you wish. This way,
                                  you'll be able to try the new capability and see if it works for you

                                  Luca

                                  Boris Folgmann wrote:
                                  > Hi Luca,
                                  >
                                  > Luca Passani schrieb:
                                  >
                                  >> Boris, can you tell me a bit more about the problems you are seeing on
                                  >>
                                  >
                                  > We've ported an application from our own i-mode database to WURFL to be
                                  > able to support all phones that understand some HTML dialect. No WML support.
                                  >
                                  > Our output complies to XHTML but for now we decided to omit the DOCTYPE at
                                  > the top, so we can additionally use <marquee> and <blink> in quirks mode.
                                  > So far this seems to work well.
                                  >
                                  > In the past we always had accesskeys since they usally worked on i-mode
                                  > handsets. Now we're checking access_key_support.
                                  >
                                  > In fact I mixed it up. For the E90 access_key_support is false, so no
                                  > access keys are used. Sorry! An example is the Siemens M55.
                                  > access_key_support is true, but in fact the numbers call special bookmarks
                                  > of the browser. This means that you will end up on a totally wrong page
                                  > when you try to select a link using an access key.
                                  >
                                  >
                                  >> are you recommending people to use accesskey in the page? or will your
                                  >> users discover the problem by themselves?
                                  >>
                                  >
                                  > We have hints like [1] or an emoji (if supported) automatically rendered in
                                  > front of the link, input field or button. This encourages the user to try
                                  > the access key. If they don't work or even call a totally different special
                                  > function of the browser, this is of course confusing to the user.
                                  >
                                  >
                                  >> also, what should the capability be called? xhtml_disable_accesskey ?
                                  >>
                                  >
                                  > The question is: are there devices that support access keys for WML pages
                                  > but not for XHTML pages or the reverse? Or is the value of
                                  > access_key_support simply wrong for the M55 and other devices?
                                  >
                                  >
                                • Boris Folgmann
                                  ... That would be fine. Do you think it s ok to use the same capability for cHTML devices? cu, boris
                                  Message 16 of 21 , May 14, 2009
                                  • 0 Attachment
                                    Luca Passani schrieb/wrote:
                                    > Have a look at the capability xhtml_avoid_accesskeys I just created. I
                                    > can do an extraction of the wurfldb for you if you wish. This way,
                                    > you'll be able to try the new capability and see if it works for you

                                    That would be fine. Do you think it's ok to use the same capability for
                                    cHTML devices?

                                    cu,
                                    boris
                                  • Luca Passani
                                    ... Are you serving CHTML to XHTML devices or are those devices happy with XHTML? assuming it s the former, adding a chtml_avoid_accesskeys capability would be
                                    Message 17 of 21 , May 14, 2009
                                    • 0 Attachment
                                      Boris Folgmann wrote:
                                      > Luca Passani schrieb/wrote:
                                      >
                                      >> Have a look at the capability xhtml_avoid_accesskeys I just created. I
                                      >> can do an extraction of the wurfldb for you if you wish. This way,
                                      >> you'll be able to try the new capability and see if it works for you
                                      >>
                                      >
                                      > That would be fine. Do you think it's ok to use the same capability for
                                      > cHTML devices?
                                      >
                                      Are you serving CHTML to XHTML devices or are those devices happy with
                                      XHTML?

                                      assuming it's the former, adding a chtml_avoid_accesskeys capability
                                      would be consistent with what we have done in the past.
                                      Question for you: do you have a list of CHTML devices for which this
                                      capability should be set to false?

                                      BTW: I have just emailed you a WURFL snapshot with the new capability.
                                      Please let me know how it goes.

                                      Luca
                                    • Boris Folgmann
                                      Hi Luca, ... The output complies to DTD XHTML Mobile 1.0, but we are omitting the at the top, so cHTML or iHTML devices can read it, too. On XHTML
                                      Message 18 of 21 , May 26, 2009
                                      • 0 Attachment
                                        Hi Luca,

                                        Luca Passani schrieb/wrote:
                                        >>> Have a look at the capability xhtml_avoid_accesskeys I just created. I
                                        >> That would be fine. Do you think it's ok to use the same capability for
                                        >> cHTML devices?
                                        >>
                                        > Are you serving CHTML to XHTML devices or are those devices happy with
                                        > XHTML?

                                        The output complies to DTD XHTML Mobile 1.0, but we are omitting the
                                        <!DOCTYPE> at the top, so cHTML or iHTML devices can read it, too.
                                        On XHTML devices this enables quirks mode which allows us to use e.g.
                                        <marquee> or <blink>. These tags are not in the DTD and are thefore not
                                        accepted by a XHTML browser if the <!DOCTYPE> is specified.
                                        Simple CSS is only used for XHTML devices.

                                        We're currently doing a field analysis if there are some pure XHTML devices
                                        that have problems with this approach.

                                        > assuming it's the former, adding a chtml_avoid_accesskeys capability
                                        > would be consistent with what we have done in the past.

                                        Ok.

                                        > Question for you: do you have a list of CHTML devices for which this
                                        > capability should be set to false?

                                        I assume that nearly all of them support accesskeys, but I'll check that,

                                        > BTW: I have just emailed you a WURFL snapshot with the new capability.
                                        > Please let me know how it goes.

                                        Tnx. Works well.

                                        cu,
                                        boris
                                      • Luca Passani
                                        ... Personally, I don t particularly like this approach. It is the kind of hacky way to fix problems which will bit you in the backside when you least expect
                                        Message 19 of 21 , May 26, 2009
                                        • 0 Attachment
                                          Boris Folgmann wrote:
                                          >
                                          > The output complies to DTD XHTML Mobile 1.0, but we are omitting the
                                          > <!DOCTYPE> at the top, so cHTML or iHTML devices can read it, too.
                                          > On XHTML devices this enables quirks mode which allows us to use e.g.
                                          > <marquee> or <blink>. These tags are not in the DTD and are thefore not
                                          > accepted by a XHTML browser if the <!DOCTYPE> is specified.
                                          > Simple CSS is only used for XHTML devices.
                                          >
                                          > We're currently doing a field analysis if there are some pure XHTML devices
                                          > that have problems with this approach.
                                          >

                                          Personally, I don't particularly like this approach. It is the kind of
                                          "hacky" way to fix problems which will bit you in the backside when you
                                          least expect it. And when it does you will need to throw more hacks at
                                          the problem which will quickly make your application hard to maintain.
                                          IMO, a better option would be to make space for a cHTML view from the
                                          beginning and serve proper XHTML-MP to XHTML(MP) devices.

                                          Luca
                                        • Boris Folgmann
                                          Hi Luca, ... You re right. Here s the replacement for XHTML we re using: /* marquee is a CSS3 feature */ div.marquee {overflow: auto; overflow-style:
                                          Message 20 of 21 , Aug 4, 2009
                                          • 0 Attachment
                                            Hi Luca,

                                            Luca Passani schrieb/wrote:
                                            >> On XHTML devices this enables quirks mode which allows us to use e.g.
                                            >> <marquee> or <blink>. These tags are not in the DTD and are thefore not
                                            > Personally, I don't particularly like this approach. It is the kind of
                                            > "hacky" way to fix problems which will bit you in the backside when you

                                            You're right. Here's the replacement for XHTML we're using:

                                            /* marquee is a CSS3 feature */
                                            div.marquee {overflow: auto; overflow-style: marquee; white-space: nowrap;}
                                            div.blink {text-decoration: blink; display: inline;}

                                            WURFL defines a capability that determines if overflow-style: marquee is
                                            supported. If not it will simply not scroll.

                                            cu,
                                            boris
                                          Your message has been successfully submitted and would be delivered to recipients shortly.