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

Re: org.json.JSONML.java

Expand Messages
  • Douglas Crockford
    ... If you need that transformation, you should do the replacement before calling JSONML. I don t understand why you would want to encode JSON names with the
    Message 1 of 18 , Jul 24, 2008
    • 0 Attachment
      --- In json@yahoogroups.com, "Mark Joseph" <mark@...> wrote:
      > The prefix, in my
      > example is "P6R", but the prefix needs to be replaced by
      > its matching URI. So the "P6R:tagName" in an XML
      > processor
      > gets replaced as "URI:tagName" since the prefix is not
      > really important. So is that expansion done in JSONML?

      If you need that transformation, you should do the replacement before
      calling JSONML. I don't understand why you would want to encode JSON
      names with the URI form.
    • Stephen M. McKamey
      The reasoning for the URI form is that the prefix form is only an alias for the URI form. The prefix alias can be made to be anything whereas the true
      Message 2 of 18 , Jul 24, 2008
      • 0 Attachment
        The reasoning for the URI form is that the prefix form is only an
        alias
        for the URI form. The prefix alias can be made to be anything
        whereas
        the true namespace is the URI. Therefore to actually encode the
        entire
        namespaced tag or attribute name (without the associated namespace
        lookup context) it would need to include the full URI rather than the
        prefix.

        --- In json@yahoogroups.com, "Douglas Crockford" <douglas@...> wrote:
        >
        > --- In json@yahoogroups.com, "Mark Joseph" <mark@> wrote:
        > > The prefix, in my
        > > example is "P6R", but the prefix needs to be replaced by
        > > its matching URI. So the "P6R:tagName" in an XML
        > > processor
        > > gets replaced as "URI:tagName" since the prefix is not
        > > really important. So is that expansion done in JSONML?
        >
        > If you need that transformation, you should do the replacement
        before
        > calling JSONML. I don't understand why you would want to encode JSON
        > names with the URI form.
        >
      • Atif Aziz
        Like XML 1.0, JsonML does not need to care about namespaces. In XML 1.0, a colon (:) is a perfectly valid character in the element name. XML namespaces is
        Message 3 of 18 , Jul 24, 2008
        • 0 Attachment
          Like XML 1.0, JsonML does not need to care about namespaces. In XML 1.0, a colon (:) is a perfectly valid character in the element name. XML namespaces is nothing more than a standardized convention and interpretation of element and attributes names. Some XML parsers allow you to turn namespace handling on or off. Likewise, a JsonML encoder can look at XML without any regard for namespaces though the story could vary for a decoder. Here's an example. Suppose the following XML:

          <root xmlns:ns1="http://www.example.com/ns1"
          xmlns:ns2="http://www.example.com/ns2">
          <child />
          <ns1:child />
          <ns2:child />
          </root>

          In JsoML, this would become:

          [
          "root",
          {
          "xmlns:ns1": "http://www.example.com/ns1",
          "xmlns:ns2": "http://www.example.com/ns2"
          },
          ["child"],
          ["ns1:child"],
          ["ns2:child"]
          ]

          A JsonML decoder can decide to deliver the element and attribute names verbatim without any special interpretation of colon (:) in JSON object member names. If the user cares about URI + local name then he or she can put/pipe the XML delivered by the JsonML decoder through an XML parser implementation that will interpret and expose those properties. That's the approach Jayrock currently takes:

          http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/JsonML/JsonMLEncoder.cs
          http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/JsonML/JsonMLDecoder.cs

          - Atif


          From: json@yahoogroups.com [mailto:json@yahoogroups.com] On Behalf Of Mark Joseph
          Sent: Thursday, July 24, 2008 10:08 AM
          To: json@yahoogroups.com
          Subject: Re: [json] Re: org.json.JSONML.java


          Sorry I am just not getting this. The prefix, in my
          example is "P6R", but the prefix needs to be replaced by
          its matching URI. So the "P6R:tagName" in an XML
          processor
          gets replaced as "URI:tagName" since the prefix is not
          really important. So is that expansion done in JSONML?

          Best,
          Mark

          On Thu, 24 Jul 2008 05:35:53 -0000
          "Douglas Crockford" <douglas@...> wrote:
          > --- In json@yahoogroups.com, "Mark Joseph" <mark@...>
          >wrote:
          >> Just so I understand your stuff resolves the XML
          >> namespace prefix (like "P6R:tagName") to the URI
          >> "http://www.p6r...:tagName", right? Because the URI is
          >> the real
          >> namespace name.
          >
          > It transforms <P6R:tagName> into
          >{"tagName":"P6R:tagName"} or
          > ["P6R:tagName"] and back.
          >

          -------------------------
          Mark Joseph, Ph.D.
          President and Secretary
          P6R, Inc.
          http://www.p6r.com
          408-205-0361
          Fax: 831-476-7490
          Skype: markjoseph_sc
          IM: (Yahoo) mjoseph8888
          (AIM) mjoseph8888
        • Mark Joseph
          So when I read the original email about JSONML it was so that any XML document could be encoded. And sorry without namespaces that is not going to work very
          Message 4 of 18 , Jul 25, 2008
          • 0 Attachment
            So when I read the original email about JSONML it was so
            that any XML document could be encoded. And sorry
            without namespaces that is not going to work very well.
            Any XML tool, and we build several, that does not support
            namespaces these days is not going to be used by anyone
            doing serious work. So I suggest that the JSON group does
            not ignore namespaces but figures out how to deal with
            them,

            Best,
            Mark
            P6R, Inc


            On Thu, 24 Jul 2008 23:22:28 +0200
            Atif Aziz <atif.aziz@...> wrote:
            > Like XML 1.0, JsonML does not need to care about
            >namespaces. In XML 1.0, a colon (:) is a perfectly valid
            >character in the element name. XML namespaces is nothing
            >more than a standardized convention and interpretation of
            >element and attributes names. Some XML parsers allow you
            >to turn namespace handling on or off. Likewise, a JsonML
            >encoder can look at XML without any regard for namespaces
            >though the story could vary for a decoder. Here's an
            >example. Suppose the following XML:
            >
            > <root xmlns:ns1="http://www.example.com/ns1"
            > xmlns:ns2="http://www.example.com/ns2">
            > <child />
            > <ns1:child />
            > <ns2:child />
            > </root>
            >
            > In JsoML, this would become:
            >
            > [
            > "root",
            > {
            > "xmlns:ns1": "http://www.example.com/ns1",
            > "xmlns:ns2": "http://www.example.com/ns2"
            > },
            > ["child"],
            > ["ns1:child"],
            > ["ns2:child"]
            > ]
            >
            > A JsonML decoder can decide to deliver the element and
            >attribute names verbatim without any special
            >interpretation of colon (:) in JSON object member names.
            >If the user cares about URI + local name then he or she
            >can put/pipe the XML delivered by the JsonML decoder
            >through an XML parser implementation that will interpret
            >and expose those properties. That's the approach Jayrock
            >currently takes:
            >
            > http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/JsonML/JsonMLEncoder.cs
            > http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/JsonML/JsonMLDecoder.cs
            >
            > - Atif
            >
            >
            >From: json@yahoogroups.com [mailto:json@yahoogroups.com]
            >On Behalf Of Mark Joseph
            > Sent: Thursday, July 24, 2008 10:08 AM
            > To: json@yahoogroups.com
            > Subject: Re: [json] Re: org.json.JSONML.java
            >
            >
            > Sorry I am just not getting this. The prefix, in my
            > example is "P6R", but the prefix needs to be replaced by
            > its matching URI. So the "P6R:tagName" in an XML
            > processor
            > gets replaced as "URI:tagName" since the prefix is not
            > really important. So is that expansion done in JSONML?
            >
            > Best,
            > Mark
            >
            > On Thu, 24 Jul 2008 05:35:53 -0000
            > "Douglas Crockford" <douglas@...> wrote:
            >> --- In json@yahoogroups.com, "Mark Joseph" <mark@...>
            >>wrote:
            >>> Just so I understand your stuff resolves the XML
            >>> namespace prefix (like "P6R:tagName") to the URI
            >>> "http://www.p6r...:tagName", right? Because the URI is
            >>> the real
            >>> namespace name.
            >>
            >> It transforms <P6R:tagName> into
            >>{"tagName":"P6R:tagName"} or
            >> ["P6R:tagName"] and back.
            >>
            >
            > -------------------------
            > Mark Joseph, Ph.D.
            > President and Secretary
            > P6R, Inc.
            > http://www.p6r.com
            > 408-205-0361
            >Fax: 831-476-7490
            > Skype: markjoseph_sc
            > IM: (Yahoo) mjoseph8888
            > (AIM) mjoseph8888
            >

            -------------------------
            Mark Joseph, Ph.D.
            President and Secretary
            P6R, Inc.
            http://www.p6r.com
            408-205-0361
            Fax: 831-476-7490
            Skype: markjoseph_sc
            IM: (Yahoo) mjoseph8888
            (AIM) mjoseph8888
          • Stephen M. McKamey
            I definitely agree. XML w/o namespaces is virtually useless these days. There is nothing about the format of JsonML that inhibits namespaces. Implementations
            Message 5 of 18 , Jul 25, 2008
            • 0 Attachment
              I definitely agree. XML w/o namespaces is virtually useless these
              days.

              There is nothing about the format of JsonML that inhibits
              namespaces. Implementations either need to support some storage of
              the prefix/namespace aliasing (as Atif is showing) or be able to
              resolve the namespaces and use the fully qualified node names (as
              Mark is asking for). It seems that both are valid, just as they are
              in XML.

              Ignoring or dropping namespaces is not a good idea.

              --- In json@yahoogroups.com, "Mark Joseph" <mark@...> wrote:
              >
              > So when I read the original email about JSONML it was so
              > that any XML document could be encoded. And sorry
              > without namespaces that is not going to work very well.
              > Any XML tool, and we build several, that does not support
              > namespaces these days is not going to be used by anyone
              > doing serious work. So I suggest that the JSON group does
              > not ignore namespaces but figures out how to deal with
              > them,
              >
              > Best,
              > Mark
              > P6R, Inc
              >
              >
              > On Thu, 24 Jul 2008 23:22:28 +0200
              > Atif Aziz <atif.aziz@...> wrote:
              > > Like XML 1.0, JsonML does not need to care about
              > >namespaces. In XML 1.0, a colon (:) is a perfectly valid
              > >character in the element name. XML namespaces is nothing
              > >more than a standardized convention and interpretation of
              > >element and attributes names. Some XML parsers allow you
              > >to turn namespace handling on or off. Likewise, a JsonML
              > >encoder can look at XML without any regard for namespaces
              > >though the story could vary for a decoder. Here's an
              > >example. Suppose the following XML:
              > >
              > > <root xmlns:ns1="http://www.example.com/ns1"
              > > xmlns:ns2="http://www.example.com/ns2">
              > > <child />
              > > <ns1:child />
              > > <ns2:child />
              > > </root>
              > >
              > > In JsoML, this would become:
              > >
              > > [
              > > "root",
              > > {
              > > "xmlns:ns1": "http://www.example.com/ns1",
              > > "xmlns:ns2": "http://www.example.com/ns2"
              > > },
              > > ["child"],
              > > ["ns1:child"],
              > > ["ns2:child"]
              > > ]
              > >
              > > A JsonML decoder can decide to deliver the element and
              > >attribute names verbatim without any special
              > >interpretation of colon (:) in JSON object member names.
              > >If the user cares about URI + local name then he or she
              > >can put/pipe the XML delivered by the JsonML decoder
              > >through an XML parser implementation that will interpret
              > >and expose those properties. That's the approach Jayrock
              > >currently takes:
              > >
              > >
              http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/Jso
              nML/JsonMLEncoder.cs
              > >
              http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/Jso
              nML/JsonMLDecoder.cs
              > >
              > > - Atif
              > >
              > >
              > >From: json@yahoogroups.com [mailto:json@yahoogroups.com]
              > >On Behalf Of Mark Joseph
              > > Sent: Thursday, July 24, 2008 10:08 AM
              > > To: json@yahoogroups.com
              > > Subject: Re: [json] Re: org.json.JSONML.java
              > >
              > >
              > > Sorry I am just not getting this. The prefix, in my
              > > example is "P6R", but the prefix needs to be replaced by
              > > its matching URI. So the "P6R:tagName" in an XML
              > > processor
              > > gets replaced as "URI:tagName" since the prefix is not
              > > really important. So is that expansion done in JSONML?
              > >
              > > Best,
              > > Mark
              > >
              > > On Thu, 24 Jul 2008 05:35:53 -0000
              > > "Douglas Crockford" <douglas@...> wrote:
              > >> --- In json@yahoogroups.com, "Mark Joseph" <mark@>
              > >>wrote:
              > >>> Just so I understand your stuff resolves the XML
              > >>> namespace prefix (like "P6R:tagName") to the URI
              > >>> "http://www.p6r...:tagName", right? Because the URI is
              > >>> the real
              > >>> namespace name.
              > >>
              > >> It transforms <P6R:tagName> into
              > >>{"tagName":"P6R:tagName"} or
              > >> ["P6R:tagName"] and back.
              > >>
              > >
              > > -------------------------
              > > Mark Joseph, Ph.D.
              > > President and Secretary
              > > P6R, Inc.
              > > http://www.p6r.com
              > > 408-205-0361
              > >Fax: 831-476-7490
              > > Skype: markjoseph_sc
              > > IM: (Yahoo) mjoseph8888
              > > (AIM) mjoseph8888
              > >
              >
              > -------------------------
              > Mark Joseph, Ph.D.
              > President and Secretary
              > P6R, Inc.
              > http://www.p6r.com
              > 408-205-0361
              > Fax: 831-476-7490
              > Skype: markjoseph_sc
              > IM: (Yahoo) mjoseph8888
              > (AIM) mjoseph8888
              >
            • Tatu Saloranta
              ... Me three. Most recent xml parsers have started dropping support for non-namespace documents, too, essentially considering XML Namespaces to be integral
              Message 6 of 18 , Jul 25, 2008
              • 0 Attachment
                On Fri, Jul 25, 2008 at 10:19 AM, Stephen M. McKamey <jsonml@...> wrote:
                > I definitely agree. XML w/o namespaces is virtually useless these
                > days.
                >

                Me three. Most recent xml parsers have started dropping support for
                non-namespace documents, too, essentially considering XML Namespaces
                to be integral part of core xml processing.

                Of course, I really hope that the current craze of trying to use json
                via xml tools will die soon and we will have good "native" json
                processing systems above parser/generator level; but until then
                conversions should be done properly.

                -+ Tatu +-
              • Atif Aziz
                ... An encoder can quite dumbly format an XML document into JsonML without explicitly dealing with namespaces (provided that the XML API provides access to
                Message 7 of 18 , Jul 25, 2008
                • 0 Attachment
                  > XML w/o namespaces is virtually useless these days.
                  > Ignoring or dropping namespaces is not a good idea.

                  Definitely. I wasn't trying to question the merits of namespaces. My point was pretty much what you stated more clearly here:

                  > There is nothing about the format of JsonML that inhibits
                  > namespaces.

                  An encoder can quite "dumbly" format an XML document into JsonML without explicitly dealing with namespaces (provided that the XML API provides access to the QName and does not hide so-designated namespace declarations when you enumerate over attributes) and no information would be lost. In this sense, the encoder is acting like a pass-through and simply mapping XML nodes into JSON values as per JsonML transformation rules. All that would matter is that the input XML is well-formed though not necessarily valid. This should answer the question raised earlier, "So is that expansion done in JSONML?" The expansion is not and does not need to be done in the encoded JsonML.

                  A decoder implementation, on the other hand, has two API choices. Either it can directly support the abstractions introduced by XML namespaces via its API or it can delegate that work to a downstream and existing XML API. It's a detail of the implementation rather than the format.


                  From: json@yahoogroups.com [mailto:json@yahoogroups.com] On Behalf Of Stephen M. McKamey
                  Sent: Friday, July 25, 2008 7:20 PM
                  To: json@yahoogroups.com
                  Subject: [json] Re: org.json.JSONML.java

                  I definitely agree. XML w/o namespaces is virtually useless these
                  days.

                  There is nothing about the format of JsonML that inhibits
                  namespaces. Implementations either need to support some storage of
                  the prefix/namespace aliasing (as Atif is showing) or be able to
                  resolve the namespaces and use the fully qualified node names (as
                  Mark is asking for). It seems that both are valid, just as they are
                  in XML.

                  Ignoring or dropping namespaces is not a good idea.

                  --- In json@yahoogroups.com, "Mark Joseph" <mark@...> wrote:
                  >
                  > So when I read the original email about JSONML it was so
                  > that any XML document could be encoded. And sorry
                  > without namespaces that is not going to work very well.
                  > Any XML tool, and we build several, that does not support
                  > namespaces these days is not going to be used by anyone
                  > doing serious work. So I suggest that the JSON group does
                  > not ignore namespaces but figures out how to deal with
                  > them,
                  >
                  > Best,
                  > Mark
                  > P6R, Inc
                  >
                  >
                  > On Thu, 24 Jul 2008 23:22:28 +0200
                  > Atif Aziz <atif.aziz@...> wrote:
                  > > Like XML 1.0, JsonML does not need to care about
                  > >namespaces. In XML 1.0, a colon (:) is a perfectly valid
                  > >character in the element name. XML namespaces is nothing
                  > >more than a standardized convention and interpretation of
                  > >element and attributes names. Some XML parsers allow you
                  > >to turn namespace handling on or off. Likewise, a JsonML
                  > >encoder can look at XML without any regard for namespaces
                  > >though the story could vary for a decoder. Here's an
                  > >example. Suppose the following XML:
                  > >
                  > > <root xmlns:ns1="http://www.example.com/ns1"
                  > > xmlns:ns2="http://www.example.com/ns2">
                  > > <child />
                  > > <ns1:child />
                  > > <ns2:child />
                  > > </root>
                  > >
                  > > In JsoML, this would become:
                  > >
                  > > [
                  > > "root",
                  > > {
                  > > "xmlns:ns1": "http://www.example.com/ns1",
                  > > "xmlns:ns2": "http://www.example.com/ns2"
                  > > },
                  > > ["child"],
                  > > ["ns1:child"],
                  > > ["ns2:child"]
                  > > ]
                  > >
                  > > A JsonML decoder can decide to deliver the element and
                  > >attribute names verbatim without any special
                  > >interpretation of colon (:) in JSON object member names.
                  > >If the user cares about URI + local name then he or she
                  > >can put/pipe the XML delivered by the JsonML decoder
                  > >through an XML parser implementation that will interpret
                  > >and expose those properties. That's the approach Jayrock
                  > >currently takes:
                  > >
                  > >
                  http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/Jso
                  nML/JsonMLEncoder.cs
                  > >
                  http://svn.berlios.de/svnroot/repos/jayrock/trunk/src/Jayrock.Json/Jso
                  nML/JsonMLDecoder.cs
                  > >
                  > > - Atif
                  > >
                  > >
                  > >From: json@yahoogroups.com [mailto:json@yahoogroups.com]
                  > >On Behalf Of Mark Joseph
                  > > Sent: Thursday, July 24, 2008 10:08 AM
                  > > To: json@yahoogroups.com
                  > > Subject: Re: [json] Re: org.json.JSONML.java
                  > >
                  > >
                  > > Sorry I am just not getting this. The prefix, in my
                  > > example is "P6R", but the prefix needs to be replaced by
                  > > its matching URI. So the "P6R:tagName" in an XML
                  > > processor
                  > > gets replaced as "URI:tagName" since the prefix is not
                  > > really important. So is that expansion done in JSONML?
                  > >
                  > > Best,
                  > > Mark
                  > >
                  > > On Thu, 24 Jul 2008 05:35:53 -0000
                  > > "Douglas Crockford" <douglas@...> wrote:
                  > >> --- In json@yahoogroups.com, "Mark Joseph" <mark@>
                  > >>wrote:
                  > >>> Just so I understand your stuff resolves the XML
                  > >>> namespace prefix (like "P6R:tagName") to the URI
                  > >>> "http://www.p6r...:tagName", right? Because the URI is
                  > >>> the real
                  > >>> namespace name.
                  > >>
                  > >> It transforms <P6R:tagName> into
                  > >>{"tagName":"P6R:tagName"} or
                  > >> ["P6R:tagName"] and back.
                  > >>
                  > >
                  > > -------------------------
                  > > Mark Joseph, Ph.D.
                  > > President and Secretary
                  > > P6R, Inc.
                  > > http://www.p6r.com
                  > > 408-205-0361
                  > >Fax: 831-476-7490
                  > > Skype: markjoseph_sc
                  > > IM: (Yahoo) mjoseph8888
                  > > (AIM) mjoseph8888
                  > >
                  >
                  > -------------------------
                  > Mark Joseph, Ph.D.
                  > President and Secretary
                  > P6R, Inc.
                  > http://www.p6r.com
                  > 408-205-0361
                  > Fax: 831-476-7490
                  > Skype: markjoseph_sc
                  > IM: (Yahoo) mjoseph8888
                  > (AIM) mjoseph8888
                  >
                • Kyle Alan Hale
                  I d like some clarification: the tagName/childNodes syntax is the key difference between Stephen s JSONML and my alternative, JSoda
                  Message 8 of 18 , Aug 1, 2008
                  • 0 Attachment
                    I'd like some clarification: the tagName/childNodes syntax is the key
                    difference between Stephen's JSONML and my alternative, JSoda
                    <http://jsoda.info/ <http://jsoda.info/> >. In fact, your "Object Form"
                    example is a perfect example of a JSoda object. Was this intentional,
                    or a coincidence? Intentional or not, I'd prefer that you didn't refer
                    to such syntax as JSONML, since the lack of such syntax is the sole
                    reason why I created JSoda. To my knowledge, JSoda predates this new
                    "Object Form". If that's true, I'd appreciate compliance with the JSoda
                    license: <http://jsoda.info/License <http://jsoda.info/License> >. I'll
                    eat my hat before I let JSoda be referred to as JSONML.

                    To give some background, I posted JSoda last year as an alternative to
                    JSONML. As I have been using it over the last year, I have realized
                    that JSONML has a smaller size than JSoda, and so is more fit for
                    storage or transmission. Actually, the end result of the realization
                    was that XHTML is much smaller than either, so I use it for storage of
                    DOM representations.

                    However, I still feel that JSoda's syntax is much more suited as a
                    format for a DOM builder than JSONML, because of the use of the tagName
                    and childNodes properties, leading to (in my opinion) a much more
                    intuitively nested XHTML representation.

                    Read more about how to use JSoda as a format for a DOM builder here:
                    <http://jsoda.info/Object.toDOM <http://jsoda.info/Object.toDOM> >, and
                    about JSoda syntax here: <http://jsoda.info/JSoda+is#simple
                    <http://jsoda.info/JSoda+is#simple> >.

                    --- In json@yahoogroups.com, "Douglas Crockford" <douglas@...> wrote:
                    >
                    > JsonML is an isomorphic transformation between JSON and XML. With this
                    > transformation, JSON is able to encode XML document structures. The
                    > Array Form represents a node as an array whose first slot is the
                    > tagName and second slot optionally contains an object of attributes.
                    > The remaining slots contain the node's children.
                    >
                    > So
                    >
                    > <div id="demo" class="JSONML"><p>JSONML is a transformation
                    > between<b>JSON</b>and<b>XML</b>that preserves ordering of document
                    > features.</p><p>JSONML can work with JSON arrays or JSON
                    > objects.</p><p>Three<br/>little<br/>words</p></div>
                    >
                    > is equivalent to
                    >
                    > [
                    > "div",
                    > {
                    > "class": "JSONML",
                    > "id": "demo"
                    > },
                    > [
                    > "p",
                    > "JSONML is a transformation between",
                    > [
                    > "b",
                    > "JSON"
                    > ],
                    > "and",
                    > [
                    > "b",
                    > "XML"
                    > ],
                    > "that preserves ordering of document features."
                    > ],
                    > [
                    > "p",
                    > "JSONML can work with JSON arrays or JSON objects."
                    > ],
                    > [
                    > "p",
                    > "Three",
                    > ["br"],
                    > "little",
                    > ["br"],
                    > "words"
                    > ]
                    > ]
                    >
                    > I added the Object Form to JSONML.java. It represents a node as an
                    > object. The node's attributes are the object's properties. The
                    > "tagName" property is the tagName, and the "childNodes" property is an
                    > array of objects.
                    >
                    > {
                    > "childNodes": [
                    > {
                    > "childNodes": [
                    > "JSONML is a transformation between",
                    > {
                    > "childNodes": ["JSON"],
                    > "tagName": "b"
                    > },
                    > "and",
                    > {
                    > "childNodes": ["XML"],
                    > "tagName": "b"
                    > },
                    > "that preserves ordering of document features."
                    > ],
                    > "tagName": "p"
                    > },
                    > {
                    > "childNodes":
                    > ["JSONML can work with JSON arrays or JSON objects."],
                    > "tagName": "p"
                    > },
                    > {
                    > "childNodes": [
                    > "Three",
                    > {"tagName": "br"},
                    > "little",
                    > {"tagName": "br"},
                    > "words"
                    > ],
                    > "tagName": "p"
                    > }
                    > ],
                    > "class": "JSONML",
                    > "id": "demo",
                    > "tagName": "div"
                    > }
                    >




                    [Non-text portions of this message have been removed]
                  • Douglas Crockford
                    ... The similarity between JSONML Object Form and JSoda is completely coincidental. Your jsoda license it limited to the use of your code. I am not using your
                    Message 9 of 18 , Aug 1, 2008
                    • 0 Attachment
                      --- In json@yahoogroups.com, "Kyle Alan Hale" <kylealanhale@...> wrote:
                      >
                      > I'd like some clarification: the tagName/childNodes syntax is the key
                      > difference between Stephen's JSONML and my alternative, JSoda
                      > <http://jsoda.info/ <http://jsoda.info/> >. In fact, your "Object Form"
                      > example is a perfect example of a JSoda object. Was this intentional,
                      > or a coincidence? Intentional or not, I'd prefer that you didn't refer
                      > to such syntax as JSONML, since the lack of such syntax is the sole
                      > reason why I created JSoda. To my knowledge, JSoda predates this new
                      > "Object Form". If that's true, I'd appreciate compliance with the JSoda
                      > license: <http://jsoda.info/License <http://jsoda.info/License> >. I'll
                      > eat my hat before I let JSoda be referred to as JSONML.


                      The similarity between JSONML Object Form and JSoda is completely
                      coincidental.

                      Your jsoda license it limited to the use of your code. I am not using
                      your code. I have never looked at your code. If you had secured a
                      trademark on JSoda, you might be able to control the way people can
                      use the term. There is no way you can compel anyone to use your term.

                      What you do with your hat is your business.
                    • Stephen M. McKamey
                      Kyle, I think it is important to note that what we are trying to do here is foster innovation and build upon each other s ideas to produce a better set of
                      Message 10 of 18 , Aug 1, 2008
                      • 0 Attachment
                        Kyle,

                        I think it is important to note that what we are trying to do here is
                        foster innovation and build upon each other's ideas to produce a
                        better set of tools to perform our projects and day jobs. If it is
                        just your hope to get famous by bashing others, good luck to you.
                        (However, I suggest you beware of going against Crockford!)

                        JSoda may suit your needs but it didn't meet some of the requirements
                        which went into the design of JsonML. (In fact the first couple
                        revisions of the JsonML grammar looked nearly identical.) JsonML
                        isn't limited to purely XHTML, even though it does that very well. It
                        is meant to be a round-trip-able format between valid JSON and any
                        valid XML fragment. Key considerations were two-way conversion and
                        compactness of form.

                        JsonML *wasn't* meant to be the way that everything that *could* be
                        expressed in XML *should* be expressed in JSON. This is the point
                        that I believe Crockford was stating when he started this thread.
                        Both "object form", as Crockford calls it, and JsonML ("document
                        form") are useful for different purposes. Just because you've named
                        an obvious object model doesn't necessarily mean people will jump out
                        of their seats to use it.

                        I hope that this makes the situation clearer,
                        Stephen

                        BTW, I think Noam Chomsky
                        (http://en.wikipedia.org/wiki/Chomsky_hierarchy) would have an issue
                        with your "Not a Language!"
                        (http://jsoda.info/JSoda+is#not+a+language) argument. Formal Language
                        / Automata Theory doesn't limit the use of the word "language" to mean
                        "programming language". As such, JSON and JsonML are both languages
                        with explicit grammars. Off topic, but this bugs me every time I
                        click over to your site to remind myself what JSoda is.


                        --- In json@yahoogroups.com, "Kyle Alan Hale" <kylealanhale@...> wrote:
                        >
                        > I'd like some clarification: the tagName/childNodes syntax is the key
                        > difference between Stephen's JSONML and my alternative, JSoda
                        > <http://jsoda.info/ <http://jsoda.info/> >. In fact, your "Object Form"
                        > example is a perfect example of a JSoda object. Was this intentional,
                        > or a coincidence? Intentional or not, I'd prefer that you didn't refer
                        > to such syntax as JSONML, since the lack of such syntax is the sole
                        > reason why I created JSoda. To my knowledge, JSoda predates this new
                        > "Object Form". If that's true, I'd appreciate compliance with the JSoda
                        > license: <http://jsoda.info/License <http://jsoda.info/License> >. I'll
                        > eat my hat before I let JSoda be referred to as JSONML.
                        >
                        > To give some background, I posted JSoda last year as an alternative to
                        > JSONML. As I have been using it over the last year, I have realized
                        > that JSONML has a smaller size than JSoda, and so is more fit for
                        > storage or transmission. Actually, the end result of the realization
                        > was that XHTML is much smaller than either, so I use it for storage of
                        > DOM representations.
                        >
                        > However, I still feel that JSoda's syntax is much more suited as a
                        > format for a DOM builder than JSONML, because of the use of the tagName
                        > and childNodes properties, leading to (in my opinion) a much more
                        > intuitively nested XHTML representation.
                        >
                        > Read more about how to use JSoda as a format for a DOM builder here:
                        > <http://jsoda.info/Object.toDOM <http://jsoda.info/Object.toDOM> >, and
                        > about JSoda syntax here: <http://jsoda.info/JSoda+is#simple
                        > <http://jsoda.info/JSoda+is#simple> >.
                        >
                        > --- In json@yahoogroups.com, "Douglas Crockford" <douglas@> wrote:
                        > >
                        > > JsonML is an isomorphic transformation between JSON and XML. With this
                        > > transformation, JSON is able to encode XML document structures. The
                        > > Array Form represents a node as an array whose first slot is the
                        > > tagName and second slot optionally contains an object of attributes.
                        > > The remaining slots contain the node's children.
                        > >
                        > > So
                        > >
                        > > <div id="demo" class="JSONML"><p>JSONML is a transformation
                        > > between<b>JSON</b>and<b>XML</b>that preserves ordering of document
                        > > features.</p><p>JSONML can work with JSON arrays or JSON
                        > > objects.</p><p>Three<br/>little<br/>words</p></div>
                        > >
                        > > is equivalent to
                        > >
                        > > [
                        > > "div",
                        > > {
                        > > "class": "JSONML",
                        > > "id": "demo"
                        > > },
                        > > [
                        > > "p",
                        > > "JSONML is a transformation between",
                        > > [
                        > > "b",
                        > > "JSON"
                        > > ],
                        > > "and",
                        > > [
                        > > "b",
                        > > "XML"
                        > > ],
                        > > "that preserves ordering of document features."
                        > > ],
                        > > [
                        > > "p",
                        > > "JSONML can work with JSON arrays or JSON objects."
                        > > ],
                        > > [
                        > > "p",
                        > > "Three",
                        > > ["br"],
                        > > "little",
                        > > ["br"],
                        > > "words"
                        > > ]
                        > > ]
                        > >
                        > > I added the Object Form to JSONML.java. It represents a node as an
                        > > object. The node's attributes are the object's properties. The
                        > > "tagName" property is the tagName, and the "childNodes" property is an
                        > > array of objects.
                        > >
                        > > {
                        > > "childNodes": [
                        > > {
                        > > "childNodes": [
                        > > "JSONML is a transformation between",
                        > > {
                        > > "childNodes": ["JSON"],
                        > > "tagName": "b"
                        > > },
                        > > "and",
                        > > {
                        > > "childNodes": ["XML"],
                        > > "tagName": "b"
                        > > },
                        > > "that preserves ordering of document features."
                        > > ],
                        > > "tagName": "p"
                        > > },
                        > > {
                        > > "childNodes":
                        > > ["JSONML can work with JSON arrays or JSON objects."],
                        > > "tagName": "p"
                        > > },
                        > > {
                        > > "childNodes": [
                        > > "Three",
                        > > {"tagName": "br"},
                        > > "little",
                        > > {"tagName": "br"},
                        > > "words"
                        > > ],
                        > > "tagName": "p"
                        > > }
                        > > ],
                        > > "class": "JSONML",
                        > > "id": "demo",
                        > > "tagName": "div"
                        > > }
                        > >
                        >
                        >
                        >
                        >
                        > [Non-text portions of this message have been removed]
                        >
                      • Fang Yidong
                        Why JSoda? No way to me ... :-) ... -- JSON: Action in AJAX! JSON - http://www.json.org JSON.simple - http://www.json.org/java/simple.txt
                        Message 11 of 18 , Aug 1, 2008
                        • 0 Attachment
                          Why JSoda? No way to me ... :-)

                          --- Douglas Crockford <douglas@...>:

                          > --- In json@yahoogroups.com, "Kyle Alan Hale"
                          > <kylealanhale@...> wrote:
                          > >
                          > > I'd like some clarification: the
                          > tagName/childNodes syntax is the key
                          > > difference between Stephen's JSONML and my
                          > alternative, JSoda
                          > > <http://jsoda.info/ <http://jsoda.info/> >. In
                          > fact, your "Object Form"
                          > > example is a perfect example of a JSoda object.
                          > Was this intentional,
                          > > or a coincidence? Intentional or not, I'd prefer
                          > that you didn't refer
                          > > to such syntax as JSONML, since the lack of such
                          > syntax is the sole
                          > > reason why I created JSoda. To my knowledge,
                          > JSoda predates this new
                          > > "Object Form". If that's true, I'd appreciate
                          > compliance with the JSoda
                          > > license: <http://jsoda.info/License
                          > <http://jsoda.info/License> >. I'll
                          > > eat my hat before I let JSoda be referred to as
                          > JSONML.
                          >
                          >
                          > The similarity between JSONML Object Form and JSoda
                          > is completely
                          > coincidental.
                          >
                          > Your jsoda license it limited to the use of your
                          > code. I am not using
                          > your code. I have never looked at your code. If you
                          > had secured a
                          > trademark on JSoda, you might be able to control the
                          > way people can
                          > use the term. There is no way you can compel anyone
                          > to use your term.
                          >
                          > What you do with your hat is your business.
                          >
                          >



                          --
                          JSON: Action in AJAX!

                          JSON - http://www.json.org
                          JSON.simple - http://www.json.org/java/simple.txt



                          ___________________________________________________________
                          雅虎邮箱,您的终生邮箱!
                          http://cn.mail.yahoo.com/
                        • Kyle Alan Hale
                          Point(s) well taken. Especially the idea of applying Chomskyan classifications to these subsets. However, some further clarifications: * I neither want to
                          Message 12 of 18 , Aug 1, 2008
                          • 0 Attachment
                            Point(s) well taken. Especially the idea of applying Chomskyan
                            classifications to these subsets. However, some further clarifications:

                            * I neither want to "get famous" nor bash others, especially Doug. I
                            agree that we're all here to build upon each others' ideas to more
                            effectively solve our day-to-day problems.
                            * The goal of JSoda is to provide a syntax that JsonML lacks. To
                            paraphrase you: JsonML may suit your needs but it didn't meet some of
                            the requirements which went into the design of JSoda.
                            * Doug was mistaken about the terms of the licensing; it also covers
                            the documentation, which includes the syntax. As Doug made quite
                            clear, I have no real control over the use of JSoda; in fact, the
                            license clearly promotes the free use and implementation of these
                            ideas. One can't protect an idea, only a product. However, it should
                            be noted that intellectual property laws are effective whether
                            registered or not, and ignorance of pre-existing intellectual property
                            does not exempt one from those laws.

                            My only goal in responding to Doug's original post was to clarify that
                            his Object Form isn't JsonML, as your site will confirm with its
                            very clear grammar form. If it is to be an alternate syntax for
                            JsonML, then let's make it that way, officially, and I'll swallow my
                            hat. In that case, I offer a ready-to-go DOM builder for the
                            JsonML/JSoda syntax: http://jsoda.info/JSoda.toDOM

                            In the mean time, I'll update http://jsoda.info/ to reflect your point
                            on these subsets as languages, and to trim other unnecessary fluff,
                            including anything that could be construed as bashing. I like to
                            write in a fairly relaxed style, but I can see how some points on the
                            site could be misconstrued. I apologize for that.

                            --- In json@yahoogroups.com, "Stephen M. McKamey" <jsonml@...> wrote:
                            >
                            > Kyle,
                            >
                            > I think it is important to note that what we are trying to do here is
                            > foster innovation and build upon each other's ideas to produce a
                            > better set of tools to perform our projects and day jobs. If it is
                            > just your hope to get famous by bashing others, good luck to you.
                            > (However, I suggest you beware of going against Crockford!)
                            >
                            > JSoda may suit your needs but it didn't meet some of the requirements
                            > which went into the design of JsonML. (In fact the first couple
                            > revisions of the JsonML grammar looked nearly identical.) JsonML
                            > isn't limited to purely XHTML, even though it does that very well. It
                            > is meant to be a round-trip-able format between valid JSON and any
                            > valid XML fragment. Key considerations were two-way conversion and
                            > compactness of form.
                            >
                            > JsonML *wasn't* meant to be the way that everything that *could* be
                            > expressed in XML *should* be expressed in JSON. This is the point
                            > that I believe Crockford was stating when he started this thread.
                            > Both "object form", as Crockford calls it, and JsonML ("document
                            > form") are useful for different purposes. Just because you've named
                            > an obvious object model doesn't necessarily mean people will jump out
                            > of their seats to use it.
                            >
                            > I hope that this makes the situation clearer,
                            > Stephen
                            >
                            > BTW, I think Noam Chomsky
                            > (http://en.wikipedia.org/wiki/Chomsky_hierarchy) would have an issue
                            > with your "Not a Language!"
                            > (http://jsoda.info/JSoda+is#not+a+language) argument. Formal Language
                            > / Automata Theory doesn't limit the use of the word "language" to mean
                            > "programming language". As such, JSON and JsonML are both languages
                            > with explicit grammars. Off topic, but this bugs me every time I
                            > click over to your site to remind myself what JSoda is.
                            >
                            >
                            > --- In json@yahoogroups.com, "Kyle Alan Hale" <kylealanhale@> wrote:
                            > >
                            > > I'd like some clarification: the tagName/childNodes syntax is the key
                            > > difference between Stephen's JSONML and my alternative, JSoda
                            > > <http://jsoda.info/ <http://jsoda.info/> >. In fact, your "Object
                            Form"
                            > > example is a perfect example of a JSoda object. Was this intentional,
                            > > or a coincidence? Intentional or not, I'd prefer that you didn't
                            refer
                            > > to such syntax as JSONML, since the lack of such syntax is the sole
                            > > reason why I created JSoda. To my knowledge, JSoda predates this new
                            > > "Object Form". If that's true, I'd appreciate compliance with the
                            JSoda
                            > > license: <http://jsoda.info/License <http://jsoda.info/License> >.
                            I'll
                            > > eat my hat before I let JSoda be referred to as JSONML.
                            > >
                            > > To give some background, I posted JSoda last year as an alternative to
                            > > JSONML. As I have been using it over the last year, I have realized
                            > > that JSONML has a smaller size than JSoda, and so is more fit for
                            > > storage or transmission. Actually, the end result of the realization
                            > > was that XHTML is much smaller than either, so I use it for storage of
                            > > DOM representations.
                            > >
                            > > However, I still feel that JSoda's syntax is much more suited as a
                            > > format for a DOM builder than JSONML, because of the use of the
                            tagName
                            > > and childNodes properties, leading to (in my opinion) a much more
                            > > intuitively nested XHTML representation.
                            > >
                            > > Read more about how to use JSoda as a format for a DOM builder here:
                            > > <http://jsoda.info/Object.toDOM <http://jsoda.info/Object.toDOM>
                            >, and
                            > > about JSoda syntax here: <http://jsoda.info/JSoda+is#simple
                            > > <http://jsoda.info/JSoda+is#simple> >.
                            > >
                            > > --- In json@yahoogroups.com, "Douglas Crockford" <douglas@> wrote:
                            > > >
                            > > > JsonML is an isomorphic transformation between JSON and XML.
                            With this
                            > > > transformation, JSON is able to encode XML document structures. The
                            > > > Array Form represents a node as an array whose first slot is the
                            > > > tagName and second slot optionally contains an object of attributes.
                            > > > The remaining slots contain the node's children.
                            > > >
                            > > > So
                            > > >
                            > > > <div id="demo" class="JSONML"><p>JSONML is a transformation
                            > > > between<b>JSON</b>and<b>XML</b>that preserves ordering of document
                            > > > features.</p><p>JSONML can work with JSON arrays or JSON
                            > > > objects.</p><p>Three<br/>little<br/>words</p></div>
                            > > >
                            > > > is equivalent to
                            > > >
                            > > > [
                            > > > "div",
                            > > > {
                            > > > "class": "JSONML",
                            > > > "id": "demo"
                            > > > },
                            > > > [
                            > > > "p",
                            > > > "JSONML is a transformation between",
                            > > > [
                            > > > "b",
                            > > > "JSON"
                            > > > ],
                            > > > "and",
                            > > > [
                            > > > "b",
                            > > > "XML"
                            > > > ],
                            > > > "that preserves ordering of document features."
                            > > > ],
                            > > > [
                            > > > "p",
                            > > > "JSONML can work with JSON arrays or JSON objects."
                            > > > ],
                            > > > [
                            > > > "p",
                            > > > "Three",
                            > > > ["br"],
                            > > > "little",
                            > > > ["br"],
                            > > > "words"
                            > > > ]
                            > > > ]
                            > > >
                            > > > I added the Object Form to JSONML.java. It represents a node as an
                            > > > object. The node's attributes are the object's properties. The
                            > > > "tagName" property is the tagName, and the "childNodes" property
                            is an
                            > > > array of objects.
                            > > >
                            > > > {
                            > > > "childNodes": [
                            > > > {
                            > > > "childNodes": [
                            > > > "JSONML is a transformation between",
                            > > > {
                            > > > "childNodes": ["JSON"],
                            > > > "tagName": "b"
                            > > > },
                            > > > "and",
                            > > > {
                            > > > "childNodes": ["XML"],
                            > > > "tagName": "b"
                            > > > },
                            > > > "that preserves ordering of document features."
                            > > > ],
                            > > > "tagName": "p"
                            > > > },
                            > > > {
                            > > > "childNodes":
                            > > > ["JSONML can work with JSON arrays or JSON objects."],
                            > > > "tagName": "p"
                            > > > },
                            > > > {
                            > > > "childNodes": [
                            > > > "Three",
                            > > > {"tagName": "br"},
                            > > > "little",
                            > > > {"tagName": "br"},
                            > > > "words"
                            > > > ],
                            > > > "tagName": "p"
                            > > > }
                            > > > ],
                            > > > "class": "JSONML",
                            > > > "id": "demo",
                            > > > "tagName": "div"
                            > > > }
                            > > >
                            > >
                            > >
                            > >
                            > >
                            > > [Non-text portions of this message have been removed]
                            > >
                            >
                          Your message has been successfully submitted and would be delivered to recipients shortly.