You are viewing a plain text version of this content. The canonical link for it is here.
Posted to j-dev@xerces.apache.org by Libor Kramolis <li...@sun.com> on 2001/10/03 19:34:52 UTC
[XNI] Native tree model
Hello Andy.
In your Xerces2 Workshop (after this year's JavaOne) presentetion
(http://www.apache.org/~andyc/speeches/20010611/xerces2ws.ppt), slide 19,
you are speaking about 'native tree model' as future extension of XNI.
I would like to ask you, what did you mean by that. Do you have any proposal
or plan about it? What is it? What it will be? What I can imagine about NTM
name is something like java oriented and extended DOM. If it is right I
would like to participate on it. We developed own Tree API for XML (TAX) in
XML module for NetBeans and we could share experiences.
Thank you for your respond.
Best regards,
Libor
--
Libor Kramolis | libor.kramolis@sun.com
NetBeans/Sun Microsystems, XML module | http://xml.netbeans.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Andy Clark <an...@apache.org>.
Libor Kramolis wrote:
> Applications which use XML as *data* storage, they do not need (and
> sometimes do not want) to know, what was structure of data source. They only
We have the DOM Views module for presenting different views
on the entire document structure. Do you need more than this
functionality?
--
Andy Clark * IBM, TRL - Japan * andyc@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Libor Kramolis wrote:
> So when you want create tree structure for *data* oriented applications, you
> can have just three nodes - element, attribute, text. Nothing more. While
> *structure* oriented application need tree structure about eighteen kinds of
> nodes. And *TAX* is structure oriented model of XML and DTD documents,
> which we use in NetBeans XML module for two way editing.
>
> I hope, it is comprehensible. I am looking forward to any questions,
> suggestions, comments. Thanks.
hi,
this is very nice distinction - and it is very relevant to SOAP as from this point of
view one could say that SOAP processors need data and not structure oriented APIs for
accessing XML? Otherwise one would be building data oriented API on top of structure
oriented.
thanks,
alek
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Libor Kramolis <li...@sun.com>.
Aleksander Slominski wrote:
> Libor Kramolis wrote:
>
>
>>Yes. There is a lot of tree APIs around XML and it is hard to choose one of
>>them. W3C DOM is the most used because it is standard API but it is not so
>>much comfortable for java developers and this is reason, why there are JDOM,
>>dom4j, etc. Those APIs are still *data* oriented only and this is reason,
>>why we are developing *TAX*.
>>
>
>>TAX provides *structural* oriented API, because NetBeans XML module provides
>>tree editing support and it needs more info about parsed XML and DTD
>>documents. And this is reason why we use XNI instead of SAX for build TAX tree.
>>
>
> hi,
>
> could you exlain what do you mean by sata and structural oriented (and maybe give
> an example)?
XML document has some *structure*, which is described by w3c recomendation
(tags, attributes, references, etc.). This structure describes some *data*.
And this is that difference between those views at XML documents.
Applications which use XML as *data* storage, they do not need (and
sometimes do not want) to know, what was structure of data source. They only
want to get tree structure of elements, which can have attributes and some
content. Attribute is pair of name and value; content is list of another
elements or/and literal data. That's it. No DTD, no schema, no references,
no comments are important for manipulation with data.
While tool, which supports tree editing of XML and DTD documents, needs
*structure* of ALL nodes, which could be in XML and DTD files used.
Including header, DTD declarations, used entity and character references,
comments, etc. Important is also order of element's attributes and
information about attributes specified in DTD.
So when you want create tree structure for *data* oriented applications, you
can have just three nodes - element, attribute, text. Nothing more. While
*structure* oriented application need tree structure about eighteen kinds of
nodes. And *TAX* is structure oriented model of XML and DTD documents,
which we use in NetBeans XML module for two way editing.
I hope, it is comprehensible. I am looking forward to any questions,
suggestions, comments. Thanks.
>
> http://xml.netbeans.org/tax/ does not exaplain anything...
You are right. We plan to improve its pages especially with UML diagrams. I
will announce about it.
>
> thanks,
>
> alek
>
>
Thanks for your question.
Regards,
Libor
--
Libor Kramolis, Software Engineer | libor.kramolis@sun.com
NetBeans/Sun Microsystems, XML module | http://xml.netbeans.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Aleksander Slominski <as...@cs.indiana.edu>.
Libor Kramolis wrote:
> Yes. There is a lot of tree APIs around XML and it is hard to choose one of
> them. W3C DOM is the most used because it is standard API but it is not so
> much comfortable for java developers and this is reason, why there are JDOM,
> dom4j, etc. Those APIs are still *data* oriented only and this is reason,
> why we are developing *TAX*.
> TAX provides *structural* oriented API, because NetBeans XML module provides
> tree editing support and it needs more info about parsed XML and DTD
> documents. And this is reason why we use XNI instead of SAX for build TAX tree.
hi,
could you exlain what do you mean by sata and structural oriented (and maybe give
an example)?
http://xml.netbeans.org/tax/ does not exaplain anything...
thanks,
alek
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Libor Kramolis <li...@sun.com>.
Andy Clark wrote:
> Libor Kramolis wrote:
>
>>I would like to ask you, what did you mean by that. Do you have any proposal
>>or plan about it? What is it? What it will be? What I can imagine about NTM
>>
>
> I was just highlighting the fact that XNI may be extended in
> the future by adding "extension" packages for doing various
> things related to or optimized for the XNI framework. As an
> example, I used a "native" tree API. However, I don't have
> any ideas or plans at the moment for inventing such a thing;
> there's plenty of work remaining for the XNI Core and Parser
> Configuration framework.
>
>
>>name is something like java oriented and extended DOM. If it is right I
>>would like to participate on it. We developed own Tree API for XML (TAX) in
>>XML module for NetBeans and we could share experiences.
>>
>
> There are a lot of various tree APIs around. The most common
> one being the W3C DOM, of course. But other examples would
> be DOM4J which is an alternative, Java-only tree API; and
> even Xalan has their own internal tree model which they
> call DTM. And there could still be others.
>
Yes. There is a lot of tree APIs around XML and it is hard to choose one of
them. W3C DOM is the most used because it is standard API but it is not so
much comfortable for java developers and this is reason, why there are JDOM,
dom4j, etc. Those APIs are still *data* oriented only and this is reason,
why we are developing *TAX*.
TAX provides *structural* oriented API, because NetBeans XML module provides
tree editing support and it needs more info about parsed XML and DTD
documents. And this is reason why we use XNI instead of SAX for build TAX tree.
> Of course, if we were to work on a general tree model as
> an XNI extension, then the goals I would have are: to be
> complete in terms of the document infoset; have high-
> performance in creation, access, and memory-usage; and
> be easily extendable. Not an easy task, that's for sure
> which is probably why there are so many different tree
> models out there.
You are writing about document infoset. Do you mean by this W3C XML
Information Set (http://www.w3.org/TR/xml-infoset/)? This could be great but
XML InfoSet does not describe all parts of XML spec, see
http://www.w3.org/TR/xml-infoset/#omitted.
I can imagine that relation between NTM and DOM will be same as relation
between XNI and SAX. I hope, I am not too wrong. I think that XNI is event
oriented API, full describing XML and DTD documents. And I also think, that
NTM could be object model API for full describing those documents. And TAX
does or would like to do something like this.
And this is reasons why I would like to participate. We do not want to
create another tree API, which will be only one of twenty similar APIs. We
would like to participate on design java oriented tree API, interfaces - not
implementation (as JDOM). What about some JSR?
>
> Do you have a link to this Tree API for XML you were
> referring to?
>
TAX is part of NetBeans XML module. It is created as NetBeans independent
library. This is short time when it was open sourced, so web pages are not
too detailed, coming soon. You can look at http://xml.netbeans.org/tax/ and
source code in CVS http://xml.netbeans.org/source/browse/xml/tax/.
Thanks for your feedback, I appreciate it.
Best regards,
Libor
--
Libor Kramolis | libor.kramolis@sun.com
NetBeans/Sun Microsystems, XML module | http://xml.netbeans.org/
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Re: [XNI] Native tree model
Posted by Andy Clark <an...@apache.org>.
Libor Kramolis wrote:
> I would like to ask you, what did you mean by that. Do you have any proposal
> or plan about it? What is it? What it will be? What I can imagine about NTM
I was just highlighting the fact that XNI may be extended in
the future by adding "extension" packages for doing various
things related to or optimized for the XNI framework. As an
example, I used a "native" tree API. However, I don't have
any ideas or plans at the moment for inventing such a thing;
there's plenty of work remaining for the XNI Core and Parser
Configuration framework.
> name is something like java oriented and extended DOM. If it is right I
> would like to participate on it. We developed own Tree API for XML (TAX) in
> XML module for NetBeans and we could share experiences.
There are a lot of various tree APIs around. The most common
one being the W3C DOM, of course. But other examples would
be DOM4J which is an alternative, Java-only tree API; and
even Xalan has their own internal tree model which they
call DTM. And there could still be others.
Of course, if we were to work on a general tree model as
an XNI extension, then the goals I would have are: to be
complete in terms of the document infoset; have high-
performance in creation, access, and memory-usage; and
be easily extendable. Not an easy task, that's for sure
which is probably why there are so many different tree
models out there.
Do you have a link to this Tree API for XML you were
referring to?
--
Andy Clark * IBM, TRL - Japan * andyc@apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org
Can't get off mailing list
Posted by "Michael N. Christoff" <mc...@eldan.com>.
How do I get off this maling list? I've sent about 40 messages to
xerces-j-dev-unsubscribe@xml.apache.org and it keeps on sending me emails!
Mike Christoff
---------------------------------------------------------------------
To unsubscribe, e-mail: xerces-j-dev-unsubscribe@xml.apache.org
For additional commands, e-mail: xerces-j-dev-help@xml.apache.org