Re: RE: Swixml and SAX, Jelly

[email protected]
Newsgroups gmane.comp.embedded.carlsbad-cubes
Message-ID <[email protected]>
Brian,

The one thing I don't see swixml being useful for is a
pluggable architecture like Eclipse or my own. I say
this with not a lot of experience with swixml, so
perhaps I am wrong. But the way Eclipse works, each
plugin provides a piece of the UI within it's
plugin.xml descriptor, but not like how swixml would
expect. Core plugins to eclipse provide a number of
extension points, each defining an xml schema that is
expected of any plugin wishing to extend it. Each
plugin to extend it must provide a validated bit of
xml interwoven within the <extension point/> node
inside the plugin.xml. For example, to add a menu, you
extend the actionSet and provide a menu name, action
class, etc. The benefit to this design is that the
entire UI is built/rendered, but not a bit of plugin
code is created until a menu item or button is clicked
to trigger the plugin's creation. You can literarlly
load hundreds of plugins, building a pretty complex UI
and very little overhead in object creation, other
than the swing elements.

Unless I misunderstood Kate (when she slammed me for
being a beginner :D ) my understanding is that swixml
is not able to build a full ui from various pieces of
xml each parsed at different times, including the
ability at runtime to parse and apply a bit of xml
into a Swing component, add it and update the screen.
In my plugin engine, in Eclipse, and in some others,
dynamically loading, unloading and reloading of
plugins is possible, and a UI needs to be able to
remove and/or add UI componenets at any point within
the application lifecycle.

I wish ya'll would give some examples or reasons why
you think xmlpull is so complicated and would
convolute the architecture. I still find working with
direct String comparisons much easier than working
with Document elements and such.



--- [email protected] wrote:
> Kate,
> 
> I partially agree and partially don't agree to your
> last email.
> 
> 1) I do agree that we are talking about the future
> architecture of 
> Swixml.
> 2) I agree that JDOM is easy to work with and
> SAX/XML Pull could 
> complicate the architecture overall.
> 
> 3) I don't agree with the forum on the need for
> SwixML to "become" a 
> plug-in framework
> 	of it's own.
> 
> 	Why would you ever want to create SwiXML as a
> "plugin framework"?
> 		Swixml is an XML to Java generation tool.  (XUL)
> 		 This is something that should be plugged into a
> "plugin" framework.
> 
> 		In my belief, Swixml, shouldn't ever need to
> become a framework that 
> supports plug-ins.
> 		It should be an XML to Java Parser.
> 
> 		I do believe, Swixml should be integrated into any
> number of existing 
> frameworks, like Eclipse, or something like that.
> 
> 4) Secondly, I would like to have any application
> that I create, 
> utilize a minimum number of XML files.
> 	The first file, would have the splash screen, and
> minimum set of 
> widgets.
> 	I  then will parse the first file.
> 
> 	Second, I will have a main XML file with most of
> widgets for the bulk 
> of the application
> 	that will be used frequently.
> 
> 	Third, I will have a number of files that contain
> specialized widgets, 
> panels, etc, that are
> 	called in-frequently (or called in separate
> threads).
> 
> 	So, I do feel, that my files will get quite large.
> 
> 
> 5) 	The issue, of whether Swixml becomes more than
> just Swing DEFINES 
> how far the architecture
> 	must move to support the goals.
> 
> 	If, Swixml will support other tags, non-swixml and
> non-SWT, such as 
> Jelly, Luxor, XUL, etc.
> 	then the issue of whether XML Pull, SAX or JDOM is
> better will need to 
> be truely addressed.
> 
> 6) I am almost ready to send my version 2.0
> candidate code to everyone.
> 
> 	I have put over 3 months of time and many hours
> into refactoring 
> version 128 of the code.  This refactoring
> 	allows much code reuse to support SWT but stops
> short of introducting 
> namespaces.
> 
> 	As I have truely worked with Swixml, and have
> touched/refactored 
> nearly every portion
> 	of the code, I have found that the current
> architecture is limited.
> 
> 	Currently, any Java Bean can be associated with the
> current 
> architecture and used just using the register tag.
> 
> 	But advanced namespace processing, like in Jelly,
> would not work in 
> the current architecture.
> 	That is where a change in the architecture towards,
>  XML Pull or SAX 
> could be beneficial.
> 
> 	The current architecture doesn't know how to
> process specialized, 
> non-Swing, non-bean, classes.
> 		(Yes, I know you could create a bean class to do
> it, but you would 
> have to change
> 		the code if you needed parent/child processing
> like swing).
> 
> 	The current architecture also doesn't know how to
> parse different 
> "hierarchies" of tags, like Jelly.
> 
> ??) Could JDOM work with a more advanced
> architecture, namespaces and 
> different parsers??
> 	Yes.
> 	Could XML Pull or SAX work better?
> 	Yes, but the architecture would need to shift
> pretty substantially to 
> accommodate.
> 
> Again, just my 2.5 cents.
> 
> Brian
> 
> 
> _______________________________________________
> Forum mailing list
> [email protected]
>
http://carlsbadcubes.com/mailman/listinfo/forum_carlsbadcubes.com


__________________________________
Do you Yahoo!?
Find out what made the Top Yahoo! Searches of 2003
http://search.yahoo.com/top2003
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.