jon * wrote:
> 
> on 4/2/00 9:49 AM, Kevin A. Burton <[EMAIL PROTECTED]> wrote:
> 
> > I never said HTML was dead.  Just that webmacro was dead.  IMO it is
> > inferior to XML publishing by a long shot.  I might as well starting
> > using JSP.
> 
> No. WM is a much better solution than JSP and in some ways the whole XSLT
> stuff because it is so limiting!!!

IMO my goal is to have all XML data everywhere.  Ideally it would be
nice to see things like OPL or Castor JDO just map things back to the
database as an XML persistence layer.  The data would still be
represented in UTF-8 but its structure would be high performance.

When you use something like XSLT you gain much more because all the data
is structured.  JSP/WM don't provide document structure and don't
natively have support for XSLT and only can through a kludge IMO.

... starting to sound like a broken record. :(
 
> I like the fact that the designers can't write stupid functions in some
> funny language...WM is simply a very very very very basic MVC model.
> 
> With XSLT or even JSP, you can embed code within your pages. That is
> horrible in my mind. You shouldn't even allow that stuff to happen because
> it *will* get abused.
> 
> I talked with Stefano about this at ApacheCon for an hour...He feels that it
> is ok that this stuff is in there because anyone smart won't abuse it. The
> same thing could be said for JSP. JSP can be used in a manner in which it is
> a good technology. The fact that it can be easily abused it the downside
> imho. So, take the easily abused parts out! They just are not needed!

I agree.  These things should be removed.  IE like multiple inheritance
in Java or pointers.  Don't let people do it.  

The problem is the blend over.  Where do you draw the line.  Do you let
people make a bean call and that is it?  I would like to see some of
these coding mechanisms removed from XSP for this reason.  Taglibs are
the solution IMO.
 
> > - it isn't XML
> 
> big deal! not everything HAS to be XML!

Yes.  It does.  The only thing that shouldn't be is raw Java.  But even
then I think that portions of a .java file should be marked up with RDF
for extended Javadoc support... flame on that one. :)
 
> > - no validation
> 
> tell me the truth, how often do you need validation like that (REAL WORLD!)?

Often.  Seriously about 50 percent of the time.  Say you have a headless
client (like a BHO under IE) and it is fetching XML.  You want to
validate the XML because really anything can happen in transit.  You
could screw up an fetch the wrong URL, etc.
 
> > - no XSLT without a kludge
> 
> XSLT is a kludge in some ways because it allows you to embed java code (or
> whatever language) to make up for anything lacking in it.

No it doesn't.  The only things you can do is have for and if.  But they
are rarely used.  Most of the time you are just doing pattern matching
and XPath queries.  

It is really intended to be a thin layer (sheet) on top of your logic.  
 
> > - still encourages logic within your page (this won't go away for a
> > while though).
> 
> that is bad imho.

I agree.  But it is a slippery slope.  I only think taglibs should be
allowed.  You need some way to dynamically produce a document.  However
some people still haven't joined the new school :)

Kevin

-- 
Kevin A Burton ([EMAIL PROTECTED])
http://relativity.yi.org
Message to SUN:  "Please Open Source Java!"
"For evil to win is for good men to do nothing."


------------------------------------------------------------
To subscribe:        [EMAIL PROTECTED]
To unsubscribe:      [EMAIL PROTECTED]
Problems?:           [EMAIL PROTECTED]

Reply via email to