OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] URI Design

[ Lists Home | Date Index | Thread Index ]

M. David Peterson wrote:
> Hi Philippe,
> Yes, I remember the announcement for this project.  It looks
> interesting but in looking through things there were a number of tems
> that raised questions and concerns so I thought it best to give it
> time to mature before worrying much about looking into things any
> deeper.
> I will point  pout that your comment:
>>this is exactly what you can do with Active Tags :
> ...can be quickly and easily countered by stating "yeah, and its
> exactly what you can do with XML."  Don't get me wrong.  I'm not
> trying to slam your project nor your efforts.  Obviously XML contains
> the syntax that can then be used to create projects like Active Tags
> to do whatever it is you want to do with an XML-based syntax but at a
> higher, and specifically defined element level.  You have done a
> really nice job of providing a nice high lever abstract which was
> compelling enough for me to want to at least scan through the pages a
> bit more to see what it was that made this project stand out.  I think
> that with any type of declaritive style XML-based language there is a
> definitely need to showcase the fact that with this defined set of XML
> elements and attributes the benefits of using this specification
> instead of raw XML and an XSLT/XQuery processor have got to be
> painfully obvious.

Active Tags doesn't intend to replace XSLT which is much more efficient 
to process an XML tree and reorganize it at will ; Active Tags is 
usefull to drive complete publishing processes with several steps 
including XSLT.

Active Tags is somewhat close to XQuery, but relies on an XML syntax, 
and as you suggest, an XML-based syntax can do whatever we want to ; 
thus, there is a smart application in Active Tags called Active Schema
and I don't think that XQuery can be used as a support to design a 
schema technology. Active Tags does.

  e.g. SVG provides the ability to create an
> extremely advanced user interface by using a syntax that has been
> simplified to the point that a minimal amount of information is needed
> to create some pretty spectacular results.  While I recognize the fact
> that within your abstract you state a primary focus to:
> <snip/> ...interoperable with other non-XML technologies 
> I wonder just what it is (beyond RDBMS; the term "Web" that follows
> doesn't make any sense...  are you suggesting that XML, XPath, XSLT,
> XQuery, etc... can't interoperate with the web???

oh no ! I realize that the term "web" is somewhat confusing : I was just 
thinking about "HTTP" ; HTTP is not an XML technology :)

> I got as far as the definition/business case against XSLT/for Active
> Tags and abrubtly stopped right then and there as your statements are
> apparently based somewhat on XSLT 1.0 

So much the better, if people recognize some XSLT in Active Tags, they 
would understand it easily
I hope that they also recognize Ant, JSP, taglibs, Cocoon, etc
There is a little of each in Active Tags

but even with 1.0 it seems
> you've got some confusing points that may or may not be valid.  I
> recognize that XSLT 2.0, XPath 2.0, and XQuery are not currently a
> member of the recommended working draft family at the W3C.  But theres
> absolutely no way you can exclude technologies that are as close to
> recommended as you can get before being called recommended, with the
> added benefit of several implementations of all three of these
> technologies in released commercial and open-source projects that are
> considered production ready releases.  When you add these 3 specs to
> the mix I am seeing very little, if any benefit to using a
> specification that doesnt offer me anything else that I cant do and
> adds to the mix yet another API I have to learn.  I would just rather
> "invent the API as I go" so to speak using raw XML and an XSLT
> 2.0/XPath 2.0/XQuery 1.0 processor as I can write that code in my
> sleep and I know that for the most part its bug free.

well, if you don't find anything usefull in Active Tags, don't force you 
to use it !
Active Tags allows to handle objects that are definitively NOT related 
to XML ; IMHO solutions based on XML serialization is not always a good 
idea, for example, when you're dealing with huge objects. Instead, 
Active Tags provides means to handle such objects like if they were XML, 
and some XML operations becomes applicable on them : crossing them with 
XPath, updating them with Active Update (almost the same as XUpdate)
Tell me if I'm wrong, but I don't think that XSLT 2.0, XPath 2.0 and 
XQuery provide such features. Anyway, I don't pretend to fight against 
W3C, I just try to supply a smart toolkit.

> But to be fair, I did stop at the definition of XSLT and for all I
> know what I have missed out on is nothing less than spectacular.  But
> I can't take the chance that it may not be all that spectacular simply
> because without proper comparison against the 2.0 spec theres no way I
> can evaluate what benefits there may be and as such can't spin the
> cycles necessary to read this until it does contain that proper
> comparison.
> Still, I want to make sure you know I am in no way condemning your
> work or suggesting there is little if any value.  I LOVE people with
> passion and those in whom take it upon themselves to make our lives
> easier and better and more fun to live.  

you're welcome !

I have that same type of
> drive inside of me so to condemn your efforts would be an obvious
> mistake of which I have no desire to make.  With the proper
> comparisons made against the current competing technologies, which at
> this point as far as I can tell are XT2/XP2/XQ, this could be
> something really cool and really worthwhile.  Can you take the time to
> make those comparisons or would that be something better suited for a
> time when the implementation/processing engine is complete and as such
> you are able to focus on these types of details?

Once again, Active Tags differs significantly : I present it like a 
framework for XML Native Programming, which covers different topics and 
applications than { XSLT2, XPath2, XQuery }. For example, Active Tags 
can be the glue for components that are exposed as XML-friendly objects. 
  I thought about such architecture where Active Tags could be an engine 
for applications that uses components.

I think that most things will become clearer for people interested in 
Active Tags when I will release the first alpha version of the 
implementation I'm working on, with tutorials, and "hello world" 
examples (http://disc.inria.fr/perso/philippe.poulard/xml/reflex/) ; in 
few weeks...

> Either way, I look forward to learning more at a point that it is a
> bit more clear just what it is that you have accomplished with the
> development of this technology.
> Best regards,
> <M:D/>

thanks for your *Active* participation, David :)


           (. .)
|   Philippe Poulard    |


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS