Re: [xml-dev] XML, Rich Internet Apps

[AndrewWatt2000@aol.com]

In a message dated 24/11/2002 17:27:43 GMT Standard Time, chuck@tumeric.net writes:


> Thanks. It's all just a big accident. There's a lot of creativity in
> programming, especially in XSLT. I'd love to see more graphic designer
> types make the leap, but numbers scare most of us.


:) ... Yep.

It's not very usual to find people who are comfortable both with code and with thinking visually. I think that's one reason why many designers have found SVG a little large to bite in one chunk. There is almost an ongoing tension in the brain to think visually but express it in code. But it's well worth persisting with SVG.



> Yes, the systems are there, but they need to be cobbled together. I can
> do that, and anyone on this list can, I presume, but most graphic
> designers I know wouldn't have a clue where to begin. And then there's
> the whole design pattern question. Do we even want designers (I am one,
> so I am not bashing these people here) to be involved in data modeling?
> Probably not. So here is my fantasy all purpose publishing IDE:
>
> * It can read XML data sources
> * It has a drag and drop interface with a property and methods
> management panel similar to what you see in development IDEs that lets
> you place an object (node) onto a panel and position it as you wish.
> * It has a timeline
> * It outputs it to at least SVG, but my real fantasy is that it outputs
> to XSL-FO and HTML and has a hook into something like FOP. And if we
> really want to get silly, maybe it even outputs to SMIL. 



Something that outputs SVG and XSL-FO would be great. As you mention later Adobe has a document server and a graphics server at the end of beta so I think we will see one possible source of high quality SVG/XSL-FO/PDF output very soon from Adobe.

You mention HTML as part of the desired output. Why?

In the open source sector Cocoon can provide the XSLT/SVG/XSL-FO/PDF trip - the capital outlay is nil but there is a lot of developer time to get fully up to speed with Cocoon. But it is being continually developed so I think it is one to watch.

There seems to a be a price opportunity for someone here. Surely there is a space between $0 and $20,000 (or is it $40,000?) for someone to offer something else in this space?

> I know I'm wishing for the moon here. One problem is that I haven't seen
> any kind of visual-based XSLT tool I would be willing to trust, and the
> above would require one built in to the process. I've looked at a few,
> and have quickly gone back to coding, but that may stem from my internal
> bias against visual editors that began somewhere along the time I tried
> out my first WYSIWYG HTML editor. But it probably has more to do with
> that I just find so much creative challenge in building XSLT docs by
> hand.
>
> So basically, I guess my fantasy is sort of an amazing, magical XSLT
> transformation tool that outputs XML from a visual interface into a
> variety of formats -- SVG, HTML, SMIL, PDF (through XSL-FO). I have to
> assume that is either an amazingly difficult piece of software to write,
> or that nobody has seen an appropriate ROI value in it.


Or that you have slightly underestimated the time to get it to market? :)

> Anyway, I would at least like to see an SVG tool that is like that. 


Agreed.

Sort
> of a data-driven WebDraw. 


Sounds highly desirable.

And with avenue.Quark and Adobe's InDesign
> 2.0's XML plug-in we're close to repurposing page layout, but I haven't
> figured out how to keep designers out of the process using those tools
> (this goes back to my "we really don't want graphic designers modeling
> XML, do we?"). 


In a sense designers must be *in* the process. I see the question as being one of how to achieve good design along with good coding (by hand or automatic) without each getting in the other's way.


> > I don't see why not! (heh) - the majority of applications are dead
> easy if
> > you know XSLT and/or DOM.
>
> Okay, you're right -- I can write stuff to make all of this work, and I
> have. What I meant was that I am not capable of writing an executable to
> make it all happen at once. Although, if I had Adobe's money (or one day
> of Bill Gates' earnings), I bet I could put the team together that
> could.


Give it time! :)



> I've done some data-driven SVG myself using XSLT. I wish I had a little
> more time to play with that. 


Absolutely. If only we had 48 hours days! :)

> So yes, I agree that all the systems really are in place for everything
> I am talking about. What I am dreaming of is binary executables that are
> within the grasp of the people who could really make the stuff we talk
> about on these lists get used every day. Making things like SVG, XSL-FO,
> and even XSLT more accessible to non-programmers would be a very good
> thing. 


Yep.

  I think that the data-driven part of this equation is very
> important. 


Yep.

I don't see any particular benefits to simply cobbling
> together SVG for the sake of SVG because it happens to be XML. 


The huge advantage of SVG over generic XML is that human beings (for at least certain things) think visually.

"An SVG picture is worth 1,000 XML words!". (c)(tm)

I might
> as well use Flash. In other words, I don't want arbitrary content in my
> SVG files. 


Surely this is a question of good design.

Part of the problem with Flash content has been bad / gratuitous design. Ironically gratuitous Flash design has attracted attention and attracted awards. A user who is beyond the shiny bauble level wants graphics which communicate effectively. SVG should be a tool in achieving effective communication. Not poking the visitor to a site / user of an application in the eye with gratuitous animation.

I would prefer that writers create and update their content
> in an XML document and bring it in to whatever design program is being
> used.
>
> I've read about Adobe's Document Server, and maybe that does all or some
> of what I'm talking about. But I haven't seen it, and it's 20,000 US
> dollars. Not a small investment for people to make. And that doesn't
> include the cost of hiring an IT person or two to make it work, or the
> training it would require to get staffers to use it properly. It sounds
> more like a product for Lockheed than an ad agency (my typical
> hang-out). 


"The rest of us" have to look to someone seeing a gap in the market between the $0 of Apache Cocoon (which is *not* (yet) safe to let a "pure" graphic designer near unsupervised) and the $20/40,000 of Adobe's Document Server and/or  Graphics Server.

I am optimistic we will see several such tools in the not too distant future. Whether the first generation of such tools are good enough or close enough to your ideal tool remains to be seen.

Andrew Watt