Spotted in At Least You're not Totally Engulfed in Flames. This guy is playing with the XML DOM tree in VW (which is exceptional btw!) and he's trying to make an RSS feed by building up the XML document.

James Robertson has modelled the RSS objects so that he can generate an RSS feed from regular objects, not XML.

And I generate RSS using SimpleWeb which has an API like the following:

self rss: [:rss |
	rss description: 'This is my description'.
	rss item: [:item |
		item title: 'This is my item'.
		item description: 'This is the description of the item']]. 
 

So on and so forth. Both techniques work well for reducing noise in the code and allowing you to abstract things further than just an XML DOM tree.

Today Software WithStyle released a beta version of XML WithStyle. XML WithStyle is a WYSIWYG XML content editor with the 'look' of the published page and the 'feel' of a word processor. For more information on XML WithStyle, please see xmlwithstyle.com.

We are keen for you to put XML WithStyle through its paces and report bugs to us as well as any comments you may have on its features and usability. To this end, we have established the XML WithStyle Beta Program.

The beta program provides you with an early preview of the XML WithStyle product prior to the commercial release of version 1.0. It allows you to:

1. conduct beta testing of the software to assist us in identifying defects,
2. provide constructive feedback about the product to help us better serve your needs, and
3. conduct a preliminary evaluation of the product prior to the 1.0 release.

If you're not a technical person and don't know anything about XML, don't worry! XML WithStyle is actually designed for you! It is targeted at non-technical business users. To make things straight-forward, it offers a familiar word processor-style user interface with a rich, WYSIWYG workspace. This makes editing just like typing straight into a published web page or printed document (subject to the look and feel used).

XML WithStyle does differ from the word processor and HTML editors you might be familiar with because it is a structured editor. This means that it ensures your content is structured according to pre-configured rules - according to the XML vocabulary you are using. This means that you can focus on the structure of your content and what it means rather than what it looks like. This can be a big paradigm shift if you are used to being creative with fonts and colors but working this way has big payoffs for organizations that want to maximize the value of the content their staff author.

For the more technically minded, XML WithStyle showcases the breadth of the WithStyle UI technology including deep structured editing capabilities as well as CSS-base rendering. Aside from several commodity components such as HTML Tidy and ASpell, all of XML WithStyle is written in VisualWorks Smalltalk.

Largely because of the additional work to port the APIs to non-Smalltalk components across to other platforms, the XML WithStyle beta is available for Windows operating systems only at this stage. We intend to make the product available for additional platforms in the future as this is one of the big strengths of VisualWorks Smalltalk. In fact, I'm using the XML editing functionality of XML WithStyle right now - in the guise of Bottom Line - to type these words on Mac OS 10.4. See our case study on Bottom Line to learn about how that application benefits from the same XML editing technology as XML WithStyle.

Although we have a good deal of work to do before the commercial 1.0 release, we feel that XML WithStyle provides a rich set of functionality at this point in a market space that continues to lack genuinely usable and cost-effective products for non-technical users.

We hope that you take the time to download the XML WithStyle beta and send us your comments. This will help us to make the commercial release even better and tailored to your needs.

Download now and enjoy using our latest creation!

-- From Rowan Bunning, Product Manager of XML WithStyle.

The discussions on whether you should be allowed to parse non-conforment XML has been raging for a few weeks now. I generally like to keep out of this war. But I think I should put my 2c in because of my involvement in the XML technology.

First of all, I do not agree with the advocates of parsing XML data that's non-conforment simply because they can. They should do respect the goals of XML. But - if your business requirement is to interact with the web at-large and you need fault tolerence with that, then handling non-comforment XML is something you should do.

In WithStyle, we prefer real valid XML - in fact, we penalise the CPU by first trying to parse a document as XML, then failing that, we use LibTidy to try and convert it in to valid XML then parse it, failing that, we then use LibTidy to try and convert it in to HTML, then in to XHTML, then parse it as a valid XML document. Failing that, we have one last resort that's out of date - but if we get here, you can bet your bottom dollar we're not going to parse it.

Now I guess I should point out that this guy has decided to stop reading James's blog because James is an advocate of parsing dodgy XML. In actual fact, James uses the same set of rules that WithStyle uses - he doesn't really pay much attention to this because he's too busy with his business requirement of handling dodgy data. But in actual fact, his code prefers valid XML.

Okay, now on to the XML specification. The spec is quite specific about saying that normal XML processing should cease if a fatal error is encountered. But, it does not stop you trying to do other kinds of processing. To that end, I submit that the approach WithStyle takes is valid:

• Can't parse it as XML, try a non-normal parsing approach to turn it in to valid XML
• Can't parse it as XML, try a non-normal parsing approach to turn it in to HTML, then XHTML
• Still can't parse it as XML, give up

Note that we do not try to correct errors as we parse the XML. We give up, like the spec says we should. Instead, we then try and adjust the data until it becomes something that is valid XML.

So at the end of the day, WithStyle will only parse valid XML - but it has a few tricks up its sleeve to convert dodgy data in to XML.

One thing I should mention is that all WithStyle technology always puts out valid XML. That's one thing we've been very strict on. We don't intend to add to the plethora of malformed XML content on the internet.

So, while James may like to get on his high horse about dealing with invalid XML, I try to avoid that argument entirely. I know that my code underneath isn't breaking the rule - and neither is James's. So, perhaps that guy Sjoerd Visscher might like to subscribe to James's blog again.. who knows. Hopefully he's subscribed to mine.

Spotted in Smalltalk Tidbits, Industry Rants .

James seems to think this is witty. What disturbs me is that Sriram is anal enough to think that people should be writing HTML or XML in the first place. Bleh, what a dumb idea. Lots of little tags.. that's insane.

Most sane people just don't do that sort of thing any more. They make their webpages using some WYSIWYG program, like Macromedia's Contribute or heck, even Frontpage. As much as I don't like this programs, I'd recommend them over somebody sitting down in notepad and writing out tags.

So I have to wonder when Sriram talks about "Missing one tag meant that hundreds of feeds couldn't be imported." .. is he saying that some moron sat there and write out hundreds of tags by hand? Christ! What drugs was that guy on?

XML is not for people, it's for computers. XML IS NOT HUMAN READABLE

Spotted in Raw . Here's some heavy ready when it comes to XML, XHTML and web API's that are alternate to WebServices.

Spotted in fullasagoog.com full roast blend . Here's a discussion about "First Class XML" .. This is something that is not a problem in Smalltalk. All objects are first class. In fact, we can specialise the behaviour of individual elements in an XML document for our program.

Spotted in Web Things, by Mark Baker

Just recently W3 has approached WS-Addressing. Mark Baker is on the ball, pointing out that this is a waste of a spec, where URI's can do just as good a job. It appears as though they've taken what a URI actually is (ie: its object model) and defined that as XML. Then thrown in namespaces to extend what a URI is capable of containing in terms of information.

What is amazing is that they managed to get this URI modelling wrong! They don't have tags for the components of the address, only the parameters.

Tell me.. why is it that CSS isn't defined in XML? Or XPath isn't defined in XML? It's very simple really. It's impracticable to have every descriptive language defined in XML. So why would we want to write URI's in XML? Well, honestly, we wouldn't - URI's work, are standard and are well supported all around.

Spotted in fullasagoog.com full roast blend . It looks as though the Australian Department of Immigration is starting to take web standards more seriously. Good news for all.