Internet Alchemy

Microsoft has released a new version of MSXML for public review. It’s the first version to support the XSLT recommendation. The change documentation lists the XSL features not supported in this new release:
xsl:transform

• xsl:import
• default-space, xsl:strip-space, xsl:preserve-space
• xsl:apply-imports
• named templates, xsl:call-template
• xsl:namespace-alias
• named attribute sets
• xsl:number
• case-order attribute
• lang attribute

and in the XPath implementation:
following

• following-sibling
• preceding
• preceding-sibling
• namespace
• lang
• sum
• floor
• ceiling
• round

Well, at least the XSLT engine appears to be usable and they even include a stylesheet to transform the old IE5 XSL style sheets into the shiny new XSLT ones.

XHTML has finally made it as a recommendation, paving the way for modular XHTML to build a future for mobile and TV based browsing. For the uninitiated, these are the differences between HTML and XHTML:

Documents must be well-formed

No overlapping tags and all tags must close properly

Element and attribute names must be in lower case

This is because XML is case-sensitive, and someone had to make a decision one way or another

For non-empty elements, end tags are required

No more leaving off </td> or </p>

Attribute values must always be quoted

This was always good practice, I guess only the lazy web designer leaves off the quotes these days (unless you’re trying to save 90 bytes per page)

Attribute Minimization

This is quite a big change – it means that things like ‘checked’ or ‘nowrap’ are now invalid

Empty Elements

You need to terminate empty elements with a / like this: <br />. This applies to all empty elements such as IMG, META, HR or BR

Whitespace handling in attribute values

All multiple spaces in attributes now reduce to a single space

Script and Style elements

Ugly, ugly , ugly change. Because of the way they’ve declared <script> and <style> you have to escape your code in <![CDATA[ . . . ]]> constructs.

SGML exclusions

This just means that they can’t use a DTD to describe some of the restrictions they are placing on XHTML such as no nesting <a> elements.

The elements with ‘id’ and ‘name’ attributes

You can’t use a ‘name’ attribute as an identifier any more, you should be using ‘id’. This affects named anchors and Netscape JavaScript that does image rollovers

Of course no current browser actually understands XHTML at the moment, but this is the way things are going so start making some of those changes now to save work in 12 months time. Read the HTML Compatibility Guidelines as well to help keep your changes compatible with existing browsers. Good luck.

If you need to read up on techniques for traversing tree structures in SQL, Joe Celko is your man (as ever). Here are his three articles about constructing, traversing and manipulating hierarchical trees in SQL.

WapWarp has responded to the Geoworks patent announcement with an open letter.

What a shame it would be if developing (content and software) for mobile phones were to become the preserve solely of large corporations. What a shame if innovation were to be quashed. What a shame if WAP were to fail miserably as a format because one company chose to raise the barrier to entry above the means of most innovators.

YourWAP have released a beautifully crafted WAP emulator. You can emulate the three main WAP phones including the Nokia 7110 and unlike the SDKs from Ericsson and Nokia it’s all self contained and very quick.

At last common sense has prevailed and Netscape are now allowed to distribute an international version of Communicator with 128-bit encryption built in. Of course I’ve been using the patch from the good people at Fortify up until now. I guess their little niche has vanished now, but they did a very good job filling the void up until now. Thanks for all the hard work.

I found this Durlacher report today. It’s a good analysis of why building community is important. In fact they go further and argue that ‘…communities are not simply a nice to have adjunct for a web-based business, but that their creation is central to a sustainable business model‘.

Spotted on WAP Today was this piece of news about a patent claim by Geoworks on aspects of WAP and WML. Geoworks have a page devoted to the claim and have written a detailed white paper. The patent was issued in 1994 and according to the abstract is:

A method for invoking a user interface for use with an application operating in a computer system which involves providing in the computer system a generic object class that corresponds to a class of function that is to be performed using the user interface; specifying in the application instance data in the form of a generic object specification that corresponds to the generic object class, the instance data including attribute criteria and hint criteria; providing in the computer system at least one specific user interface toolbox and controller that operates in the computer system to provide a selection of possible specific user interface implementations for use in performing the class of function; and providing in the computer system at least one interpreter that corresponds to the at least one specific user interface toolbox and controller.

Now what does that all mean. At least we have the white paper which does go some way to clarifying it:

The Geoworks Patent is based on modern programming techniques where it is better to write application programs to a
generic user interface specification rather than to a specific user interface specification. Before the invention of the
Geoworks Flex UI, programmers would either implement an application’s user interface with a specific interface toolbox,
thereby binding their application to that system and its limitations, or they would implement the interface in terms of a
least-common-denominator abstract interface that didn’t allow their application to take advantage of any of the strengths
of the systems on which the application would run. The Geoworks Flex UI took a different approach.




At the core of the Geoworks Patent is a unique abstraction of a program’s user interface. An application’s interface is
defined in terms of generic user interface elements. Each element has a set of attributes that define its basic
characteristics, as well as a set of hints, which give higher-level semantic information that tells the system the
designer’s intentions.




When the application is run, an interpreter uses the attributes, hints, and the location of the generic user interface
element to determine how the element can best be represented, given the features of the specific user interface. The
attributes specify requirements that must be met, while the hints suggest characteristics the specific implementation
may have, but they are not requirements; the specific user interface need only make a “best effort” to adhere to them.
The generic elements are passed to a controller in the specific user interface to realize the generic element, using
elements from the specific user interface toolbox, under the instruction of the specific user interface interpreter.

Hmmm. Sounds suspiciously like a web browser to me. Read the full paper for more of the same. They also handily provide a table listing who is likely to infringe this patent:
WAP-Enabled Wireless Telecom Services, WAP-Enabled Mobile Devices, WAP Microbrowsers, WAP-Enabled Servers, WAP Application Development Tools, Packaged WAP Applications, Custom WAP applications and Consulting, WAP ASPs, WAP sites. I think that just about covers the whole industry.




The licence fee for the patent is $20,000 per annum, but generously they state in a footnote: Companies with annual gross revenue of less than $1 million may be eligible for a fee waiver.. So that’s OK then, just 2% of revenue to host a WAP site.



There is one saving grace though: there are only two patents – US and Japanese. Europe, WAP powerhouse that it is, isn’t covered. Of course this can only be good news for us Europeans: while the US and Japanese markets dry up under the terms of this patent, us GSM types will be forging ahead making the best use of our WAP phones.

ZopeStudio is the start of something new. It’s an interface for managing Zope with a new twist: it’s built in XUL, the interface definition language designed for Mozilla. This is the birth of the true universal internet application. We all knew it was going to be the web browser, but we didn’t a web browser that was simply an application shell with bolt on functionality.

A report published by Inktomi has estimated that there are now 1 billion web pages out there. This is why search engines have to change their indexing strategy. More context is needed for each search. Where am I when I search? What have I seen before? What have other people done when performing this search?