- From: Lorenzo De Tomasi <lorenzo.detomasi@libero.it>
- Date: Wed, 29 Jan 2003 01:01:19 +0100
- To: <www-html@w3.org>
The big problem seems that no one has clear the goal and tasks of xhtml. Tell me if what I write now can be a good proposal. 1. It seems that the goal of the xhtml text module (but not only of it) coincide with the goal of DocBook. So probably using the same rules of DocBook for the xhtml body module would make happy content editors. In my opinion DocBook needs to be simplified and made more flexible to permit everyone to learn it easily. If we create a module that groups the modules that we can use both for xhtml and pdf/printable documents we obtain something that is similar to Docbook. I call it document 'body module' and I structure it in this way: <body module> <text module> <list module> <table module> <object module> <hypertext module> <note module> </body module> The <note module> is an hybrid between hypertext and list (I'll talk about it in another mail). So xhtml documentation should be structured in this way: <xhtml> <structure module> <metainformation module> <style sheet module> <linking module> <scripting module> <body module> <text module> <list module> <table module> <object module> <hypertext module> <note module> </body module> </xhtml> All these modules are rendered by the browser with a presentational style (that may not be the most important thing for some people). You can notice that these modules are the ones that most interest 'Academic Users' as defined in Veith Risak <veith.risak@chello.at> e-mail titled 'xhtml2 and user groups': <cite> "Academic users" are interested in clearly structured documents, with high temporal stability (think of reviewed documents of highest quality of content, which can be cited for a long time, ... They are very near to the original ideas (Tim Berners Lee) of html. XHTML is useful for them, because it requires valid source code instead of "tag-soup". Presentation details matters much less, but a connection/linking with MathML would be useful. Further useful components are: - footnotes - internationally agreed standards for quality attributes (from "proposal" up to "frozen document" and from "unmature idea" to "proved theorem", ...)" </cite> I think that the actual xhtml body module is walking in the right way using more general tags than DocBook, like nested <section>s, <h>s, <ol>/<ul>+<li> in place of <chapter>, <title>&<subtitle> and <OrderedList>/<ItemizedList>/<VariableList> +<ListItem>. Body module can be considered a basic module, but professional content editor need also an expert version with few little new tags, as suggested by Veith Risak; in this sense I think that DocBook and xhtml must converge more, so that xhtml body module becomes the basic module and DocBook the expert expanded module - using the same tags and not different as now. The final product would be powerful as or -better - more than the actual DocBook. The goal is to create&structure contents only one time in an xml doc and then create an xslt doc to generate an xhtml doc or a pdf doc with the same contents imported in both; using the same tag structure for both would simplify the creation of the xslt doc. For example if I create an xml file like this <content> <section> <h>title</h> <p>text</p> </section> </content> Advantages are - that I have only to import the <content> content without modifications; - that I can easily link a standard css file to obtain a good layout for my contents; - that if I write an xhtml doc I can create an xml doc equivalent to DocBook only with few cut&paste operations; - that I must learn only one language and no different tag names with the same function; - that I can also ask to a client to write contents in an xhtml body module doc and directly use it both to create an xhtml and pdf version using different xslt docs. With the current specifications of xhtml2, if I have to write a document, the best solution is that I learn DocBook and I write an xml doc using DocBook rules to structure contents; then - or I write a Css doc to give it a layout for the web (I have to learn xml + css); - or I write an Xslt doc to transform it in xhtml (I have to learn xml + xslt + xhtml); - or I write an Xslt doc to give it a layout for printable Pdf (I have to learn xml + xslt + pdf language); 2. Veith Risak defines also "Commercial users" <cite> "Commercial users" are much more interested in presentation aspects, the look and feel for readers, interaction aspects, security... Most of them are concerned with styles, css, ... as you can see from the discussion. The contents must be updated very often (e.g. stock-exchange applications), temporal stability is not so important." </cite> They are interested in learning css, javascript, etc. They need and want to learn the other modules: <structure module> <metainformation module> <style sheet module> <linking module> <scripting module> They want each site is rendered identical by every browser, without problems of visualization and accessibility. A big problem in creating websites is that they are rendered differently by each browser and webdesigners have to create different versions for each browser or to consider all the variables. A different rule may destry a webpage. Every webdesigner hates this! The problem would be solved if W3C would give 'precise to pixel' directives in how rendering each tag+css command and in how to set default browser settings. So precise to pixel definitions of css rules for each tag of default browser settings can reduce problems. Directives can be only suggested and not obliged, but I think that it's very important to write/show them. I think that every browser manufacturer would adopt W3C suggestions as they exactly do with xhtml. Now the defaults are imposed by the most diffused browser i.e. Microsoft Internet Explorer. I think that a suggested standard by W3C - not based on monopoly - is more democratic. I think also that the promotion of a well designed PUBLIC DOMAIN font family covering all characters of Unicode and serif and sans-serif would be very useful. If good, also if not imposed, it would be probably adopted as substitute of Times New Roman (not well legible on a monitor, but good for printed docs) and Verdana (not well legible on printed docs, but good for monitors) by most browsers. I think that style sheets should consider the possibilty of adding/removing antialiasing from text, because in general (but it depends to the font choosen) big texts are more legible with antialias and small texts without. If a browser has one of the 2 options as default for every text it's not good. Designer should have the right of choice to apply or not antialiasing to a text style. Tell me if you share my vision or criticize it without problems. Thankyou very much. ____________________________________________________________________________ Lorenzo De Tomasi, student of Information Architecture, Interface Design and Visual Design via Bellaria 6, 21018 Sesto Calende (Varese), Italia phone: +39 (0)331 924649 mobile: +39 329 3941065; +39 333 8979304 e-mail: lorenzo.detomasi@libero.it; lorenzo.detomasi@email.it website: http://biografica.tzone.it ICQ uin: 11313132 Yahoo! Instant Messenger id: tummait
Received on Tuesday, 28 January 2003 19:01:26 UTC