- From: Joe English <jenglish@crl.com>
- Date: Mon, 31 Oct 1994 09:05:47 -0800
- To: www-html@www0.cern.ch
The question of how to let users add their own tags to HTML comes up periodically. Wayne Wohler (one of the designers of IBMIDDoc) recently posted a message to comp.text.sgml describing one way to implement this in SGML. With his permission, I'm forwarding it to www-html in hopes that it will generate some new ideas. (IMHO, HTML should start stealing more ideas from well-designed applications like IBMIDDoc and InfoMaster, and fewer from Microsoft RTF.) --jenglish@crl.com --Begin Forwarded Message-- Date: 27 Oct 1994 15:14:22 GMT From: wohler@vnet.ibm.com (Wayne L. Wohler) Organization: IBM Information Development Strategy and Tools Subject: Emphasis markup methodology Summary: Description of a means of marking emphasis by intent Keywords: Emphasis markup Newsgroups: comp.text.sgml Message-ID: <38og4eINNiri@afshub.boulder.ibm.com> The recent discussion of emphasis markup style using FOSIs motivated me to share how IBMIDDoc supports the markup of emphasis. While the following description uses phrases as an example, it would also work for other elements and indeed the markup supports the application of these ideas to all elements. What concerned us was the idea that while we can easily support the markup of material that needed emphasis for one reason or another, in most cases this markup did not capture why material needed to be emphasized. In our older markup system (BookMaster), the abstract nature of emphasis was capture by the idea that emphasis was numbered (hp1-hp9) where you needed to establish a convention for use of different element types. Unfortunately, in practice, people quickly became familiar with the default style effect of each and used it in that way. In IBMIDDoc, Eliot and I (mostly Eliot in this case) came up with a classing scheme that allows the user to indicate intent easily. This is done by defining CLASSES of element. These classes define a class of element, what element(s) the class may apply to, what style may be associated with the element/class combination. The content of the definition explains what the class means so that you have more information about the class than simply the title or name. Here is a class for 'slang'. <ClassDef ClassName='Slang' ElementTypes='Ph' Style='Quotes'> <Title>Slang <Semantic>Identifies all terms which while generally understood, at least in the community reading this document, is not a formally correct or &otq;proper&ctq; usage of the language. </ClassDef> I a document sample I recently marked up I defined the following other classes for phrases as well: AbstractTerm FixedTerm Idiom TechnicalTerm Slang Emphasis ExampleAnnotation Cliche Acronym Trademark Notice that there is one called emphasis still, which is reserved for words I would expect to hear with emphasis and for which no other class applied. To use a class, we defined a common attribute (called CLASS) which refers by name to the class definition: Administrative tasks (sometimes referred to as <ph clas=slang/administrivia/) are often handled through other As you can see, it doesn't require much more effort than marking emphasis in more traditional ways. It does require a knowledge of the different classes that an element may take and a willingness (in the case of emphasis) to figure out why you want to emphasize the material. I have found that that isn't as easy as it may appear. Why bother? One application I think that may be useful, if this approach were used consistantly (a big if, admittedly) is that it would flag certain phrases for human language translators (people, but maybe programs too) that a term or phrase needs special treatment. This in addition to the obvious application of changing the presentation style for different classes of phrase. Comments or suggestions? Wayne L. Wohler Internet: wohler@vnet.ibm.com Dept EBR/025Z IBMMAIL: USIB29WX@IBMMAIL Information Development Strategy and Tools Phone: 1-303-924-5943 IBM Corporation PO Box 1900 Boulder, Colorado 80301-9191 Disclaimer: This posting represents the poster's views, not those of IBM --End Forwarded Message--
Received on Monday, 31 October 1994 18:12:40 UTC