- From: Paul Topping <ptopping@lightside.com>
- Date: Mon, 12 Sep 94 15:55 PDT
- To: www-html@www0.cern.ch
Brian Gaines wrote (gaines@cpsc.ucalgary.ca): > Since images are fetched separately the best way to deal with this is > for the image data structure to carry the information. Intelligent clients > can then offer exact placement, and others can just put them where they > wish. > > GIFs have provision for cutomizable annotation which is both easy to > add and decode. Other graphic formats have similar facilities. > > It is better to carry such layout information in the data structures > of the graphic objects rather than overload HTML with duplicate information. I agree that it is usually good to avoid duplication of information. However, in this case unless the HTML specification has its scope expanded to cover the specifics of these customizable annotations, this approach will not receive any of the advantages of standardization. In particular, intelligent clients will probably not address this issue with any consistency. In Brians's scenario, an HTML document that includes an IMG with an alignment attribute (top, bottom, middle), an "intelligent client" would not have much choice but to follow the alignment instruction rather than searching thru a GIF file for a baseline annotation. Only if the IMG construct included no alignment at all would the client feel confident in dealing with baseline alignment. Also, the baseline info is independent of the graphic format; the client can process this instruction even if it is relying on other software to render the image. If it is done in the IMG construct we have only to specify one way to do it and clients can implement it once to cover all supported graphic formats. If it is embedded in each different picture format, the clients must actually process image data rather than simply handing the image off to some generic picture-rendering code. As to the issue of putting formatting info in HTML docs, this feature would tie the formatting info to the image which is, by its nature, formatting information already. I don't want to beat a dead horse on this issue, but here is one more idea --- perhaps "baseline" could be an acceptable value of the IMG alignment attribute. This would direct the client to look for baseline info within the graphic. Perhaps the HTML spec could specify how such info is to be embedded in the most popular HTML-related graphic format (is this GIF?). I firmly believe baseline alignment is not just a special-purpose feature. Almost every document processor ever implemented is built around the assumption that the combination of text and graphic elements on a page can be used to express just about any visual information we can imagine. HTML, SGML, and word processors place a structure on the text elements. Draw and paint programs create and maintain structure within each graphical element. Unfortunately, our current technology allows very few relationships between graphic and text elements. The graphic can be placed horizontally with some control (we can put spaces in front of the graphic to move it to the right). Why don't we have a similar level of control for moving a graphic vertically relative to a text line? Paul Topping Design Science, Inc. ptopping@lightside.com
Received on Tuesday, 13 September 1994 00:56:13 UTC