- From: Nir Dagan <nir.dagan@econ.upf.es>
- Date: Sat, 06 Jun 1998 00:07:05 GMT
- To: w3c-wai-ua@w3.org
- CC: dagan@upf.es
I have some comments on the following topics in the UA guidelines: 0. tables 1. objects and longdesc 2. image maps 3. frames 4. meta refresh 5. etc... 0.tables ------------ PRIORITY 1 Make it optional to render table summary. It seems to me somewhat important to allow users to read about the organization of a table, if they so wish. 1. objects -------------- I find the following highly confusing: "4.2.2 ...The innermost text of the OBJECT is considered its alternative text. If the author has not supplied alternative content the browser should display "Image"." In my understanding of the spec. the content of the object, and not the innermost text is the "alterantive text" of an object. The above guideline contradicts the specs. Consider <OBJECT>text1 <OBJECT>text2</OBJECT></OBJECT> if both types of objects are turned off /unsupported, text1 text2 should be rendered, not only text2. If only the outmost object is not rendered then text1 followed by the second object should be rendered. Moreover the guideline leads to the following pardox: the markup <OBJECT data="foo.png"><IMG src="bar.gif" alt=""></OBJECT> should be rendered as nothing when png and gif are turned off. but: <OBJECT data="foo.png"></OBJECT> should be rendered as "Image". In my view it should be assumed that an author wants nothing to appear in the second case too, when png is turned off. Empty OBJECT is valid HTML, and the browser shouldn't assume authors make mistakes if the document is valid. Browsers don't fix spelling mistakes or rude words, and shouldn't "fix" empty OBJECTS. With IMG it is a different strory. alt is _required_ by the spec. So the guidelines suggest how to recover from the author's _syntax error_. longdesc: "4.2.3 When an IMG element has a value for the "longdesc" attribute and the user has turned off the display of images, render a D-link inline to give access to the long description." I think that users should have the option to turn off the presentation of longdesc as D-link. 99% of people who turn images off don't want to be bothered with images. Not making it optional, will discourage authors (e.g., me) from writing longdesc. ( I don't write d-links since they reduce usability; I will write longdesc only if I know that they don't reduce usability and are hidden unless explicetily set by the user to show them). 2. image maps. --------------------- Guidelines may mention supporting alt text to AREA elements. All browsers that I know of support only the alt of the IMG used as a map, which is good for nothing. This is priority 1. You can't navigate a site without a navigation bar. 3. frames. ------------- Somethings to add: Allow to turn off frames. When frames are turned off: 1. render the NOFRAMES element(s) 2. Make it an option to render the FRAME elements as a list of links. When doing so use the value of the title attribute of the FRAME as the link description, if not available, use the value of name, if not available use the value of src. With frames on: Allow users to override author's setting of scrolling=no, and noresize. 4. Meta refresh ---------------------- UA who support autoloading of documents with <META http-equiv=Refresh ...> should allow turning this feature off, possibly with presenting the url of the new page as a regular link. ("possibly", as UAs are not required by HTML to support this in any way) 5. etc. --------- Among the usability improving things (inspired from Opera): 5.1 allow user to define different text decoration for visited and unvisited link (e.g. underline unvisited, do not underline/strike through visited). This is good for color blind people or monochrome screens. Distinction by color is not sufficient. 5.2 allow users to set "always show image border when used as link" (override border=0) as well as "always show active frame border". Regards, Nir Dagan Assistant Professor of Economics Universitat Pompeu Fabra Barcelona (Spain) email: dagan@upf.es Website: http://www.econ.upf.es/%7Edagan/
Received on Friday, 5 June 1998 11:04:41 UTC