- From: David Wagner <dwagner@kevric.com>
- Date: Thu, 30 Dec 1999 16:07:13 -0600
- To: "'eldred@mediaone.net'" <eldred@mediaone.net>
- Cc: "www-html@w3.org" <www-html@w3.org>
On Monday, December 27, 1999 5:18 PM, Eric Eldred [SMTP:eldred@mediaone.net] wrote: > David Wagner wrote: > > > ... > > I think an accesskey to switch among alternate stylesheets is a good idea. > > Why? Well, one can make an alternate stylesheet to show only the heading > > levels, and it would be a view of the table of contents. You may have one > > stylesheet for editing, showing the content of all the INS and DEL elements, > > > > and another for viewing what will be the final document. Of course, > > choosing > > the print medium stylsheet will give you a print preview, something not all > > browsers do. Having a quick key to switch among these is a good thing, > > whether > > there is a navigational element shown or not. > > but how does the user know which accesskey to > press to switch stylesheets? I would let users know (perhaps in a help file) what these shortcut keys are. My pages are often already crowded with information. If I have a good useful document available in different styles for various media and accessibility requirements, with a few translations available, and part of a set of documents in a collection (with prev/next/chapter/toc/index/glossary/bibliography... links), I would need dozens of Anchors on every page to display all these choices. My thinking is I have a help file explaining the standard keys used to access these common LINKs from any given page. > > i think the idea behind ACCESSKEY was to put > in the user's display some indication just > before a link, that retrieving the link is as > easy as pressing that key (or maybe in > combination with CTRL or ALT or CLOVERLEAF.) > for example, as Lynx does, the octothorpe > before the nav bar at the top of the page. > > this could work easily with A ANCHORS, but > i don't see how it would work with LINKS > in the HEAD element that are not displayed. > > it might work for nav bars LINKS, if they > were displayed the same as Lynx, but there > is still a complication in that the big 2 > browsers don't support that. adding the > "feature" to enable accesskeys for all links > would seem to me backwards incompatible. I had been crossing my fingers the big 2 would display LINKs as appropriate for the operating system, as menu items, toolbar buttons, voice prompt, or such. The MSIE 4/5 "Links" toolbar at first glance looked like it would do so. Capturing the ACCESSKEY behavior fo LINKs would, as you point out, probably not work in current browsers. > > on the other hand, if the user agent displays > the HEAD LINKS at the top of the page with > optional ACCESSKEYS to switch stylesheets > and so on, then adding the ACCESSKEY to LINK > would be easy and useful. but it would be > redundant, since placing the LINK in the > HEAD is an alternative to an ANCHOR in the > displayed text, implying non-visibility in > the display. once it is visible in the > display it is really the same as an ANCHOR > and so the ACCESSKEY in an ANCHOR is already > allowed. a page author can just write for > ANCHORS instead of LINKS if that is the > desired behavior. I agree. LINKs in the HEAD should not be displayed on the page. But I don't think page authors can tie stylesheets in with A elements without scripting the behavior to override the browser's normal processing. > > as an alternative, the behavior of Lynx > might be something different that other > browsers might wish to copy. a global > option in Lynx is to have all links preceded > by a number. this shows on the display and > works quite the same as ACCESSKEY except that > one enters a number instead of pressing a > key. > > there was a suggestion a month or two ago > here that ID (for example for an H1 element) > have the semantics of TABINDEX--thus the > next TAB in order would be indicated by > that attribute of an element, instead of > by TABINDEX. nobody commented on that, > but i think it wouldn't work because an > ID attribute value should not begin with > a digit, while the TABINDEX attribute value > is supposed to be an integer. and ID and > TABINDEX have different purposes that should > not be confused. > > but allowing ACCESSKEY to be an integer > instead of or as well as a key would > allow a new browser to mimic the behavior > of Lynx, maintain backwards compatibility, > and also provide some usefulness for user > agents for accessibility. if that is not > desired, then making visible the HEAD LINKS for > navigation, with TABINDEX allowed as > an attribute, would also be good. then > Lynx could change from ID to TABINDEX for > its tabbing order, and MSIE change from NAME to > TABINDEX. but likely this will be opposed > because of backwards compatibility concerns? MSIE 5 allows accesskey="5" [Alt-5] or whatever digit you like. I agree to the need to keep ACEESKEY, TABINDEX, and especially ID attributes seperate, and could not possibly live with a limit of 10 links on each page. (See above.) > > and as far as LINKS in HEAD that are not > nav bars, the user agent would have to add > code to do something once the LINK is > selected. as for example, switch to another > language (on a new page). but without a > prompt, this would be hard to implement. > wouldn't it be better in that case just to > try to do it with ANCHOR in the visible page? I don't think it would be that difficult to program a toolbar displaying LINKs in MSIE5; there are plenty of similar custom toolbars around, and MSIE 4/5 looks like it will even support entire stylesheet changes and additions as well as simple navigation. I guess what I am getting at is there are many useful little things, relatively easy to implement, to put on pages especially when each page forms part of a larger work. Many of these things would be very useful for only a few users, but I really can't clutter the 640x480 screens of the many for benefit of the few. However, I really want to make my work shine for every user, and having a bunch of magic keys available seems like a good way to do this.
Received on Thursday, 30 December 1999 17:10:09 UTC