- From: David Woolley <david@djwhome.demon.co.uk>
- Date: Sun, 5 Aug 2001 14:48:37 +0100 (BST)
- To: w3c-wai-ig@w3.org
> > Why not a link to the performer's page - many of which have song samples on > the page. That won't pose copyright concerns and will keep the page This is regrettably becoming a gray area. A number of commercial web sites already forbid deep links and attempt to restrict the ways in which home page links are presented. This is obviously contradictory to the web, but commercial sites are normally not web sites but rather "web browser viewed" sites. As I mentioned in another reply, the dangers of linking to other sites is the cover story on the current issue of one UK e-business magazine. Using anything other than a straight, unframed, <a> element (or, though I don't reccommend this, location.href=/location.replace) reference will normally be considered tantamount to creating a derived work by the company lawyers, whatever the actual copyright law status. bgsound/embed from the band's site is a no-no without explicit permission. In a commercial world, it seems unavoidable that embedding will be brought under copyright law, if not already the case. Deep linking is more of a problem; the web is severely compromised without deep linking, but many business models are dependent on going through home pages. What I would hope for is some sort of compromise in which sites must actively indicate++ that a page may not be deep linked before there is any protection. I would hope that home page linking could be made something that couldn't be excluded, but I could see, for example, the band wanting the ability to prevent links to its site from critical sites, or sites incompatible with their image. > uncluttered. The first bullet in 3.4 allows us to "provide a graphic > equivalent or link to content that contains a graphic equivalent". Can we > specify linking to any alternative content - not just graphics? I have no problems with generalising this point. ++ This needs to be done with HTTP delivered meta data. Refer based methods are vulnerable to forgery and cause conflicts with privacy requirements as some people use bogus Referer headers, or none at all, to prevent click trailing. Whilst a commercial site may make the ability to click trail a condition of access, one doesn't want to couple this with deep linking restrictions. The meta data must be delivered at the HTTP level (although many people seem unaware of how to set HTTP meta data) because the resource may not be HTML (e.g. it may be a sound file).
Received on Sunday, 5 August 2001 12:38:06 UTC