- From: Misha Glouberman <misha@the-wire.com>
- Date: Wed, 21 Aug 1996 23:34:54 -0400
- To: www-annotation@w3.org
>Jacob Palme wrote: >Do I understand that the intention is to allow anoyone to create >annotations on any web document, and that these annotations will >then be visible to anyone who looks at the web document? That's the intention of my proposal. The W3C proposal allows for something very similar to what you describe but with a small difference- if you add a comment to a page it won't quite be visible to _anyone_ who looks at the document- it'll be visible to anyone subscribed to the same annoation set who looks at the document. I believe this small difference to be critical. I gather that in both proposals, annotations happen independently of the desires of the document owner, though I may be making unwarranted assumptions about w3c-annot. Is there a finalized position on this? > >Can this be implemented without support from the web server for >the document being annotated? If you mean "Do you envision this being a feature", then yes. If you mean "Do you really think this is viable", my answer remains that I don't know. Certainly it's possible in principle. And certainly the challenges of scale are substantial. I'm banking on a few different ways that the question of scale could be tackled: 1) The idea of a universally commentable web is found compelling enough by enough sysadmins (or enough of their users) that a highly distributed Usenet-like model could be supported. 2) The scale of the enterprise, while seen as daunting to system designers, is seen as appealling to web entrepreneurs, who want to get as many hits as they can. Web comments are set up as an ad-based for-profit system. Much as advertising on the web (and elsewhere) bugs me, it does provide a nice model for scalability of online projects- more users translates pretty directly into more revenue, which can be reinvested in more hardware etc. Either of these could be assisted by: 3) The system gets set up as a workable hack, piggybacking onto existing large writable data collections on the web like Usenet and Deja News. This would be a lousy solution in anything but the very short term, but it could help build interest that would help bring about one of the other two models. I think the potential payoff is big enough that it's worth trying as hard as possible to see whether the challenges of scale can be dealt with. > In the Web4groups project, we plan to allow the authors >of pages control of whether discussions are allowed or not on >their papers. I know. I'm curious: what are cases where you think it'd be desirable for hosts to be able to disallow comments? My paper's pretty much based on the opposite assumption. The pages that I'm most interested in annotating are the ones least likely to allow annotation- hate speech, spurious advertising claims, and misleading political rhetoric, for example. I don't want any of these to make themselves immune to comment just by inserting a special html tag. All of this (including a short reference to the inclusion of the <nodiscussion> tag in your web4groups paper) is described a little better in my paper, at http://www.muchmusic.com/muchmusic/cyberfax/annot.html >And Daniel LaLiberte wrote: >One might set up a >parallel DNS tree for annotation services, but this potentially has >the same problem since every server might have some legal power >over the corresponding annotation server. I don't think I understand what you're getting at here, and I'm really interested in knowing. If you could actually get enough people to host 3rd-party annotation servers, a parallel DNS tree of some sort seems like a perfectly fine solution for allowing the browser to find the appropriate annotation server for a given page. Are you imagining that web servers would volunteer to run their own annotation servers, and then try to control their content? Or are you suggesting that servers would have some inherent legal claim on 3rd party annotation servers? >And Wayne Gramlich wrote: >I honestly do not think that there is any way to "force" people >to view public annotations. Either the content provider (i.e the >web server) has to agree to participate or the content reader >(i.e. the web browser) has to explicitly do something to solicit >the annotations. A good thing too. I don't think anyone wants to lose the option of viewing a page without its annoations... The question, then, is whether there's a way to force pages into being annotatable, for those users who do want to see the comments. - Misha
Received on Wednesday, 21 August 1996 23:34:10 UTC