Re: RESOURCE

Erika Madja (madja@IRO.UMontreal.CA)
Thu, 10 Oct 1996 14:37:49 -0400 (EDT)


Date: Thu, 10 Oct 1996 14:37:49 -0400 (EDT)
From: Erika Madja <madja@IRO.UMontreal.CA>
Message-Id: <199610101837.OAA25016@fressel.iro.umontreal.ca>
To: timbl@w3.org
Subject: Re: RESOURCE
Cc: html@w3.org, www-html@w3.org


> Erika Madja wrote:
> > 
> > 1- Can  the <meta>  element  been used to  convey  application-specific
> > parameters such as  multimedia QoS  parametres in our case ?
> > Existing  exemples I have seen usually  define informations  such as
> > documents author, keywords etc. that are relevant for indexing and catalogin
> > documents.
> 
> YEs, certainly, the META element can be used for all kinds of
> information, including
> application-specific parameters using X-tags which you use as
> exerimental -- but
> you may find that the QoS parameters you interested in are of interest
> to a larger community too.
 
What is a X-tag ? 

> 
> > 2 - Can the information  described by  the  <resource> elements  by  accessible
> >   to  a plug-in ? How ?  as I understand, it can be extract by the user agent.
> >   Now, what interaction is provided to pass this information to a plug-in.
> 
> That is the tricky bit.  If ou try to extend a browser by implementing a
> generic object
> which itself takes on the task of chosingthe best specific resource,
> then you don't
> currently have access to any resource information which the browser
> has.  In fact, it 
> would be very useful to b able to get at htat and maybe even add to hte
> database of
> related resources. But browsers don't currently parse <resource> --
> there isn't even
> a general agreement to go that way.  But people are taking about it. I
> don't know
> waht APIs current browse have for "tell me ll you know about this URL"
> -- could
> be someone on the Cc list will be able to help here.

I find out that Netscape currently  has in it API, a  GetURL method 
that "allows  the  plug-in to retrieve arbitrary data from the network or file
system and process the  data itself" 
(http://home.netscape.com/eng/mozilla/3.0/handbook/plugins/pi3.htm).
Since <resource> will typically refer to generic document, I'm not sure
this will work. The  GetURL method  has certainly to be  rewrite to 
to handle this. Concerning generic document, it's not clear to me,
how myimage in the following exemple should be handled by systems.

<resource href="myimage">
    <link rel="content-type-specific" href="myimage.png">
    <link rel="content-type-specific" href="myimage.gif">
    <link rel="content-type-specific" href="myimage.avi">
</resource>
 <img src="myimage"> 

It is explain in the specification of <resource> that 
"Some HTTP servers use format negotiation to select
from specific documents with different content-types
or languages." Could you please explain me how it applies 
to the example  above?


> 
> > 3 - The link relation (choice-specific relation) for the <resource> element is
> >  to be used by user agents to  decide  a choice. In the case of our project,
> > it is the responsability of plug-in to make this choice according to Qos parameters and user preferences . Is this possible ?
> 
> "choice-specific" was supposed to indicate user choice, wheras
> "resolution-specific"
> for example you could invent to mean that the choice was one of image
> resolution.
> Everyhting except choice-specific was intended in the draft to be
> resolved by 
> the machine: choice-specific explicitly suggests prompting the user.
> 
> 

May be the xxx-specific relationship for link will be useful here.
Is it intended to allows a QoS-specific relationship for example ?


> Hope this helps. I think there are two questoins here - how this should
> be implemented
> in the long run, with the help of browers, and how it can be fudged in
> the near term
> with existing browers.
> > 
  Exact. I need an answer to both of the questions  
  

 Excuse-me for my bad english.
 Thank a lot. 
 
Erika MADJA
Universite de Montreal
madja@iro.umontreal.ca