Search the web
Sign In
New User? Sign Up
rss-dev

Group Information

  • Members: 1344
  • Category: XML
  • Founded: Jul 27, 2000
  • Language: English
? Already a member? Sign in to Yahoo!

Yahoo! Groups Tips

Did you know...
Hear how Yahoo! Groups has changed the lives of others. Take me there.

Messages

  Messages Help
Advanced
Addressing RSS's logical model   Message List  
Reply Message #268 of 7450 |

Contents of this message:

* Background
* Proposal
* Example
* Summary
* Attachments: oldmodel.png, newmodel.png


BACKGROUND

The current logical model of the RSS 1.0 proposal has some deficiencies,
which others have discussed in this group, and I would like to propose a
way around them.

The problem centers around the relationship between a channel and items.
Although most items in use today belong explicitly to one and only one
channel, this is not the general case.

For instance, from XMLhack I could produce 30 channels--a general one,
and then one for each of the specialized categories that we report on.
Clearly the items would be shared among several channels.

The current 'inchannel' and 'position' properties do not permit this.
Consider the aggregated case where all the channels and items escape out
of their host files. It is not by any means inconceivable that more than
one channel contains an item with an identical URI (example: a channel
which publishes commentary). Difficulties could well arise here.

It's easiest to see with a diagram, and I attach two to this message.
The file 'oldmodel.png' shows how things are now. There are two
channels, and they share itemC--this is not possible in the current
model, and the conflict is shown as a dashed line in the relationship.
Also note the direction of the arrows in the oldmodel diagram.


PROPOSAL

I propose that the list of items in a channel should instead be a
property of the channel itself, along with their position in a channel's
line-up. This allows the explicit reuse of items independently of the
channel (which my.userland, Moreover and Meerkat all do already--there's
your use-case!)

This alternative model is shown in the file 'newmodel.png'. Note that
the rdf:_1 ... properties are created implicitly from the ordering of
item references in a channel list, not by the explicit method used by
the current 'position' property (note for language lawyers: I have also
omitted the illustration of the rdf:type property being Seq, it being
not strictly necessary in considering this model.)

(This, incidentally, would also allow the creation of a channel
description independently of the item description, allowing the use of
RSS to describe your personal "hotlist" of news stories, a bit like
Meerkat's 'mob' system.)

So, what does it look like in the syntax?

Add the list of channel items to the channel as a property, thus:

<items>
<rdf:Seq>
<rdf:li resource="http://xmlhack.com/read.php?item=500" />
<rdf:li resource="http://xmlhack.com/read.php?item=499" />
<rdf:li resource="http://xmlhack.com/read.php?item=498" />
<rdf:li resource="http://xmlhack.com/read.php?item=497" />
<rdf:li resource="http://xmlhack.com/read.php?item=496" />
<rdf:li resource="http://xmlhack.com/read.php?item=495" />
</rdf:Seq>
</items>

The ordering is also then implicit, and no 'position' property is
required.

I would then advise the removal of 'inchannel' and 'position' from the
<item> node.

(I've had it pointed out to me that they could actually stay in, and be
overridden by the items property of a channel. I tend to feel this is
a poor compromise at what is still an early stage.)

There's also a knock-on of this, for images and textinputs (much as I'd
like to hurl textinputs out of the window!). I propose the following
notations for those:

<images>
<rdf:Seq>
<rdf:li resource="http://xmlhack.com/images/mynetscape88.gif" />
</rdf:Seq>
</images>

<actions>
<rdf:Seq>
<rdf:li resource="http://xmlhack.com/search.php" />
</rdf:Seq>
</actions>

(For textinput, I prefer the more generic term 'action', as I would like
to see a future RSS module be developed to describe richer URL query
functionality than just the plain old textinput).


EXAMPLE

Here's a complete example (although it doesn't exercise 'textinput' it's
enough to get the idea!)

<?xml version="1.0" encoding="utf-8"?>
<rdf:RDF
xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns="http://purl.org/rss/1.0/"
xmlns:rss091="http://purl.org/rss/1.0/modules/rss091/"
xmlns:dc="http://purl.org/dc/elements/1.1/"
>

<channel rdf:about="http://xmlhack.com/rss10.php">
<title>XMLhack</title>
<link>http://xmlhack.com/</link>
<dc:source resource="http://xmlhack.com/" />
<description>
Developer news from the XML community
</description>
<rss091:language>en-us</rss091:language>
<dc:publisher>Edd Dumbill (mailto:webmaster@...)</dc:publisher>
<dc:creator>Edd Dumbill (mailto:editor@...)</dc:creator>
<dc:rights>Copyright 1999-2000, xmlhack team.</dc:rights>

<images>
<rdf:Seq>
<rdf:li resource="http://xmlhack.com/images/mynetscape88.gif" />
</rdf:Seq>
</images>

<items>
<rdf:Seq>
<rdf:li resource="http://xmlhack.com/read.php?item=731" />
<rdf:li resource="http://xmlhack.com/read.php?item=730" />
<rdf:li resource="http://xmlhack.com/read.php?item=729" />
<rdf:li resource="http://xmlhack.com/read.php?item=728" />
<rdf:li resource="http://xmlhack.com/read.php?item=727" />
<rdf:li resource="http://xmlhack.com/read.php?item=726" />
</rdf:Seq>
</items>

</channel>

<image rdf:about="http://xmlhack.com/images/mynetscape88.gif">
<title>xmlhack</title>
<url>http://xmlhack.com/images/mynetscape88.gif</url>
<link>http://xmlhack.com</link>
<rss091:width>88</rss091:width>
<rss091:height>31</rss091:height>
<rss091:description>News, opinions, tips and issues concerning XML
development</rss091:description>
</image>

<item rdf:about="http://xmlhack.com/read.php?item=731">
<title>Open Digital Rights Language</title>
<link>http://xmlhack.com/read.php?item=731</link>
<dc:description>Renato Iannella of IPR Systems has announced version 0.5
of the "Open Digital Rights Language" (ODRL), an XML encoding for expressing
rights over
digital assets.</dc:description>
<dc:subject>XML</dc:subject>
<dc:subject>W3C</dc:subject>
</item>

<item rdf:about="http://xmlhack.com/read.php?item=730">
<title>New release of XML-RPC for PHP</title>
<link>http://xmlhack.com/read.php?item=730</link>
<dc:description>A new version of XML-RPC for PHP has been released,
incorporating new helper functions in addition to bug fixes.</dc:description>
<dc:subject>Protocols</dc:subject>
<dc:subject>Tools</dc:subject>
</item>

<item rdf:about="http://xmlhack.com/read.php?item=729">
<title>DocBook XML v4.1.1 released</title>
<link>http://xmlhack.com/read.php?item=729</link>
<dc:description>A new version of DocBook XML, the XML version of the
popular technical documentation DTD, has been released by Norm Walsh.
</dc:description>
<dc:subject>SGML/XML</dc:subject>
<dc:subject>OASIS</dc:subject>
</item>

<item rdf:about="http://xmlhack.com/read.php?item=728">
<title>Help wanted: XML for the visually impaired</title>
<link>http://xmlhack.com/read.php?item=728</link>
<dc:description>JP Schnapper-Casteras has posted to XML-DEV asking for
volunteer developers to work Project Ocularis, a Linux distribution geared to
the needs of the visually impaired.</dc:description>
<dc:subject>Community</dc:subject>
</item>

<item rdf:about="http://xmlhack.com/read.php?item=727">
<title>RDF-based tool for managing photographs</title>
<link>http://xmlhack.com/read.php?item=727</link>
<dc:description>Bert Bos has announced the availablility of rdfpic, a
tool for managing photos using the PhotoRDF work described earlier this
year.</dc:description>
<dc:subject>RDF</dc:subject>
</item>

<item rdf:about="http://xmlhack.com/read.php?item=726">
<title>COM SOAP services</title>
<link>http://xmlhack.com/read.php?item=726</link>
<dc:description>Simon Fell has released version 1.2 of 4S4C,
a SOAP to COM integration layer.</dc:description>
<dc:subject>Protocols</dc:subject>
<dc:subject>Tools</dc:subject>
</item>

</rdf:RDF>


SUMMARY

The current model (which Ian Graham rightly urges us to focus on, not
the syntax) places the channel-item relationship wrongly in the item.
This is true of images and textinputs too. My proposal fixes the model
to properly represent this relationship. It doesn't affect the RSS 0.9
compatibility of RSS 1.0 processors any more than the existing syntax.

-- Edd


Sun Aug 27, 2000 6:26 pm

edd@...
Send Email Send Email

Attachment
oldmodel.png
Type:
image/png
Attachment
newmodel.png
Type:
image/png
Message #268 of 7450 |
Expand Messages Author Sort by Date

Contents of this message: * Background * Proposal * Example * Summary * Attachments: oldmodel.png, newmodel.png BACKGROUND The current logical model of the RSS...
Edd Dumbill
edd@... Send Email
Aug 27, 2000
6:27 pm

It appears that my message doesn't format well in the eGroups archive, so I've put it on a web page too: http://www.usefulinc.com/rss/modelProposal (Sorry also...
Edd Dumbill
edd@... Send Email
Aug 27, 2000
7:22 pm

I also like this proposal. In fact, it has the same logical model as the original spec. Talk about full circle... :) ...
Dan Libby
dan@... Send Email
Aug 28, 2000
10:04 pm

... This is why I once again recommend my proposal to use more abstract URIs, rather than specific URLs. This would also allow for multiple item descriptions...
Aaron Swartz
aswartz@... Send Email
Aug 28, 2000
11:47 pm

Someone please correct me if I'm wrong, but I think the answer may be implementation specific. The version(s) of RDF I've worked with (written by guha) would...
Dan Libby
dan@... Send Email
Aug 29, 2000
6:00 am

... I'm +1 for using rdf:Seq for <items>, but -1 for use in <images> and <textinputs>. Images and text inputs are not well modelled as sequences. They'd ...
Ken MacLeod
ken@... Send Email
Sep 26, 2000
4:21 pm

... I would like to see them rendered as you describe an Alt, personally. There are many different uses for images/textinputs and having multiple of these with...
Aaron Swartz
aswartz@... Send Email
Sep 26, 2000
6:46 pm

... If we think we need more than one of either, I recommend: Alt for images (choose one from many) Bag for textinputs (choose any/all in any order) But Ken's...
Edd Dumbill
edd@... Send Email
Sep 26, 2000
6:56 pm

My vote is for multiples for image. From experience sites normally request a choice from multiple options when this type of thing is arranged over the phone! I...
David Galbraith
david@... Send Email
Sep 26, 2000
7:22 pm

Howdy, My vote would be to keep the current 1 optional image, 1 optional textinput, and multiple items in RSS 1.0. Multiple images and textinputs may always...
Rael Dornfest
rael@... Send Email
Sep 29, 2000
4:07 am

... Just to be clear, when someone _does_ model what to do with possible multiple images and textinputs, or provide use cases, they will need to be very...
Ken MacLeod
ken@... Send Email
Sep 27, 2000
3:41 pm

Howdy, On 9/27/00 8:41 AM, Ken MacLeod at ken@... wrote: ... Agreed. This is in keeping with the centripetal model with commonalities migrating...
Rael Dornfest
rael@... Send Email
Sep 29, 2000
4:13 am
Advanced

Copyright © 2010 Yahoo! Inc. All rights reserved.
Privacy Policy - Terms of Service - Guidelines NEW - Help