To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.publishOpen lugnet.publish in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Publishing / 4592
4591  |  4593
Subject: 
Re: Attention all RSS geeks!
Newsgroups: 
lugnet.publish
Date: 
Mon, 23 Aug 2004 15:41:44 GMT
Viewed: 
3580 times
  
In lugnet.publish, Dan Boger wrote:
On Mon, Aug 23, 2004 at 04:39:07AM +0000, Kelly McKiernan wrote:
* Given the objectives listed above, is RSS the best platform to use
  in solving these issues?

I've become convinced that Atom might be a better solution.  If only
because it has better mechanisms to ensure that entries are identified
uniquely, and support for meta information about authors, etc.

I've just looked over some of the Atom docs, but it seems very similar to RSS to
me, only much newer. Is Atom supported by many current blog or web publishing
software? My thoughts were that straight RSS feeds would work within the LENNI
framework, and enhanced feeds would be better for LENNI-aware parsers, so
there's some graceful degredation. It may be that I'm not as familiar with Atom,
but I get the impression that starting with Atom would be starting more from
scratch, without an existing user base like there is with RSS.

Specifically, the docs I've reviewed are here:
http://www.ietf.org/html.charters/atompub-charter.html
http://www.ietf.org/internet-drafts/draft-ietf-atompub-format-01.txt

*   Does RSS need to be extended at all to meet these needs? Is there
    another way of using standard RSS to solve these objectives?

Probably not - which is why Atom might be a better solution.  Note, I do
say "might" because I'm not as well versed in Atom yet.

I'd like to see more info about where an unmodified Atom would be a better
solution for LENNI than RSS. If it could work "out of the box" then it could be
a better solution, but only if it's something that could be read by existing
aggregators. If there's a better spec but nobody uses it, that's kind of a
problem.

* Assuming the answers on 1-3 are "yes", is RSS 2.0 the best version
  to work with? (I think it is, since it's relatively simple like
  0.91, but can use namespaces without having to include a ton of RDF
  crud like in v1.0.)

The problem with RSS 2.0 is that it's a really really vague standard.
That means the different clients might render the exact same posts in
different ways, making the publishing a lot more complex.

Actually, I kind of like that the RSS 2.0 spec is fuzzy... it forces developers
to support core functionality and maintain backwards compatibility. Anything
that the parsers don't understand should be thrown away - so LENNI-specific
content (like AFOL) wouldn't be displayed at all in an off-the-shelf RSS
aggregator display. Hard-n-fast specs tend to make developers sloppy, so there
isn't as much attention to error-correction as in looser specs. RSS is, to put
it truthfully, a gloppy conglomeration of a bunch of different specifications in
use (seven, I think) and parsers need to be bright enough to wind through them
all.

    <bricks:fanage>AFOL</bricks:fanage>

Shouldn't this be a numerical age?  Otherwise we need to define what's
AFOL...

Good question, that's a valid discussion point. I'd initially thought filtering
by age interest would be "Kids" or "Adults" or "all" but if it makes sense to
make that more granular, that's certainly possible.

* Should the updated syntax be channel-level or item-level?

The additional tags should probably be item-level, but I'm not sure if
that's what you're asking?

Yes, that's what I was thinking as well.

<bricks:theme>Trains</bricks:theme>

How do we support multiple themes?  Comma separated list?  Make
parsing/filtering a bit more complex.  Multiple entries maybe?
* Should or can Themes be multiple choice?

You should be able to have multiple themes, but not from a set list.
New themes get added all the time, we shouldn't have to update the spec
each time just to say "Fantasy Mice" is now a valid theme :)

Exactly. I'm not sure of the best method for this - the themes list should have
some sort of standardized entry so multiple parsers can filter appropriately. We
wouldn't want to have something like "Train" and "Trains" and "choochoo" be the
same theme... there should be something set so that it's "Trains".

Question for the RSS gurus: can there be multiple item-level elements with the
same name? e.g.:
<bricks:theme>Trains</bricks:theme>
<bricks:theme>Town</bricks:theme>

* What type of "approval" process should be put in place for either
  news or calendar events, if any? What's to prevent a malicious
  person from entering XXX type of content on a KFOL feed?

Nothing at all.  And I can't imagine how you can force an approval
process?  The only thing users can do is be aware where the item came
from, and apply judgment as to what's their level of trust to that
source.  Since the feeds will be published from each individual sites,
and not from a central server, it's up to the sites to make sure their
feeds are "clean".

Maybe it's a matter of providing simple ways within aggregators to remove feeds
from your list, if the content is objectionable. Having dealt with a web site
targeted at kids, my first instinct is to provide easy mechanisms to deal with
the inevitable troublemakers.

* What have I missed? Any other suggestions? Is there something like
  this already in existence?

This is exactly what RSS (and now, Atom) was designed for.  Not sure if
we need to do much more than provide easy ways to include feeds, and
publish them...  Though, I do like the filtering capabilities that the
new tags can provide.  Again, it's possible that Atom already provides
all this, without us having to extend it (much):

  http://www.intertwingly.net/wiki/pie/Primer

I took a look but it's still pretty cryptic to me. I'm very open to looking at
other specs to base this system on, if they make more sense. At this point, I'm
still not sure of the benefits of using Atom, if you (or anybody else) have some
more information about the value of Atom for LENNI, please post here.

As I mentioned in email, I think this is a great idea, and would be very
useful to the community.  I'd be happy to help out, and provide Perl
modules for parser/aggregator (and publisher!).

Great! I appreciate your comments, looking forward to continuing this process.

Thanks,
Kelly



Message has 1 Reply:
  Re: Attention all RSS geeks!
 
(...) AFAIK, yes - most CMSs, and almost all aggregators do support atom. (...) I don't think there's much of a difference between emitting Atom or RSS/LENNI, but I don't have that strong a preference. If everyone else is dead against Atom, I'll go (...) (20 years ago, 23-Aug-04, to lugnet.publish)

Message is in Reply To:
  Re: Attention all RSS geeks!
 
(...) I've become convinced that Atom might be a better solution. If only because it has better mechanisms to ensure that entries are identified uniquely, and support for meta information about authors, etc. (...) If we're using RSS, using a custom (...) (20 years ago, 23-Aug-04, to lugnet.publish, lugnet.general, lugnet.fun.community, lugnet.off-topic.geek, lugnet.org)

31 Messages in This Thread:









Entire Thread on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact

This Message and its Replies on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact
    

Custom Search

©2005 LUGNET. All rights reserved. - hosted by steinbruch.info GbR