Subject:
|
Re: Attention all RSS geeks!
|
Newsgroups:
|
lugnet.publish
|
Date:
|
Tue, 24 Aug 2004 14:02:42 GMT
|
Viewed:
|
5825 times
|
| |
|
|
On Mon, Aug 23, 2004 at 06:15:59PM +0000, Kelly McKiernan wrote:
> True, let's hope. Yes, I'm hoping LENNI will have a very firm
> specification that is easy to implement - hence the proposed use of
> custom namespaces.
>
> Looks like Atom vs. RSS is a bigger issue than I initially thought:
> http://www.lawtechguru.com/archives/2004/02/16_rss_vs_atom_continues.html
> http://www.lawtechguru.com/archives/2004/02/13_the_great_rss_vs_atom_news_feed_debate.html
>
> http://news.com.com/2009-1032-5059006.html
>
> Just what have I stepped into? ;)
Heh, the biggest political fight of recent techies. Used to be XML, but
that has settled down considerably. I used to follow the fight, but got
really tired of it... I did come the conclusion that Mark Pilgrim (of
the Atom side) knows what he's doing considerably more than Dave Winer
(of the RSS side). But that's neither here or there.
> Even after reading all that, it's still not black-and-white for me
> to point to one over the other. I guess time for more reading. Dan,
> thanks for bringing up Atom, I had no idea it even existed before
> this, and now is the perfect time to decide core-level structure
> for LENNI.
Here's one way to decide:
http://diveintomark.org/archives/2004/08/23/kitten-fight
(warning, it IS a joke).
On Tue, Aug 24, 2004 at 05:42:13AM +0000, Kelly McKiernan wrote:
> Time to break this into a subthread, I think...
>
> [RSS v. Atom - Pro and Con Time]
>
> So now that I've spent the last few hours reading the
> <http://blogs.law.harvard.edu/crimson1/comments?u=crimson1&p=1696&link=http%3A%2F%2Fblogs.law.harvard.edu%2Fcrimson1%2F2004%2F05%2F31%23a1696
> contentious history>\[1\] of RSS v2.0 and what's become Atom (although the Homer
> Simpson in me liked its previous name, "Pie"), and one thing has become clear...
>
> It's still not clear.
(just a note, which you probably know - that's Dave's point of view.
The other side of the coin is here:
http://diveintomark.org/archives/2002/09/06/history_of_the_rss_fork
http://diveintomark.org/archives/2003/06/23/a_fresh_start
)
> Oh, and a quick aside - reading all the RSS/Atom mudslinging has given
> me a greater appreciation for the people on LUGNET, who generally
> behave themselves and act like homo sapiens, even when they disagree.
> The foofooraw over a tech spec (!) is amusing and alarming, by turns.
Quite. It's amusing for a while, but after that, when you realize that
some people ACTUALLY think that way, it gets scary.
> OK, back to biz.
>
> [Question: Should LENNI use RSS v2.0 or Atom v0.3 as a base to
> build from?]
>
> Not sure of the answer yet. I think I've got enough background to put
> together a pro/con list, though, for our purposes.
>
> [RSS v2.0 Pro & Con list]¬
> <http://blogs.law.harvard.edu/tech/rss RSS 2.0 Spec>\[2\]
>
> * [Pro:] Simple, as in Really Simple.
Yup.
> * [Pro:] Power in numbers - lots of aggregators for it.
I'm not convinced that Atom hasn't caught up, though.
> * [Pro:] It's a spec and not an API (for LENNI's purposes this is a pro, I
> think)
Can you explain this point a bit more? Just not sure why this is a pro?
> * [Pro:] Extensible via namespaces to achieve what we want (I think)
Right.
> * [Con:] Elements don't have built-in attributes like Atom (author name, etc.)
Right - which means we need to extend it further.
> * [Con:] Very loose spec, open to multiple interpretations (e.g. character type
> encoding in titles, etc.)
That's a pretty big one, IMO, especially since we want LENNI to be
international in nature.
> * [Con:] Locked, but not loaded. It's essentially a dead-end technology that
> won't grow.
Which means that one day in the future we'll have to migrate.
> [Atom v0.3 Pro & Con list]¬
> <http://www.atomenabled.org/developers/syndication/atom-format-spec.php Atom
> Spec>\[3\]
>
> * [Pro:] "Out of the box" includes better elements (author name, etc.)
Yup.
> * [Pro:] Active development community
A non-dead-end, though it will mean we might have to migrate as it
evolves.
> * [Pro:] Google uses it for Blogger
Meh - minor pro? It's used in many places, but not sure how that
affects us?
> * [Pro:] Most aggregators now support Atom
Right.
> * [Pro:] Can use namespace declaration to
> <http://bitworking.org/news/Extending_the_AtomAPI pull in lists of Themes,
> etc.>\[4\]
Looks cool - is that part of the spec yet?
> * [Pro:] Supposedly more extensible (see below)
> * [Con:] Extensibility descriptions
> <http://www.atomenabled.org/developers/syndication/atom-format-spec.php
> sparse>, at best\[5\]
True.
> * [Con:] Spec is in early phase and could change radically before finalized
I think it _could_, but I don't think it's likely to change that
radically?
> * [Con:] It's not just a spec, it's an entire API that details how HTTP/XML
> servers should parse Atom feeds and elements. This is overkill for LENNI's
> purposes (although it has a high beanie propeller RPM)
Heh. I think the fact that it includes an API might be useful - saves
us from having to develop our own? And should allow us to use existing
libraries for most of the sync we'll want to do.
> * [Con:] Less well-integrated into existing aggregators, so there's smaller
> market penetration, at least in the early stages of deployment (ooh, marketing
> speak coming out, sorry)
Sorry, my eyes glazed over :) Can you translate?
> * [Con:] More elements are required (e.g. guid), forcing more work from
> developers (but this might not be a bad thing, unless you're lazy like me)
Well, I'd definitely want to have guids - I think they're really really
important to have. I read some RSS feeds that don't include guids
(washington post's for instance), and it's really really annoying to
read the same entry multiple times, just because they fixed the spelling
of something.
> [Conclusion Thus Far]
>
> As you can see, it's about even on the pluses and minuses for each
> format. Frankly, I'm kind of put off by the bickering between each
> camp's proponents, none of the parties involved really come off
> shining. Ah well - as long as the code works, I guess.
Yah, I do wish the "world" will settle on one or the other. The more I
think about it, I get more convinced that Atom is really the way to go.
If only to avoid the dead-end that RSS is (supposedly).
> For right now, I'm slightly leaning toward RSS, but that may only be
> because I see fairly clearly the method of implementation necessary to
> get LENNI working on an RSS base. Atom may actually be closer to what
> is needed, but it's not as clear to me how to extend it, and whether
> extending the XML format will break brand-new Atom parsers or not.
I'd claim that if a newer parser fails to parse an older version of the
format, it's broken. If a feed identifies itself as Atom 0.3, and a
parser tries to force it into a 0.4 template, it deserves what it gets
:)
However, that doesn't change the fact that there isn't an official way
of extending atom yet. Need to poke around their ML to see if any
progress has been made there. I'm downloading the archive now, I'll see
what I can find.
> So here I am, sitting on the fence, waiting to be nudged one way or
> t'other. Thoughts?
I think you know which way I lean, but I'll let others chime in :)
On Tue, Aug 24, 2004 at 10:44:14AM +0000, René Hoffmeister wrote:
> My suggestions are:
>
> * Using RSS and creating specific (but standardized) tags for LEGO
> related parts of each RSS-item
Right - with the debate still going on RSS vs Atom.
> * Providing the news from all participating sites on a central place
> (e.g. legofan.org)
Well, I'd think many sites might want to keep at least some subset of
the feeds posted. But yes, there should be a central listing somewhere,
just to make it easy to find more feeds. LF works for that :)
> * Providing it not seperated by source site (like it's working on
> legofan.org at the moment), but one big RSS-index with possibilty to
> filter and show these news by the given criterias and (most
> important for sharing all the news) the possibility to get RSS-items
> generated by these criterias for the use on ones own site.
Some sort of central RSS aggregator, that will allow you to subscribe to
a custom feed as a result of a query? That would be cool, but isn't the
whole point of LENNI is to allow you to have that feed generated on your
own site in an easy way, and not rely on a central server?
Maybe I'm misunderstanding. Definitely we can do both :)
--
Dan Boger
dan@peeron.com
|
|
Message has 1 Reply:
 | | Re: Attention all RSS geeks!
|
| In lugnet.publish, Dan Boger wrote: (a lot of very informative stuff that made my head spin a bit, although I now think I know which Dan favors which does carry some weight with me, Dan's often right, technically... ) So what's the net here? Does (...) (20 years ago, 24-Aug-04, to lugnet.publish)
|
Message is in Reply To:
| | Re: Attention all RSS geeks!
|
| (...) Same here. Good point about spec changes, Jake. I'd love to do a pro/con list of RSS/Atom, but I don't know enough yet, so I guess I'll have to research that some more. My main concern is using something that can be used by the maximum number (...) (20 years ago, 23-Aug-04, to lugnet.publish)
|
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
|
|
|
|
