... I have been developing something a lot better (for java programmers) than an ontology editor that may be of interest to java syndication people, and also...
84
patrick chanezon
chanezon
Feb 1, 2005 5:53 pm
Just a few high level comments. I can't comment on the details since I would need to go in the code and I don't have any time to devote to this now. I'm very...
85
Kevin A. Burton
sfburtonator
Feb 1, 2005 11:17 pm
... Do you serialize the hashmap? Thats the only way I can imagine this working. There's a big difference in writing an RSS library and a realworld aggregator...
86
Nick Lothian
nick_a_l
Feb 1, 2005 11:20 pm
I've done some more work on the benchmarks. The code and results are now available from http://www.mackmo.com/syndbench/ (and attached as a PDF). There's no...
87
Brad Neuberg
bradneuberg
Feb 1, 2005 11:28 pm
The question is, are most feeds small or large files, and is the extra complexity of an event based API worth it for RSS feeds? SAX makes sense for XML because...
88
Nick Lothian
nick_a_l
Feb 1, 2005 11:34 pm
... Kevin, but you have to understand ROME's design goals, which are clearly different from feedparser. ... Simple... ... still be extensible so that they can...
89
Nick Lothian
nick_a_l
Feb 1, 2005 11:34 pm
... I was looking for stats on this - I though PubSub or Syndic8 had them available but I haven't found them yet. Nick IMPORTANT: This e-mail, including any...
90
Nick Lothian
nick_a_l
Feb 2, 2005 12:13 am
ROME has to work out of the box and do everything for developers, but still be extensible so that they can provide their own implementations if they have more...
91
Kevin A. Burton
sfburtonator
Feb 2, 2005 12:17 am
... Why do you assume an event API is more complex than a DOM API? It really matters what you're trying to do. At Rojo our code that uses FeedParser is...
92
Kevin A. Burton
sfburtonator
Feb 2, 2005 12:18 am
... Also... these benchmarks might be less or more fair to either FeedParser or Rome. For example if Rome more spec compliant than the FeedParser on RSS and...
93
Kevin A. Burton
sfburtonator
Feb 2, 2005 12:22 am
... 1.8M actually ;) ... Yes... true... And serialization isn't that bad. Honestly in NewsMonster we used an in-memory RDF model that I developed. It just...
94
Nick Lothian
nick_a_l
Feb 2, 2005 12:27 am
... Ha! (No, it was an internal thing) [snip] ... No, we were talking about the f-range instance-manipulation stuff (I think!) IMPORTANT: This e-mail,...
95
Kevin A. Burton
sfburtonator
Feb 2, 2005 12:45 am
OK. Rather than just cite benchmarks I ran months ago and have misplaced I figure I'll just run the benchmarks again via the current FeedParser. We have an...
96
Nick Lothian
nick_a_l
Feb 2, 2005 1:32 am
... Knew I'd seen it somewhere: From 28,774 feeds "First issue to look at is at the size of the feeds. They show a characteristics similar to a negative...
97
patrick chanezon
chanezon
Feb 2, 2005 1:46 am
Cool. Thanks Nick. P@...
98
patrick chanezon
chanezon
Feb 2, 2005 1:50 am
Very good, just publish you benchmark code and results when you're ready. Yes, an Event API, with a Bean API on top of it makes sense. When we have the event...
99
Brad Neuberg
bradneuberg
Feb 2, 2005 1:51 am
So almost all feeds are in fact quite small. What are others interpretations of this data? Brad...
100
patrick chanezon
chanezon
Feb 2, 2005 1:51 am
... Yes we were taling about the f-range IM. Glad that you consider it in your plans Nick. I agree it looks more complicated. P@...
101
patrick chanezon
chanezon
Feb 2, 2005 1:54 am
... Thanks Nick, very well expressed. I couldn't agree more. P@...
102
patrick chanezon
chanezon
Feb 2, 2005 2:15 am
Mine is that ROME is fine for a feed aggregator today. If Kevin gets the 14 times improvement he claims with SAX it's still interesting. It depends on the app...
103
Kevin A. Burton
sfburtonator
Feb 2, 2005 2:59 am
... For the record I think its a pretty easy case to make that most feeds are < 150k or so... Kevin -- Use Rojo (RSS/Atom aggregator). Visit http://rojo.com....
104
Kevin A. Burton
sfburtonator
Feb 2, 2005 3:09 am
... I've alwasy invisioned RSS and RDF as a good solid node distribution for a graph based database which is ordered on time. Thats why a range based queries...
105
Robert Cooper
kebernet
Feb 2, 2005 3:50 am
Just to chime in here, on the "Feeds are quite small" bit... Another application that some friends and I have been batting around using RDF for that involves...
106
Nick Lothian
nick_a_l
Feb 2, 2005 4:38 am
... An intersting thing about that case is that (assuming the file is a couple of GBs or bigger and you are on a 32 bit VM) you wouldn't be able to use the...
107
Henry Story
hjsatdoc
Feb 2, 2005 9:23 am
... Cool. Another useful technology moving in the same direction. (I prefer the generality of of the library I wrote [0] or that provided by rdfreactor [1], as...
108
Charles Miller
carlfishie
Feb 2, 2005 10:22 am
... If you're willing to pub up with our mildly dodgy RSS support, I could set up a space on confluence.atlassian.com in about five minutes. C -- Charles...
109
Danny Ayers
danny_ayers
Feb 2, 2005 11:06 am
[Nice work Mark!] ... Hopefully as we get closer to Atom 1.0 there will be convergence on one predominant, use-in-the-wild model. Sometime soon I must update ...
110
patrick chanezon
chanezon
Feb 2, 2005 4:06 pm
Why dodgy? I looked at http://confluence.atlassian.com/spaces/blogrss.action?key=JAVABLOGS in NetnewsWire and it looks OK. (wearing my ROME product manager hat...
111
Marc Goodner
marcgoodner
Feb 2, 2005 4:09 pm
I submitted several proposal, unrelated to this discussion, and even today after the deadline is past and you can't even edit the submission any more the...
112
Danny Ayers
danny_ayers
Feb 2, 2005 5:51 pm
On Wed, 02 Feb 2005 17:06:22 +0100, patrick chanezon <Patrick.Chanezon@...> ... Just a thought: I don't know about TWiki but at least some of the Wikis...
