thanks kris for the immediate testing. it's right, rss items without a "link" node or an "about" attribute won't make it into the aggregator (yet?). imho it just does not make any sense to have stories without a uri. that's the way the internet works. dunno why of all rss generators radio userland creates such meaningless output. but maybe this is just a set-up thing, because adam curry's rss feed (made with ru8) is looking good.

i even get an error in the browser just trying to display ronsen's feed. probably due to the low forgiveness of xml...

features are yet here to come. posting the whole story is a strange thing, because one might loose track of where the content is belonging to. still evaluating...

i don't want to post the whole story, i only want the whole story in the edit window to cut out the interesting bits.

well, i checked the rss output with the w3c rdf validator and it looks fine. however, i would be interested whether there are any differences between this particular feed and one you can read with your peerkat system.

channels always include an items section and textinputs have an attribute: "rdf:about"

i have mocked up the parser (put in a dummy attribute rdf:about), the parser (sax) now renders ok, but does not output, because of missing items section

just did a test on feedreader everything seems fine, i will look into peerkat/orchard parser source for a hint now

the funny thing is that i removed the items section just recently because the w3c validator returned a "Warning: {W101} Unqualified use of rdf:resource has been deprecated" message... (check it out yourself with this example rss10withItemsRes (text/xml, 6 KB) ).

i am a little bit confused about this. if it won't break anything on the other end, i happily will include the items section in the channel node. the about attribute in the textinput node should not be a problem, anyway. would be thankful for some more comments and evaluations, though.

you get your warning here because of your use of resource instead of rdf:resource

see a list of common validation errors from schematron at groups.yahoo.com

yes, you're totally right... i did not spotted this mistake but it's there. i mean: it was. the item resources got back into the feed and i also added the about attribute to textinput. i am curious what peerkat will say now. btw. thanks a lot for your help!

peerkat says

Feeding pop-log into path /poplog pop.antville.org

1/7 feeder: 1016995996.0 Regular Feed Completed: 3 good / 0 bad feeds, 1 new / 24 total items

Next feeding time at 2002-03-24 09:53PM

thanks alot!

btw: there might be a glitch, all items are recognized, but only the first item is taken - i look into it...

and i guess other extended characters are not recognized

i think, extended markup must be escaped or go into a cdata section ???

did you check the rss-spec: title, url, link, description all have a Model: (#PCDATA)

from the spec: Encoding While RSS 0.9 supported only ASCII encoding, RSS 1.0 assumes UTF-8. Using US-ASCII (i.e. encoding all characters over 127 as &#nnn;) is conformant with UTF-8 (and ISO-8859-1, HTTP's default header encoding).

i dont know, if you want to fix this, because noone else does - sorry!

btw: channel.description is not an empty element

i'd better be quiet now :)

cool, i am happy that it the antville feeds are working now with peerkat. in the case of the encoding i first have to check how other feed readers cope with PCDATA. would not be a big issue to implement that, then.

let me know what it's about only taking the first item (it says "1 new / 24 total items", is that a hint?).

i also will check if i can leave out empty elements – although i do not understand why that should not be allowed (just because it says so in the specs?)...

i think, peerkat is too strict, here. other feedreaders are much more gentle, maybe only because they dont know better? i'm waiting for comment on PCDATA from schematron support...

in fact, peerkat excludes feed items, if there are eg. german umlauts in them. i will rewrite the items to a &#nnn; notation and have a look then

update: rewrote feeditems to the &#nnn; notation feedimport works fine now

does that mean the use of encoding="iso-8859-1" in the xml header is not sufficient for allowing umlauts and stuff?

what is it good for, then, when i have to encode special characters using &#nnn notation, anyway?

confusing.

will contact the rss-dev group now

seems, they insist on the spec you may lookup

• groups.yahoo.com

do you know of any german rss1.0 feeds to check?

popo.at is rss10 and renders ok funny thing, anyway: author is "piefke 3000"

there is html markup in description element and there are character entities and characters above 127 are escaped to &#nnn;

some deep knowledge over there might be...

this is slightly embarrassing for me because i know this piefke 3000 guy pretty well... :)

truth is, i really assumed that this annoying and time-consuming encoding of special chars in xml is obsolete when the right encoding is used in the xml header.

seems that assumption was too daring. so whatcha say if i simply would encode the antville feeds, too?

would be a really smart move :) thanks alot...

thanks hannes, i'm glad that you like it. and i agree that the origins of the items have to be accessible more easily. you opened my eyes on this issue anyway. so please stand-by a little, the improvements are here to come...

[darn, i just noticed that i posted the following text to the wrong thread... supposed to be an answer to kris' request for posting the whole item description to another weblog. was that due to the misleading conception of too many "post your answer" links here, again? dunno.]

i would like to know what the others think about this... should one really be able to post the whole text to another weblog? even given that there is a link to the item's origin as proposed by hannes?

I personally think the post feature is extremely cool! The steps from having the bare link to posting it on one's weblog are cumbersome enough to want it. I think both "link" and "post" are really useful and should both be in.

[sorry hannes. it was not my intention to discuss the link feature, i posted to the wrong thread!]

anyway, the post feature is already there (sounds like this was not made clear enough)... when you click on the word POST on the right of each item an abstract of the item immediately will be transferred to rss.antville.org.

the question rather is whether just an abstract – containing the item title (as far as available) and a link to the source weblog story – (as implemented right now) or the whole text as it appears in the original weblog story should be available for posting it to another (e.g. one's personal) weblog.

usually, the abstract itself doesn't make a good post, whereas some sentence from the whole story might be a good teaser line. as is said before, the whole story would be better. just because it appears in the edit window doesn't mean people do actually post the whole story. and if they wanted to steal content, they could have done it by copy and paste anyway.