RSS Problems August 22, 2002 4:32 PM   Subscribe

Does anyone else use the rss feed? I've tried to use it with a news ticker with little success. [more]
posted by swift to Bugs at 4:32 PM (14 comments total)

A news ticker shows headlines, and since a typical entry in the feed looks something like:

[item]

[title]posted by trioperative August 22 8:31 AM | 38 comments[/title]

[description][![CDATA[ dumpster sticker says, [a href="http://www.theksbwchannel.com/news/1620890/detail.html">Don't throw that baby away!. Christine McGuire of the Santa Cruz county District Attorney's Office thinks it's a good idea.]]] [/description]

[link]http://www.metafilter.com/mefi/19379[/link]

[/item]


the news ticker shows the "posted by trioperative August 22 8:31 AM | 38 comments" part as the headline and nothing else.


How is the feed meant to be used?
posted by swift at 4:34 PM on August 22, 2002


Here's the deal: RSS sucks in general for weblogs.

RSS was designed (IMO) as a stricter format than weblogs frequently follow. It follows more along the lines of news stories. Not every blog post has a title and a short description like a newspaper article might. The link in RSS feeds is typically used for a single link being discussed, but many MetaFilter posts have multiple ones.

Since there are character limits to RSS, and that I couldn't extract single links from posts, my next problem is trying to determine where I should put the info I do have. In general, nothing is predictable about metafilter posts, they may start with links that can act as titles, but often they do not. About the only short piece of info I do have is who posted it and how many comments there are, so those became the titles. I'm also not going to add a title field to the posting page, just to satisfy the RSS spec, I'd rather a spec support a wider range of options.

People complained about the uselessness of reading an entire post's payload without html in it (which again, I couldn't break out as the post's link, since there could be more than one), so I decided to just dump the entire HTML of a post in the description area, limits to descriptions be damned. Many bleeding edge readers can parse the feed fine, in effect allowing remote copies of the front page's contents to be displayed elsewhere -- true syndication.

Keep in mind that these problems existed in 1999, and still, we've had little improvements to RSS. In general, we're still trying to stuff weblogs into the template made for newspapers.

On top of all this, since the site relies on user-submitted data, it is often "dirty" containing characters hostile to XML (as seen by the boxes and question marks around the site). This often breaks RSS feeds, and there's not a whole lot I can do.
posted by mathowie (staff) at 4:44 PM on August 22, 2002


My friend, a web developer, says you're just lazy.
posted by swift at 5:25 PM on August 23, 2002


My friend, a web developer, says you're just lazy.

You are joking, aren't you?
posted by timeistight at 6:42 PM on August 23, 2002


No, he really said that. And I agree. Matt's just lazy.
posted by swift at 11:12 PM on August 23, 2002


What would you have Matt do? Should he go in and personally assign a title to each post and clean the data so that your news ticker works?

Why don't you just copy all of the text from the site into your web site, and you can clean and normalize it and then publish it as an RSS feed? Let us know when you have that working.
posted by willnot at 11:56 PM on August 23, 2002


It could be probably be fixed with a script. I'm no good at scripting, but I bet Matt is. He could do it if he felt like it.

What's the RSS feed for, if it doesn't work right?
posted by swift at 5:19 AM on August 24, 2002


the problem isn't that it could be fixed if someone could be bothered, but that it's not clear what the solution is. no matter how many scripts you have, you can't change logic - if the rss format is designed in a way that doesn't map to what's on the site, you're in a mess.

(i'm a software engineer and i'm normally happy for lusers to think i'm capable of anything, if only they paid me enough (just how much are you paying matt, swift?), but even we have limits....)
posted by andrew cooke at 5:47 AM on August 24, 2002


I'm no software engineer, and this problem is apparently in the realm of impossible rocket science, but it seems like ripping out the first seven or so words of a post and using them as the rss title wouldn't be all that difficult. It's better than "posted by user x."

However, like you say, since I haven't contributed any money to Metafilter, I shouldn't expect anything to work.
posted by swift at 6:19 AM on August 24, 2002


What's the RSS feed for, if it doesn't work right?

It's mostly there so that Matt can use MacReporter.

Lots more information about MetaTalke and RSS here.
posted by willnot at 8:15 AM on August 24, 2002


Now it's all clear. Thanks, willnot.
posted by swift at 8:53 AM on August 24, 2002


NetNewsWire is much more fantastic.
posted by rhyax at 11:52 AM on August 24, 2002


swift, yes I could grab the first few words from each post, but still the restrictive format of RSS sucks for weblogs. I've considered this before but "So the Al Queda allegedly..." isn't all that informative. If your reader supports HTML rendering, you can see the entire post in the description field. This is a better solution in my opinion, because if I were to grab the first seven words of a post, and extract the first offsite link for the link field, it would be less useful than having the entire post at your disposal.

As to my laziness or lack of willingness to work on this "The reason we are good programmers is because, by and large, we are lazy." All programmers are lazy, that's why they started programming in the first place, because they were too lazy to do something by hand.
posted by mathowie (staff) at 12:47 PM on August 24, 2002


Yeah, you're lazy Matt. That's why you haven't got us our ponies yet.
posted by timeistight at 1:07 PM on August 24, 2002


« Older Comment problems   |   Why is it not ok to post repeats of old posts? Newer »

You are not logged in, either login or create an account to post comments