Unacceptable Microsoft bugs
Don Box touts himself, Sam, and Clemens as being “probably the only three live CDF feeds on the net”, and claims that “CDF is still supported in IE (including Longhorn!) and has been stable…”.
Well, in fact it’s Clemens’ tool, dasBlog, that’s providing his feed, and googling “cdf ashx” seems to say there’s a couple hundred live CDF feeds around. However, the state of CDF support has shockingly declined: I click a link to a CDF file in Firefox, it asks what to do with it, offering the default of opening it with “ChannelFile”, but when I do that, and IE’s Favorites manager opens, lets me save it as a bookmark, and then opens IE with the bookmarks open in the sidebar, ready for me to browse items from the channel, I find that if I click another CDF link in Firefox, it adds another IE bookmark folder, but the items in the first one are overwritten by the items from the second. This sort of sloppy handling of an old, nearly abandoned format simply can’t be tolerated! (Unless it because of something Firefox is doing, in which case it’s all Windows’ fault).
It’s been said a few times before, but it bears repeating: in a lot of ways, CDF rocks. If there was a cross-browser implementation, with auto-updating bookmarks that would let me easily navigate through the unread things, and if everyone had nice clean lightweight individual entry archive pages, I’d give up a separate aggregator in a second. Cross-browser, easy navigation, and a spec with a few less typos, and that’s all I want.
And speaking of the edges of the ”who invented what” permathread, even though I’ve looked at it lots of times before, I’d never really noticed just how awful the RDF of RSS 0.90 actually was. Of course it was early days, and they had plans to make things more complicated, er, more correct, later, but all those resources, treated as useless strings? RDF in the sense that it wouldn’t choke an RDF parser, but not much more.
Why not make templates for a few of the popular blogging packages, and spur demand from that end? IE is not a small market to target.
http://diveintomark.org/about/templates/index/cdf
$ lynx -head -dump http://diveintomark.org/about/templates/index/cdf | grep Last-Modified
Last-Modified: Wed, 20 Aug 2003 05:42:21 GMT
So I’v ever really been interested in CDF before, it only being implemented in IE and all, but, reading the spec, it turns out that, with a little bit of cleaning up and a few tweaks, it could be my ideal syndication format.
Do you think I’ll get away with stealing CDF, making a few changes, and claming it was all my own work? ;)
Just replace some item names, call it Atom/NG and everything will be fine ;-)
No, call it RSS/NG. After all, RSS was never more than CDF with different tag names.
Because I don’t want to be left off the syndication bandwagon, and also because I find it repugnant that a weblog based on Microsoft technology might have more features than the software I have arbitrarily chosen and have been too lazy to move off of (6A RULEZ!!!!!!!!), I humbly (this is me being humble, you should see the other side of me) present:
http://diveintomark.org/xml/cdf.xml
All other feeds are deprecated and will go 410 within 48 hours.
END THE SYNDICATION WARS! ALL HAIL CDF!
Feed Parser 3.0 beta 20 fully supports CDF.
Wasn’t XML case-sensitive back then? All the MSDN articles and the example templates I’ve seen use all-caps for everything, but the Note uses StudlyCaps for element names and all-caps for attribute names. I trust Feed Parser handles both?
Most XML formats are case-sensitive, except RSS, which freely alternates between camelCase and StudlyCaps, and OPML, which has no rules whatsoever.
The MSDN documentation on CDF as implemented by IE4 uses all-caps for element names, and child content for values. The draft submitted to the W3C is apparently an earlier draft, and uses StudlyCaps for elements and VALUE=”foo” attributes for values. Oddly, both use all-caps for attribute names.
All the CDF I’ve seen in the wild matches the MSDN documentation, thus that is what FeedParser supports, although it accidentally supports the StudlyCaps version as well. If there is a groundswell of support for the VALUE=”foo” syntax found in the earlier draft format (and knowing this community, I wouldn’t rule it out), I can add that as well.
Did I mention this shit is unit-tested? Which is more that you can say for IE, apparently.
I’m sorry, that should read ”except RSS, which freely alternates between camelCase and all-lower, and OPML, which has no rules whatsoever”.
Off-topic: did you know that when Userland stole RSS from Netscape, one of their several incompatible changes to RSS 0.91 was to make channel/image required? Also, they dropped the optional image/description element. Compare: Netscape channel vs. Userland channel, Netscape image vs. Userland image.
Mark, why didn’t you provide feedback on this during the open comment period? And was it really ”stolen” — could you explain that. To me it looks like UserLand was co-author of RSS 0.91, and all Dave did was write a spec, which it seems you or anyone else could have done. Isn’t that what Rael Dornfest did? Do you think he stole something? I think it’s pretty transparent that you’re a bitter person, and you’re trying to drag Dave’s reputation through the mud. Maybe it’s time for you to stop. It’s pretty obvious what you’re doing.
My understanding of copyright law is that things are copyrighted unless copyright is explicitly disavowed. If encouraging someone to include some of your stuff in their stuff gives you all rights to it as soon as they stop updating it, then Six Apart better not ever stop updating Movable Type, because I’ll declare it abandoned, and take it over.
Chances are it was the right thing to do for RSS, at least for someone to grab it and run when Netscape got bored with it, but it’s not as simple as it would be if Mark and I cowrote DEG, with every page saying ”Copyright Mark Pilgrim and Phil Ringnalda” and then he got bored with it and I moved it to my server.
Ack, I didn’t even notice the value vs. content diff.
There’s something severely broken in the way we go about specifying and comparing things (see also: skip hour 0), but I’m not sure what we should be doing instead. I’m always tempted to get a piece of paper and a pencil, and draw a line down the center, which can’t be the right approach.
Yesterday, I was reading Keith Robinson talking about using personas to help you decide what not to publish on your weblog, and I thought ”what on earth should you not publish?”
My reader persona’s first characteristic is going to be a very strong tendency to go overboard.