Twitter losing posts 
Flickr: “Our long national nightmare is over.”
Twitter Blocks is the kind of thing that demos well at conferences. Not too useful in real life.
John Furrier has the transcript of a Calacanis podcast chutzpah-fest at RWW. He’s a blood-thirsty samurai, so he says. “I always look at entrepreneurs as samurai. It’s a lonely pursuit at times and basically your life is to fight. And you get done with one fight. You clean the blood off your sword. You put it away. You walk 10 miles to another village. And then you got to clean up that village. A couple of people got to lose their arms. And then you clean the blood off your sword. You have a cup of tea and some rice. And then you walk to the next village.” Really, and I thought he was the sensitive type who sulks in depression for days because he was interrupted.
Economist on who’s afraid of Google.
another senior bush staffer leaves. something terrible must be coming??
dan mactough *convincingly* argues that “expanded” does not belong in opml 2.0.
apple to sell ringtones. excuse me while i yawwwwn.
sugarattack says twitter is losing posts. i noticed that too. imho they ought to use some of the $5 million to make it reliable.
I’m going to cross-post my tweets here until i notice that it’s stopped losing them.
noticed that some MSM podcasts are not bothering with the length att on enclosures. not happy about this. 😦
Proposal for ‘expanded’ attribute in OPML 2.0. Important for people developing in-browser apps that use OPML.
Dan MacTough offers a compelling argument for not including “expanded” in the core spec. (I fixed the typos, and reworded the last two sentences per another suggestion, here.)
Michael Markman on the “perfect storm” around Scoble this week.
Nathan Rein on computer-free micropodcasting. Good idea.
TechCrunch on new features for Twitter today.
Scripting News Annex 
Posted by Ron Hochsprung on August 31, 2007 at 7:21 am
Dave,
The wording of the last 2 sentences of your proposal are, IMHO, badly worded; not to mention “its” being misused.
I suggest:
The attribute is optional; if not present, the outline’s first-level sub-outlines are not visible when displayed.
Scripting News is the first blog that I check each morning…
Posted by Ron Hochsprung on August 31, 2007 at 7:30 am
Dave,
Just after sending my comment about your extended proposal, I ran across this link on The Register’s blog:
http://www.theregister.com/2007/08/31/blog_history/
Comments?
Posted by Dave Winer on August 31, 2007 at 7:42 am
Ron, I made the changes per your suggestion.
http://www.opml.org/expandedAttributeProposal
Also, about the Register bit, I know Guy from many years ago, we had dinner quite a few times when I was making software for the IBM PC and Mac in the 80s. He’s a good person, and if he says there was a blog in 1996 in Britain then I’m sure there was.
Of course there were blogs even before that and the WSJ did a pitiful job of researching that piece. But to blame it on all Americans or an American attitude is something I’d expect from some other writer at the Register, not a gentleman of Kewney’s caliber.
I suggest we pool our resources and do a careful study of the early development of blogging, and give credit where it is due.
Posted by Tom Morris on August 31, 2007 at 8:50 am
An expanded attribute sounds like a superb idea to me. There’s no straight forward way to parse the expansionState in XSLT, but it would be extremely simple to implement expanded in XSLT (as I will have to soon for the XOXO-OPML bridge work).
Posted by Dan on August 31, 2007 at 10:42 am
@Tom Morris – I feel your pain regarding XSLT.
But if you’re maintaining the node expansion state in your own application, you can always use your own namespaced attribute.
If you’re thinking more about a one-off transform (or perhaps an initial import), I’m not sure why XSLT should be expected to be able to translate what is essentially a snapshot of the node display status from one application to another and from one file format to another. XSLT can certainly preserve the expansionState information in the resulting XOXO (for example, in a comment or a hidden span). Having done that, it seems to me that XSLT has done all you should ask of it. Now it’s up to your application to figure out how to translate the display snapshot captured by the OPML Editor (for example) onto your browser. Sounds like a job for JavaScript.
By the way, is dealing with expansionState impossible with XSLT (haven’t tried) like xml:base (as far as I can tell), or is it just really tedious?
In any event, isn’t this more a shortcoming of XSLT rather than the OPML format?
Is preserving a display characteristic like the expansion state relevant or important enough to belong in the file format specification? expansionState has to stay for backwards-compatability. Would the proposed attribute further muck things up?
Posted by lemon obrien on August 31, 2007 at 11:36 am
the worst thing happening is that bush and the fed are bailing out rich people with the mortage crisis. Last month’s New Yorker had a small article about it. Most of the people loosing money have enough to loose and still be fine.
I’m really tired of being an American and getting fucked over by my government.
as for everyone leaving…its bad, consider the fact, most of the heads of our government could be prosecuted as war criminals under international law. Now consider, they’ve done nothing except rob, steal, and manipulate, for greed.
sorry for the rant…but i’m ready for the revolution; maybe i’ll lead it 🙂
Posted by Amyloo on August 31, 2007 at 12:04 pm
I think it has to mean that Snow doesn’t want to defend the White House story of the progress report.
Posted by Tom Morris on August 31, 2007 at 12:38 pm
I’ve been experimenting with XSLTing this evening. I’ve managed to take an OPML file that has an expansionState element (in fact, an outline straight out of the OPML Editor) and convert it to the proposed ‘expanded’ attribute by using XSLT 2.0’s xsl:number function. I’d like to do it in XSLT 1.0 – I may send it off to the folks on xsl-list to see if they can come up with something that works with XSLT 1.0.
I’m not sure about the implementation of ‘expanded’ (I like how XOXO does it with the compact attribute, but that’s pretty much personal preference), but I do think that it may as well go in. OPML already conveys this information already. There are numerous contexts where putting it in is worthwhile, and it’s a definitive improvement over expansionState.
Posted by Dave Winer on August 31, 2007 at 2:18 pm
The fact that you’ve been able to interpret expansionState is all the proof I needed that the attribute isn’t needed. I’m going to leave it open for another week or so, but I don’t, at this point, plan to include it in 2.0.
Posted by Tom Morris on September 1, 2007 at 4:58 am
I was under the illusion yesterday that xsl:number is an XSLT 2.0 element (I found it in the XSLT 2.0 reference, and didn’t look back in the 1.0 reference).
I’m most of the way through writing an open source OPML-to-XOXO parser in XSLT 1.0. It will support all valid OPML (and most valid XML that gets close to the OPML spec).
Posted by Joshua Allen [msft] on September 1, 2007 at 2:44 pm
Yes, expansionstate with XSLT is not the easiest thing, but it’s possible. I was tempted in the past to add my own “expanded” attribute in my own namespace, but ultimately decided against it, since 100% of other OMPL processers in the world would not be able to read it, and I would have to support expansionstate anyway for compat.
Posted by Jeff Imig on September 2, 2007 at 10:54 am
re the Long National Nightmare…
I don’t believe for a second KR is just crawling off to write his memoirs. It seems more likely he’s just being freed up so he can work full time for the RNC. Or some organization like that, ramping up for the next election.
Please prove me wrong!
Thanks, Jeff
Posted by Jeff Imig on September 2, 2007 at 10:55 am
jeez… it’s nocomputers, not nocomputer in the website link.