Saturday, May 31, 2008


So I've mentioned that I wanted to change to a three column layout here. For some reason I thought it was going to be difficult. Before trying to dive in and do it myself, and spending the next several hours cursing css (or, at least, my css abilities), I asked Google one more time. One of the first results was this page, which got me going with no trouble at all. So thanks guys.

Anyway, if you've got strong feelings for or against the current layout, let me know. I don't expect I'll hear from anybody.

Friday, May 30, 2008

World Record Layout

I updated the layout for this blog recently. Not much, really, I just added two new bits along the right side:
  1. Download Day. The folks at Firefox are trying to set a world record 'for most software downloads in 24 hours'. They've got it set up so you can pledge to download on the day Firefox 3 comes out (which, interestingly enough, is not yet determined, just "sometime in June"), so I put a link there on the right. If you're reading this in a feed reader, you can just follow the link here.
  2. Flickr Slideshow. To play around with new (for me anyway) things (or just avoid doing work), I signed up over at flickr (so now I'm on yahoo too, I guess), and posted a few photos. Those now show up in a slideshow along the right panel. There isn't much up there yet, but hopefully I'll remedy that... sometime.
I'd still like to spend some time trying to make a slightly more substantial change to the layout here. I've been thinking I want the 'blog archive' and 'labels' bits on the left (maybe 'about me' as well), as they are part of this blog, and leave the other bits (other random bits of my online life) on the right. None of the provided blogger templates are 3 panels though, and I haven't found one online. But I need css practice anyway, right?

Thursday, May 29, 2008


Sometime in April I started wondering about giving myself a haircut. Everybody "knows" it's a bad idea, so I was curious to see just how poorly it would go. I waited until the end of the semester, and then went for it. Since I was expecting disaster, I thought I'd document the process with some photos. It actually went better than expected, so I didn't bother much with the pictures. But last night I was going through some other pictures, ran across these ones, and got a chuckle out of them. So I thought you might as well.

Here's one before getting started:

And then I brushed it up, generally good for a laugh:

I'm vaguely reminded of Eraserhead.

After these two, I stood over the sink with a pair of scissors, grabbing bits of hair and chopping off those that felt long. Here's the result:

Probably a photo from the back would be a bit more like what you'd expect of giving yourself a haircut... but I'm only so flexible. I guess the trick with giving yourself a haircut is: if you cut everything short enough, you can't really go wrong.

Wednesday, May 28, 2008

Video of my Cat

I took this video a while, of my cat Kramer playing with a new catnip toy. He's generally pretty reserved.

Tuesday, May 27, 2008

OpenID Revisit

A while ago I wrote a joyful post about OpenID and how I thought it was great. Admittedly, since then I have hardly used my OpenID identity (any of them). I've remained positive about the idea, and keep hoping I'll find/make the time this summer to do enough webpage building sorts of things to get to a point where I would accept OpenID login.

So I've been brought down a bit by the article "The problem(s) with OpenID". It's a little lengthy, and apparently largely a wrap-up of many other posts from many other places. Still, it seems like a fairly important article to read.

Also today I listened to a podcast from April concerning DataPortability with guest Jonathan Vanasco, and found it to be pretty interesting. In particular, it made me think more about the idea of having many different faces online, e.g., something like a MySpace versus LinkedIn account. And, as coincidences go, just yesterday Mr. Vanasco wrote a post: "Data Sportability". All quite interesting, and I'm looking forward to the promised 'upcoming' posts.

I wish I was more thoughtful about security and privacy and things. The web sure is an exciting place. Articles like these tend to make me wonder what I should be doing with my life. There are smarter people than me, that work harder than me, all over the place, getting all sorts of amazing things done. What sort of contribution can I actually make? Where should I direct my energy, so that a contribution actually can be made?

Monday, May 26, 2008

Feed Update

I just decided to try out FeedBurner for this blog (new feed here). I'm not sure how that works for current subscribers - will you even see this message? Anyway, if you get a minute, perhaps update to the new feed.

I decided to do this, because according to the Feedjit widget here, I'm actually getting hits from... well... around (I know, plenty of them are me, seeing how the page looks. But I know I'm not all of them). So I thought I'd get some more information about that, and FeedBurner seemed like a common way to do so.

Please bear with me.

Beach Week Recap

So it turns out (I'm not sure I'm surprised) that I have a little in common with whales on the beach - it's quite clearly not my natural environment.

I made it 3 days without checking my rss reader, and for the remainder of the time I didn't read items, just starred a bunch. So today, catching up, has been pretty taxing. Luckily it was a holiday, so a fairly slow news day. I'm back down to a reasonable 20ish starred items. I don't think I'll try for any more rss-free days any time soon.

Couple of other things from the beach.
  • Running on the beach really isn't that exciting. The scenery never changes really (at least, the 5 mile stretch I was on), and it was windy (could have been a fluke, I don't know much about the beach) and sand blowing in your face isn't fun while running.
  • It occurred to me at some point during the week that though I expected the expression 'give props' was probably short for something, I had no idea what. I asked my friends (5 of us there), and none of us knew. Our librarian friend looked it up on WordNet (which was cool, I'd just started hearing about WordNet). Turns out its 'proper respect'. Makes sense.
  • We played some fun games while we were there (besides Dr. Mario, which did get brought out - what sort of nerd brings a Nintendo to the beach?). Settlers of Catan is a fun game - except I almost always get frustrated while playing. Eric had a game 'Pentago' that was also pretty cool. It's connect 5 on a 6x6 board, but after each turn you get to twist one corner of the board. We also played some Farkle.
  • I started wondering about perhaps making a blog for the calc 2 class I'm teaching in the fall. Trying to decide what it would include, and if I feel like worrying about privacy concerns and all of that. I guess UVA has some system in place for class collaboration, perhaps I should start there.
So what did my (not so) unconnected week miss?
  • I did, actually, get my Twine invite, which I was surprised by. Perhaps because I mentioned it here? I know that when I signed up for the invite, they asked for my webpage... are they monitoring to decide who gets an invite? Or was it just a coincidence? I haven't yet had a chance to play with it, but am looking forward to doing so from here (well, probably tomorrow) on out.
  • The XOXO (next version of OLPC laptop) looks like it could be cool (see here and here). I'm going to wait a while before I get too excited about it - see what it actually turns out to be. The $75 price mark was shockingly low, given the current version is at $188. Time will tell I guess. Also they're doing G1G1 again later this year (with current version laptops). I certainly don't need another, and I really, really hope they have their distribution issues worked out. Oh yeah, and I took mine to the beach as an excuse to test it in direct sunlight (like we don't get that in cville), and it worked beautifully (as expected).
  • I listened to a Semantic Web Gang podcast while I was away, with Barney Pell from Powerset. Got me excited about Powerset, and some of the people had pretty cool accents. So I downloaded a bunch more of the semantic web sorts of podcasts I've been avoiding. I still haven't gotten around to them.
  • Read a pretty interesting article from Newsweek - 'The Rise of the Rest'. Worth a look. Made me wonder what I should be doing to better prepare for global interaction. Any suggestions?
  • Found (thanks 'Discover' button in Google Reader) a cville blog I actually subscribed to, because it's about local tech. And it turns out I missed another interesting talk here at UVA. Sigh. Hopefully this is the last one I miss because I don't hear about it until after the fact.
Probably the highlight of the news I missed was the Roku Netflix set-top box (see here and here). I had basically forgotten about that news from January, but I'm kinda excited about it again. Sounds like its a really easy to use piece of hardware, in a reasonable price range... I'm debating about it. People have complained about the 'small' number of 'Watch Instantly' titles from Netflix, but I'm sure I could find enough to keep me occupied (like I need another distraction). If the Roku box was able to play videos from anywhere online (maybe embed a link in your Netflix queue?), I would have bought one already. Perhaps one of the other boxes that is set to come out later this year?

Sunday, May 18, 2008

Beach Week

I'm headed to the beach this week (in South Carolina, if you care). I'll be taking my computer(s) (that's right, I just might use my XO to read some pdfs on the beach, or take pictures?), but I'm debating about taking the week off from keeping up with my feeds. I also don't expect I'll blog or tweet (don't fret, I'll be back before too long). I wonder if I'll actually stick to 'no feed reading', how many items I'll have at the end of the week to catch up on, and how thorough I'll be in doing so. People say unplugging is supposed to be good sometimes, and I'm thinking I might test that. I just wonder how much reading I'll have to do at the other end to make up for it.

I'm wondering what I'll miss this week
  • XO-2 gets unveiled Tuesday. XP or no, perhaps it'll have some new cool hardware.
  • MicroHoo back in the news
  • All the recent fun about dataportability wrt MySpace, Facebook, and Google. Maybe somebody (my guess is not one of these three) new will come along this week and actually do dataportability, instead of just social network interactivity.
  • Something fun and exciting in the semantic web sphere? Maybe I'll get my twine beta invite (I'm not counting on it).
And what will get done, with all this time free from reading rss feeds?
  • Notes for the calc class I'm teaching in the fall?
  • Editing Lie groups notes from this semester?
  • Progress on research?
  • Sunburns? More blisters from running? Learning something about my friends?
Have a good week. Look for a stream of shared items (if you are following those. You are certainly welcome to.) and perhaps r?ttd's next weekend.

Thursday, May 15, 2008

Global Comments

I'm part of the problem. Minutes ago, I decided to 'Share with note' an item from Google Reader, instead of posting a comment on the site hosting the original article. In my defense, I didn't have a huge amount to say, I'd be surprised to think it was truly valuable, and I would have had to register at the site in order to leave a comment (instead of allowing, say, logging in via openid, in which case I would have). I want to be more a part of the online community, leaving comments on posts that strike me... but at the same time I am fairly lazy. So instead of contributing to the community, I contributed to the problem: comment tracking.

By not posting my comment somewhere associated with the original article, I have essentially forked the conversation. Somebody trying to keep up with the commentary the article generates doesn't know what I said (not a true loss, in the case, but the principle is there).

I was wondering a little about comment aggregation recently, and systems to keep all comments about an article (or links back to articles directly inspired by an article). I hope they develop more in the near future, and I vaguely wonder what, if any, will be the effect on sites like digg or slashdot (more on this below).

[Warning: this post gets a little rambly about now. If you've got better things to do, don't let me keep you.]

Another aspect of this commenting issue that has been on mind, as of late, is the one I mentioned in a recent post: there are too many venues to publish thoughts. I had something I wanted to say about the article I mentioned at this start of this article. I could have left a comment on the original page. I could have thought about what I wanted to say more and made a post here (and put a link in the comments on the original page). I could have left a tweet pointing to the article (but with 140 characters, some occupied by the obscure tinyurl, could I have left a tweet that would have actually inspired people to take a gamble at a url, for which they had no idea where it actually pointed?). I could have just 'shared' the item in my reader. All of these different venues are hosted by different services, and read by entirely different audiences.

What I dream about is a global interface. I don't want to have to post the same thing many times so that 'my audience' sees it. I want to post in one place, and then if I actually have an audience, have the system automatically send them a message (or wait for an rss request) that says something has been posted. And I can pick a subset of my audience for particular messages, and global reading permissions (I could choose, in fact, just a single person, with only that person allowed to read it, and have just eliminated my need for email/IM). Except my audience has the option of just being an audience when I talk about a particular topic (e.g., they could filter my posts down to those tagged with 'math' (future versions of this system handle the tagging for me, as the semantic web and natural language processing makes better progress)). And when I tell the system that my post was related to, or inspired by, some article I found, the system goes to that article and lets it know it inspired me, with a link to what I had to say.

Again, but from a reader's perspective: Instead of finding a blog I want to follow, and noticing that they also twitter, and... wherever else they post, I can just click 'follow', and get their one global feed (tailored the way they want for privacy). Hopefully this also limits duplication from people whose posts get published in a few locations. And then I notice that a particular source has a hard time staying on track, and talks about a diverse collection of things. I've only got so much time in the day and I don't care about the person's fish recipes (or whatever nonsense), so I can set it up to filter incoming messages/feeds from them based on topics/tags the interest me. Better yet, the system sorts all of my incoming traffic based on my APML data and the audience of the message (if I'm the only intended recipient, and I ignore it... that's not very polite), and soon starts tossing out things I never read (or at least dropping them in some folder I rarely check). And when I do find somebody new to follow, my system asks them for recommendations, compares against my current subscriptions and interests, and leads me to new interesting material. Instead of choosing subcategories of digg to follow, my system finds pages directly.

And we all live happily ever after...

Wednesday, May 14, 2008

Lockhart's Lament

I just finished reading "Lockhart's Lament" [pdf] which seems to have been posted as part of a monthly column at the MAA (if the page had an rss feed...). It was a great article about mathematics education, or lack thereof. I very much encourage you to read it. There is also, I just noticed, a followup, but I've not read that yet.

I hope I can do something, in my own classes, to address the issue brought up by the paper. Hopefully I remember to re-read this paper several times while preparing classes, and let if affect my course. I think I may also encourage my students to read it, especially if I can get my course to strike them as 'not a typical math course' (or even just certain class periods). Perhaps they'll be more on-board with atypical exercises if they have some idea where I'm coming from. Of course, his essay was geared toward K-12 classes, and I am (and hope to continue) teaching college classes, but still.

I also found the article inspiring on the 'math is awesome' level, which I've been lacking a bit as of late. Not making progress on research is frustrating, but math really is awesome. Reading 'A Mathematical Journey' by Schumer has also been helpful. I was thinking last night that this book is probably one I can return to throughout my life, and always (re?)discover wonderful things.

Math Article

I found this pdf for the Newsletter of the European Mathematical Society, which had a pretty interesting article "A Visual Approach to Calculus Problems", by Tom Apostol. As it's visual, and I don't really have the urge to generate pictures, any description I give will be lacking, so I'll just let you check out the article if you are interested. It begins on page 17 (19 of the pdf).

Tuesday, May 13, 2008

R?ttd #2

  1. I read this howstuffworks page on credit reports today. Anybody get their annual free credit report? Does it help?
  2. I was wondering about a feed reader with keyboard shortcuts that let me rate items as I passed them. Something like 'don't care', 'vaguely promising, but skipped', 'read it', 'enjoyed it', 'loved it' (actually, just 4 options would be better in terms of easy keyboard usage). The reader could then see which of their feeds were the most popular, and help get rid of ones they don't really care about. Anybody know of such a thing? Perhaps a greasemonkey script to run on top of google reader?

Monday, May 12, 2008

R?ttd #1

I'm debating about starting a series here: Random questions throughout the day (see title). Instead of forgetting the questions I was going to ask throughout the day, or post them on twitter as they occur to me, I think I may try to gather them up here. I'm not sure why. Perhaps it's so I feel like I'm actually posting things here with some sort of regularity (like that's important). Perhaps I just don't want to be limited to 140 characters (although you might like it better if I were). Perhaps I'm looking for (and successfully finding) more ways to avoid work. Or I want to express questions, but not take the time to write up a full post about any of them in particular? Anyway, here's today's (issue 1) of r?ttd:
  1. I think I'm missing something with XHTML+CSS. When I think about it, my xhtml is still too focused on the presentation. I have to put a divs or spans around things so I can place them how I want them to show up on the page, with extra divs for spacing and other decoration. But I thought the point was to separate display from content? Move display handling over to the css side? But if I don't have enough divs in place to position and decorate my page, how do I make it look like I want? I'm sure an example would help this explanation, but I should be doing real work anyway.
  2. In this article, the author states that "Programmers understand hierarchies, but many users don't think that abstractly". This is removed from the point of the article, but it still stood out. Seriously? Users don't understand hierarchies? Sigh. I read that article and also this one in the same day, which seems about as noteworthy as anything else I talk about. Anyway, I found both interesting.
  3. Am I ever going to find time to really learn about all the acronyms floating around? Should I post a list of the ones I'm most interested in learning on this blog? One of my sidebar items? Would anybody notice? Would anybody care? (If you're still reading this far into the post, I guess you might notice)
  4. Am I ever going to _do_ anything? Should I try to trim down the amount of reading I do, and instead actually go "out" (while staying indoors) and try to build something? Or "out" (wherever people go) and meet people? Anybody have any tips for good ways to trim down your feed reader list?
  5. Heading to the beach next week. Should I go without a computer? Don't people say unplugging is good sometimes? How long will it take me to catch back up on my rss feeds when I get back?
  6. This article (pdf) talks about engaging students. It seems to be mostly talking about younger students. Does it apply in the college setting? How should I engage my math students? Not entirely related, but this article talks about scientists opening up their notebooks to the world, before publishing, to expand science itself. How does it apply to mathematics? How about to Ph.D. students? Should I be posting here about the things I'm doing in my research? Would somebody come along and finish my work before I get a degree? If somebody posted a helpful comment on some key point I was stuck on, would I still be granted a degree? Was I ever going to get one anyway?
  7. Do I (do you?) have a romanticized view of fast food? Like... the people who eat there are busy getting things done, and don't have time to stop and prepare some actual meal? And, gosh, don't I wish that was me? Well, even if I'm not that busy, and never getting anything done, I can pretend by going to taco bell at night and getting some mt dew (to keep me up, so I can get more done?) and tacos?
  8. Did you make it this far through the post? Kudos.

Thursday, May 8, 2008

The Problem with Posting

There are too many venues, that's the problem. I have no audience, and still 'publish' in 3 places: blogger, twitter, and I'll go ahead and consider my reader shared items. Of course, with the recent updates to reader shared items, what is the point in posting to twitter anymore (was there every a point?)? I can post random notes to my shared items without any associated link... that's what I do on twitter (I know, I know, you can't follow anybody in reader... unless you sign up for their shared feeds). On the flip side, I can post links on twitter, but why bother if they are shared through reader? And I don't do any real amount of multimedia (pictures, video) posting, but I can't imagine things would be better if I did. And this is just my own individual decentralized posting. What about the people I follow, the random networks they're all on, and services they use. We're all on different IM/blog/microblog/social networks, and it's a mess.

Everyone (e.g., here's two) is abuzz today (I guess yesterday now, sorry) about myspace joining dataportability. And it is good news. I think it's only a first step (one of many first steps being made), which is clearly valuable. Nothing will apparently come out for a few weeks, but I'd still like to think about what's next (since I have been anyway). Perhaps something like the following:

From my main page I can post new material, manage my friend networks, maintain my profile, manage access to my material and profile, and read all of the material people have sent me. Let's start with friend networks. I've set up lots of little networks for myself (instead of signing up to whichever ones online) - some friends are in several of the networks, some are people I've never met (imported from my twitter/blog followers). Some are people I'll only communicate with via email, others are more instant message sorts of friends. Some are people that I'll never hear from, but who've decided to read what I've got to say. Others are the opposite, people I read but will probably never communicate to. The university has set up a network for the class I'm teaching, as well as a network for the faculty and staff. I've got lots of information about some of these people (close friends, online family), and little more than a username for plenty more (blog followers). I've got all of my contacts, at all levels, accessible in one place.

When I want to post something new, I distinguish it as a noted link, or a global tweet, or a local tweet, or a blog post, or a geophysical post (or...?), or I pick an individual out to send a message (email, IM? same thing). So my random twitter followers only get to see my global tweets (and noted links, or links to new blog posts I make, if I want to allow that), while my closer friends might see my local tweets (as well as global ones). Generally I put random 'what I'm doing' tweets in the local bit, and random 'interesting thought' (to me) tweets in the global bit. I've also set things up so that the world can see my blog posts and global tweets (though if I wanted to make an exception for any individual post, it'd be easy to do). My geophysical posts go out to people physically nearby (I've set it up to broadcast to listeners in c'ville, since that's where I almost always am. More advanced (mobile) users can hook things up with their phone to broadcast to local listeners wherever they are). I don't use this feature much, but I hear it's popular. Sometimes when I get bored I take a look at other people's local messages. Or post a message asking if anybody is up for a run this afternoon, or some frisbee.

So how do people see what I've posted? You must have noticed that in the above I only talk about generating content, not presentation. When somebody goes to my base address (like my server points them to the suggested presentation means (for example, a link to a blogger template), which is something I've customized (added various widgets, changed the layout, etc). But since my data is all stored in standardized formats, it is quite likely that individuals accessing my page have set up their browsers to ignore my suggestion (actually, probably set things up to not even bother asking), and will use their own template. This is a natural extension of the sorts of post-production scripts people already run (greasemonkey, ad-block plus).

People can see my global posts without any further interaction on my part, because I've set things up to automatically accept requests to join my global feeds (there's nothing stopping people from joining the rss feed for my blog, or tweets). When somebody signs up, it shows up in my list of friends, under my global followers. This list is mostly only used as a distribution list for when I write something new, but I also feel like it's polite to then sign up for my followers feeds (most of the time). When I meet new people, I can add them to my various friend networks and they will then be able to follow my posts, if they want. If somebody annoys me, I can remove them from the list, and they won't get my (vitally important) updates. I can even block them from signing up again.

What's great about this system is that it handles all of my online communication. Emails, IMs, tweets, blog postings, feed subscriptions all come and go through this personal communication channel. I mean... IMs are just rapid-fire emails, emails are just individually-audienced blog posts, RSS feeds are emails you don't respond to (but you are, of course, encouraged to make insightful comments concerning). Now I've got one system to both send and receive all of it.

Perhaps some people will have this set up through some company. Like amazon hosts everything for you, or myspace (given recent events). These companies will provide nice ways to interface with all of your data, but the better ones also make it easy to bundle up your data and take it to a different service. Open source projects will also provide these services, but you'll still have to find a host (surely there is an analogy to setting up a wiki using twiki, or a bulletin board using phpbb).

And one day, perhaps a nice cloud will come along, and I can have my setup there, so I don't have to worry about porting my data. Semantic web technologies will determine the content of my posts, and little autonomous agents will wander around the cloud, telling people that are interested in such content that I've posted something new. And dually, I'll have little agents wandering around gathering up things they think I'll find interesting (instead of wading through rss feeds for blogs that don't have a focused topic), and sending little links back to me. They'll not stop at forwarding pages, but will send me directly to the original author (so instead of looking at the digg page for an article, I just see the article (and perhaps a note it made it to digg)). The comments generated by anybody, anywhere, will all (mod privacy) be accessible to me. My little agents are turning up more and more interesting items every day...

Monday, May 5, 2008

Notes in Reader

Fun new feature in Google Reader: add notes to shared items. I've been wanting this recently, so that when I share a feed, I can attach my own little blurb. And it comes with a bookmarklet, so now I can share items without having found them from a feed. Hurray! My own little digg (minus digg's RDFa I guess).

A few things:
  1. What is the keyboard shortcut? [Update 8 May 2008: Shift-D]
  2. In the box that pops up where I enter my note, it has a check box for 'Add to shared items'. What happens when I uncheck that? Where does my note go? It doesn't seem to go to my Google Notebook(s). [Ok, it seems to just go to 'Notes' under 'Your Stuff' on the left panel of Reader, and I guess looking at it there I could then 'Share' it]
  3. I've got the shared items widget on this blogger page, but it ignores my notes. That needs to not happen.
  4. [Added after first posting] How do I un-note something?
In fact, if the bookmarklet gave me the option of not bookmarking the current page (just making a share-able note), then I can roll my shared items and twitter feed all into one. That'd be totally sweet. Google, I know you are listening...

[Update (<5 minutes after original post): Also something relatively new (I never saw announcements for it, and it didn't seem to be there a few days ago) in reader is that you can see who your friends are that can see your shared items. Of course... I don't have many :( ]
[Update (<10 minutes after original post): Could this be a step in liberating twitter? Also I noticed my first link was a link to a feedburner page, so I updated that]

Sunday, May 4, 2008

Data Portability

I realized that running our own "cloud" makes the idea of data portability irrelevant. I shouldn't have to 'port' my data anywhere. I should have my data stored where I know how to find it (which in a cloud seems like a trick), and then allow services to access it. What matters is standardized file formats that have enough semantics built in, and it seems we're headed that way.

At some point I thought these ideas were heading somewhere. Once I sat down to start writing, things started to crumble. I'm starting to think that's an important sign.

Saturday, May 3, 2008

Reader Tip

I spend most of my online time in Google Reader (or links followed from it). It's only taken about 6 months to become a complete addict (not necessarily of reader, but rss/atom in general). My feed list is currently sitting just under 200, and I go through all of my feeds religiously. If there's something I want to come back to, I star it, but at the end of any given day, I have no unread items (approx. 3-4 hundred items/day). Recently, I've been thinking about changing things up a little. There are plenty of feeds out there which I simply can't keep up with. For example, I used to subscribe to the main Digg feed, but couldn't keep up. To manage that, I just subscribed to the subcategories I was interested in. But then I found Mixx and LinkRiver, and am getting a little overwhelmed.

There was a post recently over at ReadWriteWeb with tips for feed reading. The number one tip was to oversubscribe. At the time I didn't think that was for me. But now I'm starting to think it'd be ok to not read all of my feeds. Except, some of my feeds I want to see every item from (basically all the feeds I have before adding mixx and linkriver). I knew in google reader you could specify what tag to start from. So I thought I'd make a new tag, 'start' with all of my normal feeds in it, to keep up with, and a tag 'firehose' to browse through when I get time, but not stress about reading.

The google reader 'Manage Subscriptions' interface is pretty friendly, letting you change tags and remove feeds, and quickly find feeds you are looking for. So I made a tag 'start', clicked 'select all feeds', and then under 'More Actions' chose 'add label' -> 'start'. Wait for it... wait for it... 'Ooops, an error has occurred. Please try again in a few seconds.' in the standard red bar across the top. Ok... well, trying again didn't seem to help.

Luckily, last night I'd decided to play around with how reader handles import/export of OPML files. Recall that OPML files are for outlines (and I guess are the typical way feed lists are stored), so they have a pretty obvious nested structure, and the file you'll get by exporting your feeds from reader is pretty easy to understand. There are <outline> elements that have just a title and name attribute, and these are the folders, and then within them are <outline> elements with feed information (name, url...). Let me distinguish these by saying 'folder' elements, and 'feed' elements. So here's what I found about importing:
  • If a feed element is for a feed that you are already subscribed to, then the feed simply picks up a new tag for the folder element it lies in from the OPML file. So your feed now has multiple tags and will show up in multiple folders, if it didn't already.
  • If a folder element is the same as an existing folder you have, the feeds will simply be added to that folder. That is, you won't end up with two folders with the same name or anything.
  • If your OPML file has deeper nesting than folder/feed (which is what you get from a reader export, but not what you'd expect the generality of an outline to be capable of), the folders get flattened. For example, if your file has a folder element 'folder' with folder elements 'first' and 'second', which themselves just have feed elements, then in reader you will see 3 new folders: 'folder', 'first', and 'second', and feeds from the sub-folders will show up in both 'folder' and whatever sub-folder they are part of.

This seemed to be the setup I needed to do what I want: add a tag 'start' to all of my existing feeds. I exported my feeds to an OPML file and fired up my editor. Surrounding basically my whole file (that is, surrounding all of the <outline> elements) I added a new <outline> element, with attributes 'title' and 'text' both set to 'start'. Back in reader, I then told it to import this file. A short pause, and it said everything had gone through. Back in my feed list, I now have a 'start' folder in addition to all of the other folders I already had. All of my feeds show up in two places (items only get counted once, as it should be), so I can see what folder has how many items, and I can scroll through the items just as if I was looking at 'All Items'. But now I can go back and add feeds for things like mixx, linkriver, fark... (feeds with overwhelming numbers of items), tag them as 'firehose', and not feel bad if I don't get to them all.

[Damn, the list messes up my pretty left-hand borders again. I guess I'll have to mess about with some more css some time soon. Update (12 May 2008): I just put a span around everything. I should see if I can do that in my template, to save a step.]

Friday, May 2, 2008

Rothe Diagram

Last week I attended a seminar entitled "Symmetric Group Sn and it's Kissing Cousins", given by Ta Khongsap. This came hot on the heals of Dr. Arone's treatment of the Barratt-Priddy-Quillen Theorem (relating classifying spaces of the symmetric group with stable spheres, totally awesome, I should probably return to it again here sometime). The two together made me think I should probably try to learn more about the symmetric group. In the mean time, I wanted to mention one of the interesting things Ta brought up in his talk, called a Rothe/Lehmer Diagram.

Recall that the symmetric group on n letters (denoted here by Sn) is generated by the elementary transpositions si, for i=1,...,n-1, where si just switches i and i+1 (so it is typically written (i i+1)). Given an arbitrary σ in Sn, we might like to know how to write σ as a product of the generators si (say, using the least number of generators). To do this, we begin with the Rothe diagram, constructed as follows: Consider an n by n array, and in the i-th row, put an 'X' in the σ(i)-th column. For example, the permutation (1 4 2 5 3) in cycle notation (which I'll take to be in S5) has Rothe diagram:

. . . X .
. . . . X
X . . . .
. X . . .
. . X . .

Next what Ta had us do was draw lines down from every 'X', and lines right from every 'X'. While I'm sure it looks terrible here, hopefully you get the idea:

. . . X -
. . . | X
X - - + +
| X - + +
| | X + +

We're almost there. To finish up making the diagram, work your way down the rows. In each row, work right to left as follows: if you are in the i-th row, start with si in the right-most open cell, then si+1 in the next-right-most open cell, etc. So in the example above, we'll get:

s3 s2 s1 X -
s4 s3 s2 | X
X - - + +
| X - + +
| | X + +

Now we can read off the product of the si, from top to bottom, left to right, and we see that:


(which is easy to check).

For more on this topic, I might point you toward either of these pages.

(I know this post broke my pretty borders on the left side of the page... sigh) [Update 12 May 2008: fixed it with a span around everything. Now about that ugly whitespace...]