Friends

Saturday was learning the care and maintenance of a caravan, followed by work in the garden.

Sunday was cutting the hedge, breaking the scissors, getting new scissors, breaking them too and finally in the end getting quality German indestructible scissors.

Monday was resting and going for a walk to Mechelen. Yesterday was a course and discovering a large (15cm across) infection on my leg. So I got a pack of instructions (if ... or ... or ... then go to the hospital immediately) and two packs of antibiotics.

So a quiet weekend :-S.

This evening I made my first Wikipedia edit ever. It was a simple enough thing: someone had inserted the word "lolz" into an article, and I felt compelled to rectify the situation out a sense of duty, honor, moral rectitude... okay that's all bullshit.

Why did I really fix it? Because I got to log the revert as: No lolz plz k thx.

Simple pleasures for simple minds.

http://us.rd.yahoo.com/sports/rss/mlb/SIG=12924uemf/*http%3A//sports.yahoo.com/mlb/news?slug=txredsoxcrisp&prov=st&type=lgns

Boston Red Sox center fielder Coco Crisp left Tuesday's game against the Baltimore Orioles in the sixth inning with a stomach ailment.

http://us.rd.yahoo.com/sports/rss/mlb/SIG=128im0vtk/*http%3A//sports.yahoo.com/mlb/news?slug=txredsoxdrew&prov=st&type=lgns

Boston Red Sox right fielder J.D. Drew left Tuesday's game against the Baltimore Orioles with a sprained left wrist in the bottom of the third inning.

Previously unread.

Another Tempe Brennan book. Some things are comfortingly similar, others are reassuringly different. Donät know what more I can say without giving the plot away.

Defintely worth a read, if you like kinda-sorta sciency crime fic.

I'm not a big fan of subpixel antialiased text, but it is technically neat and not difficult to implement, so for fun I started adding it to McCLIM's mcclim-truetype package. Since mcclim-truetype doesn't do hinting, the fonts are considerably more blurry, albeit less distorted, than what you get from libfreetype (and consequently mcclim-freetype). I thought subpixel AA might be of particular help due to this.

I hacked up a quick version which rendered the glyphs at 3x the normal horizontal resolution, filtered the resulting alpha mask to reduce color fringing, then combined triplets of pixels into a component alpha mask to use with Xrender. This didn't look bad - softer but more clearly defined.

Shortly thereafter I encountered this page suggesting that the subpixels of an LCD panel are not evenly spaced, but instead grouped together with a slight gap between the RGB triplets. If that's true, such a display must distort the image compared to my expectations, pinching together the triplets of subpixels. I tried to verify this on my Thinkpad's display - I think the grouping is visible, but my camera lens wasn't really up to it.


This suggests we do things slightly differently:


With that in mind, I hacked in a second mode which renders at 4x the original width, filters the result, and maps the first three of every four alpha pixels to an LCD subpixel (discarding the fourth). This should (I hope) provide less distortion in the mapping from the coordinate system of our glyphs to the physical display. This approach also looks pretty good, and it's sharper than the previous approach due to the filter width effectively shrinking by 3/4 on account of the higher resolution. I can't say either approach is clearly better.

Here are all three for comparison:


For whatever reason, the character spacing gets thrown a bit out of whack (in addition to our already not bothering to use the kerning tables). Figuring this out is next on the agenda.

As mentioned in a previous post, I really want to revamp the nklein website. With my recent obsession with Lisp, I wanted to do some of it in Lisp.

I'd like to score each of my projects on some number of scales... sort of a fuzzy tags thing. Then, I'd like to use Vecto to generate a navigation icon based on the scores on the different dimensions. This would be cached, but generated when dirty.

I'd like to add trackback pings (or something similar). I'd also like to use the LiveJournal XML-RPC interface to pull journal entries with particular tags onto a page at nklein.

( So, my work's cut out for me.... what's the complaint? )

http://us.rd.yahoo.com/sports/rss/mlb/SIG=12cm0u98p/*http%3A//sports.yahoo.com/mlb/news?slug=txredsoxfrancona&prov=st&type=lgns

Boston Red Sox manager Terry Francona will miss the upcoming two-game series in Baltimore following the death of his mother-in-law.

http://us.rd.yahoo.com/sports/rss/mlb/SIG=12he98ca8/*http%3A//sports.yahoo.com/mlb/news?slug=ap-redsox-franconaout&prov=ap&type=lgns

Boston Red Sox manager Terry Francona left the team and will miss the first two games of his team's series at Baltimore following the death of his mother-in-law. Mary Ann Lang, mother of Francona's wife, Jacque, died Monday night in Tucson, Ariz. Francona and his family traveled to Tucson on Tuesday. Bench coach Brad Mills was to run the Red Sox in Francona's absence.

AutoMotivator has a special fancy high res version so you can, in theory, make a motivational poster big enough to print and hang on your wall.

This has worked out reasonably well, but one sore point is the performance of ImageMagick for big images.

AutoMotivator works by taking the original photo, creating two PNGs for the title and inspirational text, and compositing them together against a black background with convert. On big images (2000 pixels or more on a side), it gets very slow (5-20+ seconds) and sucks up a lot of memory.

I'm not sure why it's so slow, but I suspect it's loading all the input images into memory, creating a big output image in memory, filling in the output, and then writing the output to disk.

Since my compositing needs are pretty simple, I don't have to load all that image data: producing each row of the output image only requires reading one row from each the images that single output row intersects. (This thought was originally inspired by picturetile, which must be incredibly slow.)

With that idea in mind, over the weekend I wrote a small compositing program in C. It uses the row-at-a-time input and output functions from libpng and libjpeg. The basic algorithm is:

• get the dimensions and placement of the input data sources
• note which data sources appear on each row
• create a jpeg output object, and one output row
• for each row in the output:
  • for each input source on that row:
    • merge a row from the input into the output row
  • write the output row out to disk
• finish the output jpeg

I'm really rusty at C (it's been over ten years since my last useful C program), but the end result is working well: my compositor is about ten times faster than ImageMagick at assembling big poster images. Sure, it also has a hundredth of the features, but it fortunately does exactly what I need. And the difference in user experience quality between producing a poster in two seconds and 20 seconds is huge.

http://www.librarything.com/blog/2008/05/top-ten-suggestions.php

Just over a week ago* we asked members to come up with their recommendations on "Ten Ways to Make LibraryThing Better." We promised to pick twenty-five winners, including ten winning answers and fifteen random picks.

Members heard the call, writing 259 answers for a total of 45,000 words--slightly longer than Henry James's Turn of the Screw. Last week Sonya, Abby, Casey and I got together to work on LibraryThing for Libraries. We took a break on Wednesday to (drink and) read through the answers. We couldn't pick just ten winners, so I've expanded it to 17--32 winners total. We could have easily done 50 more.

The Prizes. Winners get to chose between (1) A CueCat barcode scanner; (2) A LibraryThing t-shirt; (3) First dibs on a LibraryThing Early Reviewers book.

Winners should let Abby (abby@librarything.com) know what you want. If you want the Early Reviewer book, you're also going to need to change your Early Reviewers picks to select just one book. We're going to give you an "ER mojo" of a million, so whatever you pick, you'll get.

The Winners. Random Winners: rfb, maryanntherese, jocainster, Imprinted, circeus, jabogaer, rastaphrog, claudiuo, jjmcgaffey, arnzen, trojanpotato, surly, phoenixfire, sigridsmith

sophies_choice (7): "Let us mark which books are our favourite." I'm divided whether to make this work like author and venue favorites, or to make it a "collection."

PhoenixTerran (31): "Update debris and author pages immediately after combining/separating has occurred" A big leap is going to happen here very soon, with the introduction of a more stable "editions" layer. I'm actually doing edition-level calculations in the background today, with an eye to inaugurating the system on a limited basis tonight.

Philtill (160): We all loved Philtill's ten suggestions, which amount to "Make LibraryThing more like Tickle." There are dangers to personality tests and statistical correlatons, of course. But we love to play with data, and "tell me about myself" is one of the main reasons people use LibraryThing anyway. So, expect us to take these ideas seriously.

jocainster (28): "Add a link to the book's main page in the 'Recently Added' section." Abby had to be restrained after reading this one.

parelle (44): Parelle wrote two related suggestions--LT bookmarks and a parnership with Moo Cards. dreamlikecheese focused in on sending cards to libraries and bookshops. This is one area we're definitely going to look into.

sabreuse (152). "I was at a conference last week where I picked up several new books, but didn't have internet access all day. And I realized that I want to be able to add books by SMS, the same way I can send photos directly to flickr or add events to my google calendar by text message, both of which I do all the time. I'd love to be able to add new ISBNs to my library while I'm out shopping, or traveling, or tied up away from a computer."

nperrin (17): "Some ingenious way to link books to books about them. If I'm looking at a novel, I want to know how to find the best criticism of that novel or author."

usquam (109): "Work with publishers to get better integration of their catalogues into LibraryThing. They should have covers, contents, editions, etc - as per the new 'series' area, it would be interesting to see what we have from a particular publisher, and then have them show other editions or titles we might like or are missing."

susiebright (155): "I loved Secret Santa; it was the hightlight of my Xmas gift giving because it was so entirely unexpected. I think you should offer a 'Birthday Surprise' gift program of the same kind. You pick a 'birthday kid's name' out of the hat, and send them a book based on what you glean from their library!'" We're thinking that BirthdayThing could be hard to arrange, but doing a mid-year (June 25?) Secret Santa sounds fun. This time, members are doing the ordering!

yhoitink (9): "Add the European Library as a source." Casey is squarely behind this one.

amysisson (87): "a virtual 'badge' or 'ribbon' (like LT author) for on the profile pages of people who've contributed over a certain level(s) of info, such as CK or combining" I'd love to do something like this. I'm attracted to the Barnstar model.

papyri (95): "Provenance, ex-libris (previous owner(s)) info listing (can be done like multiple authors). Possibly including dates and locations. Privacy option for this would be nice." Sophies_Choice also suggested this be integrated with LT Local. Good stuff.

ssd7 (111) "Cross Source Searching. So, I would like to get my data from the LoC. But I would also like to just punch in an ISBN. These two desires are not always compatible since searching on ISBN's often yields nothing from the LoC. When a search returns no results why not use the LT database or Amazon to find the title and then research for the user? Or at the very least let me set up a 'priority' listing of the sources so that if LoC yields nothing, it will automagically search Amazon." ssd7 (111) also suggested "Open source the code." This continues to interest us. No promises.

hegelian (16): "OpenID might be a smarter way to login for some people."

_Zoe_ (24): "The ability to reset the unread marker at the message you've actually read up to."

zcannon (25): "A widget that works on Wordpress."

TerrierGirl (34): "Could each book's original copyright year be added to the my library, add to library screens? This would help interested potential readers place each book in time. Also, it would tell a reader when a particular book fell within that writer's career." I've wanted to do this for some time.

Notes on Method. We decided to leave off a small number of common topics, including collections, author disambiguation, HelpThing, tagging of groups, web links on book pages, more than seven columns, and a Facebook application. They are very much on our radar already. Seeing them over and over again had its effect, you can be sure.

We also left off suggestions for features completed since we asked the question, like better tags, and to avoid new features in favor of bug-fixing. It's a delicate thing, and not one we've always gotten right, I'll admit. I've been on a bug-fixing and performance kick recently.

---

*That's my story and I'm stickin' to it!
**The person you don't know is Mike, a local Portland programmer working with us part-time for a few months. Note, I was supposed to be also sitting in the chair—reading Everything is Miscellaneous—but there was a tragic head/butt airspace issue.

The web browser is the new operating system. My own computer is at this point little more than a glorified web browser, with a text editor, command prompt, python interpreter and svn thrown in for the occasional color.

Since the web browser is the new OS, it should really, really, have the equivalent of a process list. I almost always have a whole bunch of tabs open, and firefox is most of the time using a nontrivial amount of CPU doing not much of anything. I have to guess which tab is causing the problem when the CPU gets pegged, and sometimes it seems that even shutting down all tabs doesn't completely fix the problem. Could somebody please implement metrics on how much CPU each tab/window is using, and get the process separation right so that whenever a tab/window is shut down all remnants of it are completely toast?

http://www.schneier.com/blog/archives/2008/05/what_is_the_com.html

The Department of Homeland Security has a new $200 million Comprehensive National Cybersecurity Inititative (CNCI). Congress is happy to fund it, but kind of wants to know what it's going to do.

I have to admit, I'm kind of curious myself.

00:00, May 13.
Achewood posts a comic showing wikipedia lying about Finland.


10:32, May 13.
Wikipedia begins to lie.


Minor edit-war follows.

13:55, May 13.
Wikieditor Wafulz semi-protects page.


All such edits are reverted.

smason wrote:

I wrote a Missile Command clone for the multi-touch wall at Obscura Digital. Just like the original, except you can fire by touching the wall with your fingers. Save the Golden Gate Bridge from ICBMs. Fun for the whole family!

After much work, Fedora 9 is finally out. It pretty much rules, so you should really check it out. Digg it here

http://lambda-the-ultimate.org/node/2811

A History of Erlang and the accompanying Presentation Slides by Joe Armstrong are a must read for anyone interested in PL history.

Erlang was designed for writing concurrent programs that "run forever". Erlang uses concurrent processes to structure the program. These processes have no shared memory and communicate by asynchronous message passing. Erlang processes are lightweight and belong to the language and not the operating system. Erlang has mechanisms to allow programs to change code "on the fly" so that programs can evolve and change as they run. These mechanisms simplify the construction of software for implementing non-stop systems.

(Link to previous HOPL-III papers on LtU).