maradydd: (Default)
I owe the Berlin trip a proper writeup, but some highlights: talk went extremely well, saw many old friends and acquaintances, came up with yet another paper we need to write with Dan Kaminsky, had some interesting discussions about a computer science curriculum that emphasizes security from the get-go, narrowed down the scope of some tools I need to write in the very near future in such a way that I can put together a proper spec now, got invited to give our talk or something very much like it again at Dartmouth. [ profile] enochsmiles and I co-present extremely well, which bodes well for future joint presentations (which I enjoy better than solo presentations, when they go well at least).

We also sort of got stuck in Berlin after seeing [ profile] foxgrrl off at TXL, as it turns out that trains from Berlin to Leuven are not to be had after about 2 pm; the farthest west we could have gotten was Liège. A glance at a rail map suggested a wild possibility: Saarbrücken, so on a wild shot I called [ profile] oralelk's office and got him on the first ring. Despite not having had much contact at all over the last, um, five years (bad Meredith, no cookie!) he was still quite happy to have us crash on his couch for the night, even coming out to meet us at 11:30 at night, staying up to chat, and putting off going in to work until well past 11 am despite having quite a lot of work to do. It was rapidly discovered that Saarbrücken is one of the least convenient places in Germany to get to Belgium from; our options were basically the ICE high-speed train to Paris and the Thalys to Brussels, or a bus to Luxembourg and two trains for roughly a quarter the price. Thus I have now been to Luxembourg, making that eight countries so far this year.

I have also just received notification that our Black Hat talk has been accepted. Thus, I will be both there and at the Open Science Summit in Berkeley immediately thereafter, July 29-31. (Current plan is to arrive in CA on the 30th.) Unfortunately, this will mean missing DEFCON, for me at least; I'm not sure about [ profile] enochsmiles.

It is going to be a wild summer, with tools to write and a journal article to finish and a couple of big chewy proofs to prove on top of all my normal work. But I'm excited!
maradydd: (bad post!)
Yes, I am ignoring you. September and October are when the early-spring conferences have their deadlines, and I have somewhere between four and six papers to finish between now and Halloween. You do the math. (The research is done, I just have to get it out of my head and my notes and into LaTeX in a form that other people can recognise.) Yes, you can read them, I'll post links to the tech-report versions as they happen. No, there is not much you can do to help, other than leaving me alone, unless you are here to give me a back massage, are going out to pick up food for me so that I don't have to get up, or can send me backward in time. Yes, I miss you. No, I am not dead. Yes, I will be back in touch when I have time to breathe again.

Coffee still works, kind of. I am contemplating alternative delivery mechanisms with better bioavailability.

And I get to do all this over again next April/May for the fall conferences. YAY.

(this post only brought to you by the fact that FC extended their deadline a week.)
maradydd: (Default)
I just wrote an attribute grammar for the general object-encoding rules. (Not all of BER; that's next. Just section 8.1 of X.690.) Unless sections 8.2 through 8.22 throw me a curveball, which it doesn't look like they will, QED. It's fully left-recursive and uses only synthesized attributes, for all your efficient parsing needs. Boo yah.

If anyone playing along at home wants to have a look, drop me a comment.

FWIW, attribute grammars are absurdly simple in Haskell; Happy supports them natively. (Attribute grammars are somewhat intractable in strictly-evaluated languages because if you have to use strict evaluation, the data dependency graph gets impractically large very quickly. Haskell, which uses lazy evaluation, doesn't have this problem. Every day I come to love this little language more and more. Oh, and I can do a multiple-entry-point approach to support CER and DER. Seriously, how cool is that?)

Don Knuth, Haskell Curry, Simon Peyton-Jones, Simon Marlow and Andy Gill are totally my heroes forever and ever.
maradydd: (Default)
Since my last update, I've picked up the necessary hardware to build a SpiffChorder, Mikkel Holm Olsen's homebrew AVR-based chording keyboard. The circuit is now mostly breadboarded, and one of my winter gloves has been converted into a prototype that looks scarily like something out of Serial Experiments Lain. ph34r my wire tentacles!

I have also fallen madly in love with the art of wire-wrapping. It's not a technique people use very often anymore, since soldering is so convenient and cheap, and almost nobody prototypes CPUs with discrete components these days -- we have FPGAs for that. However, it's a great way to hook up components that need a flexible connexion, such as glove-mounted switches that need to tie in to a breadboard. Wrap a lead, then wrap a single header pin, and voila -- breadboardable glove-mounted switch. (Also handy for connecting panel-mount components to a breadboard, using the same header trick. This worked great for the USB-B jack.)

[ profile] joel tells me that I'm not the only one doing my part to keep wire-wrapping alive -- Steve Chamberlin did prototype his own CPU with discrete components, showed it off at Maker Faire, and taught Joel (and presumably many more people) how to wire-wrap. Keep the dream alive, Steve!

Now to figure out what I did with those 82-ohm resistors, finish that breadboarding job, and try loading Mikkel's hex image onto my ATMega8. Cross your fingers for me -- I want to take this to the hackerspace meeting tonight.

ETA: hm, that's strange, the ISP pins are definitely lined up correctly (my multimeter tells me so!), there's a 2.2k pullup on the RESET line, but I still get the blinky red LED of no love from the AVR-ISPmkII. Actually, first it's the solid red LED of no love which isn't mentioned in the manual, then when I try to load up the image via avrdude it's the blinky red LED of no love. Downloading AVR Studio now, we'll see if the official tools give a more useful error.
maradydd: (Default)
I'm building a C++ project for an unusual platform, and am having some confusing problems with my libstdc++. For some reason which I cannot fathom, I am getting an absurd number of "undefined reference to..." linker errors for symbols which are indeed undefined in libstdc++.a, but which are definitely defined in libc.a and libgcc.a. Yes, I am linking to both of those. (I know I am, because earlier I was getting some undefined-reference errors to symbols in libgcc.a from the code I'm actually compiling, and when I added -lgcc they went away.)

Any idea what's going on here? Do I need to compile libstdc++ from scratch rather than using the provided binary? (Please, God, let the answer to that be "no".)

ETA: enigmatic ld ordering issues for the lose. Thanks, [ profile] tangaroa!
maradydd: (Default)
I spent my Thanksgiving weekend in a cabin in the woods near Lake Tahoe. Four days of eating well, lazing about in a hot tub, hiking, and sleeping a lot has me feeling much more human than I did last week. However, I had functionally no internet the entire time, and while I am sort of caught up on LJ, I am not particularly inclined to go back through twenty pages of posts to find out what happened while I was gone. If there was something you wanted me to read, tell me here.

I also got some absolutely badass news about one of my clandestine projects, which I will share in detail rather soon (so, uh, I guess it won't be clandestine anymore), but right now I have actual work I need to do.
maradydd: (Default)
Today I went to TAP Plastics in San Francisco and they cut me the plastic pieces I will need to build my very own gel electrophoresis chamber. (Service while you wait, by the way. It took maybe ten minutes, during which time the friend who drove me and I wandered around the store looking at all the cool stuff they had on display and chilling in chairs shaped like giant hands. A+++ will shop there again, if only to spend more time browsing through their photos of neat stuff they made.)

Then I got home and checked the mail and found that my electroporation cuvettes arrived!

Oh, today is a happy happy day. Time to leave positive feedback on eBay for the guy I bought the cuvettes from, fix some lunch, then go down to SuperHappyDevHouse and start building.

*dances the dance of parts that are heeeeeeere!*
maradydd: (Default)
Man, it's amazing the kind of stuff you can get done when you wake up at 9am.

So far, today I have rearranged furniture (my desk is no longer taking up all the useful space in the living room!), made biscuits for breakfast, tidied up the kitchen, and gone to the Norway Festival out at the Marina with my friend Josh and his wife Chris. I thought the Belgians made some seriously rich food, but the Norwegians have them licked -- I just missed a cooking demo for a dish which basically consists of baked sour cream topped with a butter sauce. Josh and Chris got the recipe, though, so I will loose it on [ profile] enochsmiles at the soonest possible opportunity.

I also watched a yarn-spinning demo and got to try it out. I don't have anything resembling the room to house a full-blown spinning wheel, though some of the ones they had there were surprisingly compact. My spinning-fu is definitely weak, but I was surprised at how easy it was to pick up, and I now have some lumpy yarn that I made all by my little self. I also picked up two hanks of lace-weight silk/merino blend (50 grams == 630 meters, yowza), one in claret and one in a rich sapphire blue, which I aim to use for matching fingerless gloves for [ profile] enochsmiles and myself. He's taken some gorgeous photos when we've been out walking around in the mornings, but my hands get awfully cold, and fingerless gloves would help a lot without getting in the way of camera operation too badly, I figure. I found several dozen glove patterns here, and now must figure out how to convert a sport-weight pattern into something that will work with lace-weight yarn. (Knitting buffs: can I get away with just working double-stranded?) I will also probably just use [ profile] cassandrasimplx's ultra-clever trick of working on paired circular needles rather than double-pointed needles, as I can think of little more frustrating than uber-skinny needles, uber-skinny yarn and my uber-skinny fingers becoming a horrible tangled mess.

So, in a little bit here I'm going to pick [ profile] miss_education up at BART, boogie on down to SuperHappyDevHouse with her and [ profile] foxgrrl, and get my libdejector hack on. ([ profile] palecur, you down? We'll be blowing through your neck of the woods on the way there if you want a ride.) After that, the forecast calls for setting up the projector and watching goofy movies with Chris++. I have to say, it's a damn nice life.
maradydd: (Default)
The die is cast: my qualifying exam presentation is scheduled for 9 May, so it'll be the second to last major thing I do before I'm out of here for the summer.

Between now and then, I need to turn this fuzzy-search project I've been working on and bending everyone's ear about into a fifteen-page paper. Most of that will be taken up by A Concise History of Just About Everything To Do with Association and Similarity Classification; describing the project itself will just be a couple of pages, because really, the system isn't that damn complicated. (Counting the wrappers I've had to write, but not counting the libraries it's built on, the entire thing comes in at just under a thousand lines of code.) I'll also rabbit on for a bit about where I plan to take the project (read: "Get ready, because in another six months this is what I'll be giving my comps presentation over"), which should eat another page or two. The hard part -- the system -- is done. This is the bite-my-lip-and-grind-it-out part.

Current plan: knock out a first draft by Monday, then have it revised by the time [ profile] yoctohedron gets here on Thursday, so as not to be Horribly Antisocial Girl during that time.

All things considered, this is really no worse than writing up an entire syntax midterm the morning it's due, and I've done that without blinking before.
maradydd: (Default)
My advisor would have me shot if he knew I took out a couple of hours to do this, but I'm vaguely proud of it anyway.


maradydd: (Default)

September 2010

12131415 161718
26 27282930  


RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags