nosignal

Thanks for sharing these links. Honestly, charity isn't something I think much about, but have a poor opinion of. Nice to be reminded that good applications of it exist, and that it's measurable and assessable. The 80,000 Hours site in particular looks like a great resource.

Also quite a recommendation coming from you - never fails to amaze me the people you find posting on here.

This looks beautiful, but it doesn't release my mouse click on Chrome; i.e. when I click on the globe to click & drag, the rotation of the globe keeps following my mouse, and I can't click the "play" button.

Nice project though, as a geo guy who's been meaning to play with D3, it's inspiring.

With your keyboard set to dvorak.

If I had karma enough to downvote advertising, I would.

FWIW, Little Bird electronics are another Australian supplier of Arduino & associated electronics. http://littlebirdelectronics.com/products/arduino-uno-r3

No association, except for being a satisfied customer.

It's always a toss-up between ordering locally & just shipping from Sparkfun, though, especially with the exchange rate.

That blog post was fascinating — I had no idea the Fan community was so mobilised (nor that they were even a community cohesive enough to have a name, nor that the name was Fans).

I wonder how many other comprehensive communities exist like this, somewhat under the radar, serving themselves.

Am I the only one who doesn't use bookmarks?

99% of the time, if I need to find something I've read before, I'll just search for it & find it.

The exception to this is when I was working support and had several product spec sheets & other reference pages bookmarked for quick access to email links to customers. Even while I was at Uni, I'd download articles to PDF.

On the rare occasion that I like or am interested in something enough to want to bookmark it, I find I rarely actually come back to it. Just looking now, in my Bookmarks menu I have a couple of folders — apparently the result of me saying "save all open tabs as bookmarks" in the hope of coming back to it later — but never having done so.

Long-form articles I come across while I should be doing something else go to Instapaper to be read later (is this what people use bookmarks for?). Even then realistically I only actually read 20% of them.

Anything I really think a friend of mine should read, I email a link there & then. This is a 1- or 2- click operation.

Anything that I think is just too interesting/funny/important to keep to one or two friends, I put on Twitter, or here on HN. (I guess that ends up being a kind of bookmark service; I have scoured my submitted stories (and upvoted stories) to find them again.)

For my frequently visited sites without RSS, the browser URL bar is my friend - just start typing & it autocompletes. The only bookmark in my bookmark bar is the "Instapaper: Read Later" bookmarklet.

What does everyone else use bookmarks for? Is there some awesome use case I've been ignorant to since 1996? Am I just a freak for never wanting to revisit things?

Thanks for that mashup link, I'm always on the lookout for web map references so that's pretty useful. Cheers.

Thanks for putting it up. I'll have a play with it!

Is this code that you can put online somewhere? This kind of thing would a) likely be really handy for lots of people and b) is the kind of thing I'd like to be able to make myself (using browser as a readymade UI), so having an existing reference would be helpful.

...any runtimes except *nix (or cygwin), that is.

From the 1983 BBC "Fun to Imagine" series. Video for reference: http://www.youtube.com/watch?v=wMFPe-DwULM

At the risk of being pedantic, it's worth noting that some things benefit from memorisation (e.g. the alphabet) and some things benefit from "higher" learning (e.g. understanding the concepts of programming). I don't think patio11 is necessarily saying "something is better than nothing"; sometimes it is _better_ to teach someone 'if you see the shape O, you say "oh"'.

I saw Damian Conway present about this, and it was absolutely fascinating. It started as a practice of removing syntax from code position and instead encoding it in variable/function names, like Latin does.

"The paper also explains the special source filtering and parsing techniques required to efficiently interpret a programming language in which the syntax is (largely) non-positional."

Here's the Sieve of Eratosthenes:

        #! /usr/local/bin/perl -w
        use Lingua::Romana::Perligata;
        maximum inquementum tum biguttam egresso scribe.
        meo maximo vestibulo perlegamentum da.
        da duo tum maximum conscribementa meis listis.
        dum listis decapitamentum damentum nexto
            fac sic
                nextum tum novumversum scribe egresso.
                lista sic hoc recidementum nextum cis vannementa da listis.
            cis.

I'm supposed to be teaching a class on introductory Python (students are doing their exercises now) and I want to scrap the rest of the day to talk about how phenomenally awesome this news is.

I'll settle for a strange, quiet sense of elation.

No source besides my discussions with techs in Australia Post, sorry.

I see these are built on the Australian Bureau of Statistics "approximate" postcodes which are matched to census districts and therefore more useful for demographic analysis; no CDs cross postcodes in this dataset. There are a bunch of similar geographic datasets available from the ABS, including electoral boundaries, local government areas and so on: http://www.abs.gov.au/websitedbs/D3110124.NSF/24e5997b9bf2ef...

All of that data is approximated by the ABS from whatever body controls the "real" versions, but is 99% good enough for anything you'd want to do with it.

Australia Post, on the other hand, don't really use postcodes to deliver mail anyway, I've been told. They're a vestigal construct which are very prominent in people's minds but so fraught in the implementation that any systems which analyse them (including Australia Post's) work against them rather than with them. I didn't know about Canada's system — that sounds like a much better idea!

I noticed this too. It seems advice to "pick random words" should be extended to "pick random words and leave the last letter off".

Hangul, the Korean writing system, is at once phonetic and syllabic. Each "symbol" represents a syllable, but each sub-symbol represents a particular sound in the syllable; and each sub-sub-symbol encodes phonetic information about how it should be pronounced. See http://en.wikipedia.org/wiki/Hangul#Letter_design for a better explanation.

A binary representation of the Hangul system could be very interesting. Use the symbol 1 for "vowel base" and +1 for "high", +2 for "high front", +3 for "high back"… You could eventually have each byte represent a syllable. It'd be like chmod commands + phonetics.

There goes my evening…

I can see this being applied in labels etc. for computer vision. Some kind of control/orientation character could make it similar to a QR code but actually readable by humans.

Generally, though, if something's going to be read by both, you may as well just have a label with plain English + a QR code. No reason why it has to be readable by both.

Definitely. Could be offered as a one-off design, I imagine.

Rounded corners, drop shadows, subtle palettes with contrasting call to action buttons, several dozen social network sharing icons... Did I miss anything?

I had some similar thoughts while taking a night class in woodworking a few months ago. Seeing as I'm never going to blog it, I'll put them here. Hopefully, the parallels with software projects are obvious.

Have a side project on which to practice new techniques.

This course was structured to have the first 2 or 3 weeks learning basic techniques (joins, gluing, tool use) and the rest spent using those techniques to build a final project. Because I was just learning, I'd practice and practice making e.g. mortise & tenon joints on pieces of scrap wood, and then throw them away before doing the "real thing". Often these practice joints were better than the final product, and I threw away a lot of wood (and time).

While you'll always need to do some throwaway practice, it would have been better if I'd had 2 projects on the go. That way, I would have ended up with a probably-still-incomplete side project (say, a simple side table) as well as my butcher's block. Most likely I would have scrapped the side project at the end anyway, but it's possible it may have ended up of equal or better quality as my supposed "main" project. Not to mention you learn more when you have a practice project than when you're just making joins in scrap, or gluing random scraps together; if you treat your practice as a whole project then you get to not only practice the individual techniques, but also how the overall project comes together.

Adjust the plan to the materials.

Starting a woodworking project was entirely new to me so I was a little nervous. I'd heard "measure twice, cut once" a hundred times and was terrified of getting any measurements even a millimetre out. The key to a beautiful finished product was precision, I thought. As such, I approached the plan I chose as gospel. However, after selecting a plan for my project, it turned out that I couldn't get the timber in the exact dimensions the plan called for. This struck terror into my heart - how can I make a cut 130mm if my lumber is only 128mm long? At first I would try to wrangle (glue, cut, or plane) the timber into the dimensions the plan called for. This was very wasteful of time, effort & materials. After realising that, I tried just pressing on blindly and worked only with the materials, but 2 or 3 steps later, things started to not join together properly because my measurements were out.

Eventually I learned the best way was to pick a plan, gather the materials that most closely matched the plan, then go _back_ to the plan and redraw it with the measurements of the materials I'd found. This was much less effort in the long run, and allowed precision within the constraints.

It's not done until it's finished.

This one's pretty straightforward and hardly unique to woodwork, but woodwork lets you make a bad pun about it. My final project was a standalone butcher's block for the kitchen. Although the construction is complete, and everything's glued & fixed together, it's still not done. I haven't had time to do the final sanding, polish the legs, and finish the surface with oil. Being exposed to food oils and fats, the surface needs to be finished before you can use it, otherwise it'll chip, split & splinter. I didn't leave enough time in my class to finish the surface, and haven't made time to do it at home, so my project is at once 95% and 0% complete. At the moment, it's effectively just a heavy piece of fuel in the garage.

If there's some aspect of the project that conceptually seems like a "final step", make sure that you budget for it in your project scope, and don't think "well, the hard part's done, so I'll stop work" - the project isn't done until you finish it (in the case of a woodworking project, literally).

Interesting you mention this; I've recently started a new job which requires me to learn a whole new skillset (web programming with JavaScript). I've been a developer of desktop apps for a while so programming isn't new, but the language and the paradigms (async? wtf?) are challenging. This is in contrast to my early memories of "web programming" aka making a static Geocities site with notepad and wondering why everyone was so excited about the web when it was this easy to make a web page... turns out it's not that easy anymore.

The point being that it really is in those moments of confusion and frustration - where you try something and it doesn't work, or you realise you don't understand how to do what you want to do, especially when you think what you're doing is supposed to be easy - where distraction comes in and you go off track.

I've managed to realise that and am glad I'm not alone. The next step - actually doing something about it - is a little harder (hence why I'm on HN right now...!).

I've just watched two of them (8: Data Structures from SICP & Graph Algorithms from Algorithms) and enjoyed them. The format of the videos is nothing special (camcorder on a tripod at the back of the room), but the teachers are very good, presenting complex subjects clearly. The videos alone seem an excellent reference & resource. The accompanying materials (handouts etc) are sparse but good accompaniments. The courses were taught assuming reference to a textbook, so they're not 100% self-contained.

It seems most videos are available on Google Video if you (like me) had forgotten that RealPlayer even existed: http://www.google.com/search?q=Arsdigita&tbo=p&tbm=v...

(Allegedly VLC can play RealPlayer, but it wasn't working for me.)

In this specific course, not much. I personally had no maths education beyond year 12, and only what I'd picked up along the way while reading around the subject. All the maths you'll need to know is explained thoroughly enough in the lectures. I think if you already know linear algebra, you'll have a head start - I didn't know it, and it took a bit of work to get up to speed.

There is a lot of discussion of derivatives and partial derivatives in the course, however it's entirely possible to complete the course and feel comfortable with the material without understanding derivatives. At least, it was for me =)

As for machine learning more generally, I'm certain a greater understanding of the maths behind what's going on would make you a better practitioner. I'm sure there are plenty of people currently making their living applying ML techniques without being maths experts, but I'm also sure the leaders in the ML field are maths experts.

Here are my notes (on github). Nowhere near as polished as this version, and probably reveals more about my process of understanding than machine learning itself, but if we're sharing: https://github.com/mechamoth/ml-class/blob/master/ML_Notes.o...

"What would prove my assumption (or idea) wrong?"

It falls under experiment design, but it's a much more straightforward and easy to understand question. You can ask kids that question (they're quite good at it). Really helps narrow down avenues of inquiry & speed up your problem-solving.

It's only useful if the person is already interested in being less wrong, but everyone is interested in being less wrong about something. When they (you) are, this is an excellent question to ask.

I wonder about this. Apparently the letter we get at the end contains our ranking. I wonder what the distribution of scores would be; I'd assume there would be a small fraction of enrolments actually doing the exercises, a big spike at 100%, and most falling in the range of 80-100%. Personally I forgot that I hadn't done a couple of the quizzes, so did them late, and didn't complete the optional part of the first week's programming assignment on time, so although I've probably got a "good mark" on paper, I'm probably nowhere near the top 10% of students.

I wish I could +1 this more than once. This is exactly the point, to me. There's not really such a thing as "an AI job". There is application of AI techniques to existing problems. Those problems are already found riddled throughout every job everyone's ever worked in.

You can't sit back and wait for AI to "happen". What would that even look like? The benefit of these techniques is in application to existing data; you don't have to quit and found some kind of computer vision/kinect/big data startup to take advantage of the knowledge.

You should be aiming to put yourself out of a job by using AI.