Brain numbers

I’m fascinated by reading about the brain in abstract terms. Here’s a new article in The New Yorker by Jim Holt that takes a broad look at the research being done by Stanislas Dehaene in Paris. He takes high resolution 3D scans of peoples brains while they’re thinking about or doing things to try and understand how our mind works.

This might seem like madness: “well, I know the left side of my brain is responsible for jumping” (say). But by isolating separate thought processes and functions to separate brain areas, you can gauge the limits of certain types of thinking.

The aforementioned article discusses numbers in this context: turns out that there’re neurons in the brain that fire specifically to recognise groups of objects in numbers of up to about five. Any more than that and you need to count the objects individually before you can know how many there are.

Here’s a result that spins me out:

A few years ago, while analyzing an experiment on number comparisons, Dehaene noticed that subjects performed better with large numbers if they held the response key in their right hand but did better with small numbers if they held the response key in their left hand. Strangely, if the subjects were made to cross their hands, the effect was reversed. The actual hand used to make the response was, it seemed, irrelevant; it was space itself that the subjects unconsciously associated with larger or smaller numbers. […] He even suspects that this may be why travellers get disoriented entering Terminal 2 of Paris’s Charles de Gaulle Airport, where small-numbered gates are on the right and large-numbered gates are on the left.

Say what? What implications does this have for other cultures who write and do maths in other directions? What implications does it for how we read a number like “728456” where the smallest number is on the right? Could we suddenly become better at doing maths by hand if we reversed our number system?

And speaking of being better at maths (there are tangential parts of the article that discuss the immensely better mathematics teaching programs in Asian compared to the western world), I am absolutely stunned by this:

Chinese, by contrast, is simplicity itself; its number syntax perfectly mirrors the base-ten form of Arabic numerals, with a minimum of terms. Consequently, the average Chinese four-year-old can count up to forty, whereas American children of the same age struggle to get to fifteen. And the advantages extend to adults. Because Chinese number words are so brief—they take less than a quarter of a second to say, on average, compared with a third of a second for English—the average Chinese speaker has a memory span of nine digits, versus seven digits for English speakers. (Speakers of the marvellously efficient Cantonese dialect, common in Hong Kong, can juggle ten digits in active memory.)

If this is true (and the result is logical to me), then why on earth haven’t we invented new words for our numbers? In a bit of a coincidence, I’ve been thinking recently how inefficiently I count numbers in my head and lazily trying to improve how I do that. I thought the inefficiency of my technique was verbally thinking each number as it passed, which is extra slow in the teens and seventies. My plan was to try and simply visualise the numerals in my head ticking past and not say them at all; I found I could actually do it for small stretches of numbers but I’d quickly fall back on my old techniques. Does everyone count in their head by thinking of the actual words? (I probably should have mentioned by now that the article says that yes, this is the case; and furthermore, you always use the language you were taught to count in originally. Large numbers and multiplication are stored in memory, apparently, as language strings.)

Of course, this is all for “average” humans; it sounds like the experiments that have shown these results are only just getting off the ground and it would be interesting to see what things look like for autistic mathematicians (not to mention magnetic field-induced temporary autism).

Interesting stuff.

---

“one, two, tee, for, fev, six, sen, ate, nen”

“one tee + ate fev = nen ate”

Mac security

Another day, another computer security rant and rebuttal. Here’s a recent piece over at MacUser by Dan Moren.

Now, don’t let it be said that I disagree with the article in general. This is gold:

Okay, if an instant message window pops up on your computer, with a screenname you’ve never seen, asks for your Social Security Number and you give it to them, then frankly you didn’t deserve that identity in the first place.

But I feel Dan goes a little overboard here in his refutal of whatever some jackass is saying:

Mac users need security software, too. Oh, I know: Mac lovers will respond to this statement by puffing up their chests, raising their chins and sniffing, “Ha, my Mac always has been and always will be virus free. I don’t need no stinking security software.”

Who are these people? Where does one find them? Maybe it’s those same people who go to oxygen bars and drink fancy bottled waters and have weekend quail-shooting parties before jetting off to some balmy Caribbean island where they play shuffleboard and laugh about how they don’t need security software.

Okay, sure. The article is good. Make sure to teach the kids not to let in Trojan horsies. And that paragraph above made my giggle a bit. But you know freaking what? My Mac always has been and always will be virus free. I don’t need no stinking security software.

Hilarity

I don’t do short posts here much. Perhaps I should. I can’t remember whose link I first followed to “Guy in the Hat”, but I’m glad I did:

It’s almost like watching one pseudo-journalist giving another fellatio via email.

It’s funny coz it’s true. But it’s funnier because Gruber deserves it. Well, I think so, anyway.

Oh wait, does that mean I now have to swallow?

Free dictionaries and "Dict OS X"

It’s really annoying that there are hardly any free and good multilingual dictionaries. Except that there are loads of online ones. I don’t like online services much, mostly due to latency. But also because native clients have the potential to be simply better.

Speaking of which, all the dictionary programs I’ve seen are really ugly. The content available for these programs generally don’t take advantage of the fact that they have an effectively unlimited amount of space to show you as much information as you might happen to want.

Imagine a translation dictionary that gave you conjugations and declensions of every word you asked of it and showed it in use in multiple contexts. (If you wanted that much information, of course.) That’d be, like, really useful for learning a language.

Of course, it’d be harder than transcribing a print dictionary into a searchable computer program. And I’m guessing there’s not that much money to be made from dictionaries at the best of times. But it would make a good free project/labour of love for language geeks.

Well, we gotta live with what we got. And I found it quite a pain to get dictionaries locally installed that can be accessed through a native application in Mac OS X.

Not knowing where to start, OmniDictionary seems pretty ideal to begin with. By all accounts, it’s “easy” to set it up to grab local dictionary content. But Jens Ropers’ DICTataro, recommended by Omni, is down at time of writing, so no luck there. Trying to manually set up a local dictionary server was a dismal failure. The links at dict.org are all rather old (i.e., nothing is actively maintained) and the Java programs listed there simply bailed when I tried to run them.

I finally had some luck with Dict OS X. The 0.1 version number and 2004 release date notwithstanding, at least it works. It comes by default with an English dictionary, and English to Czech and vice versa translation dictionaries. (Instructions for the program itself are all in Czech but luckily self-evident.)

More dictionaries can be downloaded from freedict.org. After downloading, say, the English to French files, there are a few steps to get the dictionary into Dict OS X:

• Uncompress the eng-fra.dict.dz file. This is most easily done by renaming it to eng-fra.dict.gz and opening the file; Mac OS X’s extraction tool does the job.
• Delete the compressed file and rename the containing folder to something like “English to French.dictosx” to turn it into a bundle (the name you chose is what appears in Dict OS X’s interface). It should contain the .dict file and a .index file of the same name.
• Finally, move the newly created bundle to ~/Library/Application Support/Dict OS X/

Phew. And once you’ve gone through the tedium of doing all this, you’ll sadly discover that the dictionary is woefully inadequate:

That’s right. There’s no gender information for the nouns in this dictionary. It’s better than nothing, at a pinch. But I wonder what it would take to get a free dictionary like this off the ground. Because we certainly need something better.

‘Yahoo!’: all about the name

Many thoughts around the place on Yahoo!’s probably acquisition by Microsoft for forty billion dollars. I like Gruber’s take.

I think the way to analyse this deal is by looking at what Microsoft gets out of buying Yahoo. Does Yahoo actually provide any products that Microsoft doesn’t? By and large, no: Microsoft has been bleeding money trying to replicate Yahoo’s online presence. Search/portal/community — there’s nothing that Yahoo has in material terms, give or take, that Microsoft doesn’t already (try to) do.

Except for the whole customers thing.

This is more like Google buying YouTube: after trying and failing to complete with the largely more successful product, the easy way out is to buy the mindshare. And you can’t keep the mindshare without retaining the original product.

So Microsoft can ditch its own attempts at creating a Yahoo knock-off and simply reap the rewards of owning the Yahoo name. The seemingly obvious thing to do with the acquisition is to keep the Yahoo stuff that makes money (search, email, Flickr, …), close down the competing areas that Microsoft does better (I guess that’d be the Zimbra thing that people keep mentioning), and watch the cash roll in. Or something like that.

As a non-Yahoo and non-Microsoft devotee, it doesn’t make any difference to me. (Oh, I’m a casual user of del.icio.us, but I had no idea they were owned by Yahoo and I can give it up without losing any sleep.) I hope for the sake of it’s fans that some semblance of Yahoo is left after the merger. Surely things couldn’t turn out as badly as everyone is saying?

To-Do management: “Things”

The bane of productivity is multi-tasking. Context switching in your brain slows down pretty much everything you think about and can make it impossible to dig deep into the hard problems. That’s why Richard Stallman doesn’t use a web browser and Don Knuth doesn’t even use email.

But multi-tasking is pretty damn addictive. It’s not something that most people have the will-power to give up — nor would they want to. Outsourcing your memory to a computer is an easy step to free up some mental capacity for problem solving.

Now, I’ve never been much of a ‘diary’ person to keep track of things I have to do. Generally my brain’s been good enough for most things. The extent of the project management in my various endeavours has been a flat list of flagged emails in Mail.app. One obvious problem with this is that tasks for which I don’t receive email don’t receive as much attention. And the non-hierachical system I set up makes it easy to lose the one important one you’re looking for in a big list. (For similar reasons, Mail and iCal’s revamped To-Do list in Leopard is not really the answer to my problems.)

This is one of the reasons I’ve never had a very good system for organising my long-term goals. Especially for my PhD, to be honest. However, with my new, ubiquitous, MacBook, I’ve the liberty of exploring software to address this personal deficiency of mine. Judging by the plethora of software that’s emerging in this space, I’m not the only one.

Now, ages ago I once used OmniOutliner to try and keep something of a project management ‘outline’, but it’d didn’t stick so well in my mind; it’s pretty essential that the application that you put all your forward-thinking thoughts into is easy and enjoyable to use. If there’s any mental burden just to get started, it will never survive as a long term commitment. Bearing this in mind, let’s look at the first app that I’m trying out as a candidate to fill that particular gap in my brain: Cultured Code’s “Things”.

First thing’s first: I’m not a fan of the name. It’s about the least Google-able name of any software ever. People are already resorting to calling it ‘Things.app’ just to be able to differentiate it from, well, normal sorts of things. But it’s certainly a name that strikes a chord amongst Mac users because it sounds like a name that Apple itself would use.

They claim this is an alpha-release preview; despite this, the interface is quite polished and the application seems very stable. Having said that, I do hope that some refinement to the interface is made to simplify it some more.

But let’s start off with the good things. The killer feature of this program is being able to set repeating schedules of To-Do items. I’ve been wanting this for as long as I can remember. “Mop the floor” repeat every second Sunday. Chores are the very first thing that I forget/neglect to do when I’m having a Sunday at home. Note, however, that this feature doesn’t exist yet. But I’m predicting that it will be awesome based on their recent comments on the feature.

Second up, the workflow of the program has been exceptionally well design; adding and deleting To-Do items and projects is as easy, often, as hitting the space bar. There’s also a system-wide heads-up display from which you can add a To-Do whenever. And did I mention that the program is very attractive? Fluid to use and handsome looks makes an app that you like to spend time, or ‘nest’, in. So far so good.

Now the application gets a little complex to describe in words. To-Do items themselves can be arranged in terms of their focus: today, next, later, and postponed. This is a good breakdown, I feel. Granular enough to segregate the items that don’t belong together, but coarse enough that you don’t lose items in categories that are too fine.

This is the first layer of metadata that can be assigned to To-Do items. The second layer of metadata that “Things” gives are: projects and areas. A project is something to be completed that requires many To-Do items. An area (of expertise) is the broad categories that your projects fall into. I might be using “Things” to track my both PhD and Typesetting work, which each have various projects attached to them.

Here’s the kicker: each project can only belong to a single area of expertise, and yet in the organisation of the application, projects and areas are treated roughly as orthogonal. Take a look:

In the sidebar there’s a list of projects in an ungrouped list. In the main window, the same information is presented but this time divided into categories.

In my opinion, this duplication of information reveals an underlying flaw in the interface of the program. Projects should exist as sub-categories of Areas, which would help keep the projects logically separate in the sidebar and eliminate the redundancy — while simplifying the interface.

In brief, here is the current design in full ASCII glory:

ORGANISE
  ▾Projects
    WSPR 
    Thesis
    Experiments
    LM fonts
  ▾Areas
    Web
    Chores
    PhD
    Typesetting

And here is my proposed design:

AREAS
  ▾Web
    WSPR
  •Chores
  ▾PhD
    Thesis
    Experiments
  ▾Typesetting
    LM fonts

This efficiency of the list is only slightly improved, but the duplication of information in the main window is no longer necessary. Fewer equivalent ways to view data equals less confusion.

One of the reasons this interface should be simplified is because there’s a third layer of metadata that can be applied to all of these To-Do items: Tags. These provide an arbitrarily extensible way to group To-Do items into smaller and smaller sub-groups, if you so desire.

Tags allows you to filter items into arbitrarily granular (and non-exclusive) categories. A GTD-type approach is suggested to use tags for contexts like Home/Work/Phone/Whatever. (I’m not a GTD kind of guy, by the way.) Alternatively, it might be impractical to list every single Ruby script or LaTeX package you maintain as an individual project; tags allow you to do this without cluttering the interface of the main window.

I can’t do the tags feature justice with words, but suffice it to say that tags in “Things” are Done Right. And they will stay completely out of your way if you’d prefer to keep things simple. This is extremely important in hooking people in with a simple-seeming application that hides its power in its flexibility.

My only other complaint after a short time of using “Things” is the button bar at the bottom of the main window. I loathe buttons I can’t get rid of, even more if they’ve got labels that I can’t even hide. (For example, now that I’ve assigned Control+Space to quick-add To-Do items I guarantee I’ll never click that button ever again.) I think it’s very important to have scalable interfaces, where in this case the scalable refers to being able to hide away the complexity of unnecessary features after you’ve learned the shortcuts or decided to mentally pass on that particularly feature of the application.

To wrap up this mini-review, I’m optimistic about things. No, sorry. I’m optimistic about “Things”. Time will tell if the application is sticky enough for me to nest in it over time and create the extension of my brain that is really the goal of this sort of program. I think it will be. I’m hoping that with feedback from the community before the ‘v1.0’ release, things will only get better. Oh, sorry again. “Things” will only get better.