I don't feel this study addresses language at all; it only addresses technology. Sure, their language has no number bigger than two; but that just means they can't really *count* higher than two. If you'd gotten an adult english speaker to do the same tasks, but with numbers of objects that were too large to count in the given time, the results would probably be the same. I mean, if someone flashed you a box with 83 batteries in it, would you "remember" later, unless you counted them each time? Recognizing 10 objects is not like recognizing a flower, justice, or murder. It's based on a technology.
Of course, though counting technology is based on number, there are more sophisticated ways of using it besides rote one-two-three counting. Since we're a numerate society, we train children to count and map sets to each other and all that fun stuff early on, and drill it into their heads. I remember when I was a kid doing these things, and learning to quickly recognize, say, five objects, by grouping them into 2 and 3 objects, and sort of "imprinting" what those look like in different configurations. I just "know" what five things look like now, the way a chess player "knows" what a winning configuration looks like, or the way a chef "knows" when a chicken is perfectly cooked. I was never trained to recognize 83 batteries, so I can't, any better than people whose lives have no use for counting technology when they see four or five batteries.
Number words are just rosary beads, and people in numerate societies use them all the time. These tribesfolk simply don't, just like they don't write, program computers, drive cars, or bake souffles, but you could train them, even in their own language, using existing words for the numbers, and the difference would go away, I bet. Now *that* would be an experiment.
>If you have this much code, I bet there's some duplicated code in there.
This is the second time in the thread (which is very informative, thank you everyone, I love you all, mwa) that someone has said that if I have 300K lines of code (it's Java), there's probably something wrong. This is a non-sequitur to me, and reinforces something I've wondered about for a long time-- do people just write tiny programs out there?? It reminds me of the time I was working on a different project using C++ Builder-- after the system was complex enough to have about 300 classes, many of them involving a lot of templates, the linker kept seg faulting. Their tech support asked me how many classes I was trying to link, and when I said 300, he just laughed at me incredulously. I don't think 300 classes is a lot to expect a tool to handle. The prorgram was about 1/3 done at that point. Our system has 1700 classes and a feature set so rich that it's sometimes hard to believe it's *only* 300K lines.
For the record, we refactor all the time, and we do rapidly develop "stories" quickly trotted out for customer approval or the garbage can, and there's no "this is my code and my domain" mentality here, so we're doing a good bit of XP-ish stuff already.
Anyway, I'm not trying to get into a macho My-program-is-bigger-than-yours thang, but I do sometimes wonder how pragmatically useful some methodologies and tools are once you get beyond web sites with a shopping cart bean that you have 7 people working on.
Sorry! Venting! Continue with useful comments.
Slashdot Mirror
I don't feel this study addresses language at all; it only addresses technology. Sure, their language has no number bigger than two; but that just means they can't really *count* higher than two. If you'd gotten an adult english speaker to do the same tasks, but with numbers of objects that were too large to count in the given time, the results would probably be the same. I mean, if someone flashed you a box with 83 batteries in it, would you "remember" later, unless you counted them each time? Recognizing 10 objects is not like recognizing a flower, justice, or murder. It's based on a technology. Of course, though counting technology is based on number, there are more sophisticated ways of using it besides rote one-two-three counting. Since we're a numerate society, we train children to count and map sets to each other and all that fun stuff early on, and drill it into their heads. I remember when I was a kid doing these things, and learning to quickly recognize, say, five objects, by grouping them into 2 and 3 objects, and sort of "imprinting" what those look like in different configurations. I just "know" what five things look like now, the way a chess player "knows" what a winning configuration looks like, or the way a chef "knows" when a chicken is perfectly cooked. I was never trained to recognize 83 batteries, so I can't, any better than people whose lives have no use for counting technology when they see four or five batteries. Number words are just rosary beads, and people in numerate societies use them all the time. These tribesfolk simply don't, just like they don't write, program computers, drive cars, or bake souffles, but you could train them, even in their own language, using existing words for the numbers, and the difference would go away, I bet. Now *that* would be an experiment.
>If you have this much code, I bet there's some duplicated code in there. This is the second time in the thread (which is very informative, thank you everyone, I love you all, mwa) that someone has said that if I have 300K lines of code (it's Java), there's probably something wrong. This is a non-sequitur to me, and reinforces something I've wondered about for a long time-- do people just write tiny programs out there?? It reminds me of the time I was working on a different project using C++ Builder-- after the system was complex enough to have about 300 classes, many of them involving a lot of templates, the linker kept seg faulting. Their tech support asked me how many classes I was trying to link, and when I said 300, he just laughed at me incredulously. I don't think 300 classes is a lot to expect a tool to handle. The prorgram was about 1/3 done at that point. Our system has 1700 classes and a feature set so rich that it's sometimes hard to believe it's *only* 300K lines. For the record, we refactor all the time, and we do rapidly develop "stories" quickly trotted out for customer approval or the garbage can, and there's no "this is my code and my domain" mentality here, so we're doing a good bit of XP-ish stuff already. Anyway, I'm not trying to get into a macho My-program-is-bigger-than-yours thang, but I do sometimes wonder how pragmatically useful some methodologies and tools are once you get beyond web sites with a shopping cart bean that you have 7 people working on. Sorry! Venting! Continue with useful comments.