Slashdot Mirror

made by a nice girl

pic: http://web.media.mit.edu/~vmb/gauri.jpg
homepage: http://web.media.mit.edu/~nanda/

could this idea only come from a girl..?

pic: http://web.media.mit.edu/~vmb/gauri.jpg
homepage: http://web.media.mit.edu/~nanda/

could this idea only come from a girl..?

If the inventor is around, I bet I will never sleep. Geek girl folks .. there is a hope for all of us.

I read this a while ago. I found a pdf of the New York Times article here http://msl1.mit.edu/furdlog/docs/nytimes/2004-12-0 7_nytimes_poor_english.pdf . Cogent point: Businesses are having to send people back to school to learn how to write. Many cannot write coherent sentences or paragraphs. Just try to deal with requirements when not only do they not have a clue, but can't even write their non-clue.

Another Wikipedian, a developer with Yahoo India, mentioned this on the Wikipedia village pump last night. Being that I handle the full-length music uploads almost single-handedly (you can see my progress here) I went and eagerly tried it out. The result was very disappointing. I searched for about 20 different songs on my wishlist (at the bottom of my user page. Most of the hits were mutopia MIDIs or bizticket e-donkey links --- eg, useless. So I search for the songs + (Ogg OR mp3). The only useful hits were to the Internet Archive and to the MIT free music site, both of which I have thoroughly plundered. So like I said, this was a sizeable letdown.

I fear the Candiru Vaudellia Cirrhosa more than I fear piranhas

Not sure what you are trying to say, but copper is the best pure metal conductor. Super conductors are, of course, much better. As far as corrosion, all the elements corrode and oxidize (STM experiments tell us that even gold forms an oxide mono-layer in about 10^-9 seconds in air). But gold and platinum corrosion are not as bad on connections. In dry environments, I'm not sure copper is noticably different within 10 years.

Its roots are in MIT, a much bigger AI project.

I bet some guys from there were/is affiliated with that lab.

http://start.csail.mit.edu/

Its there since '93 ;)

The thing that is always lost in discussion of educations, is the first most important variable. Whether education succeeds or fails depends for the most part on the desire of students to learn. Good teachers or good computerized courses and software can promote the desire to learn, and can dramatically enhance the learning process, but in the absence of desire these things aren't likely to make it happen. The key reason American education seems to be failing is most American's have a low desire to learn. Most High School students place way more value on athletics, activites and social strata than they do on the education component.

If students don't see the value in education and don't want to learn next to nothing is going to help.

One motivator for learning is hunger. If students see education as a route to success, earning a living and putting food on the table chances are the will be motivated.

Another motivator might be if you are part of a successful, educated family your family may instill in you the value of it (i.e. you live like you do because your ancestors valued education, skills and hard work).

If you couch education as it is to most Americans:

A. We are rich, we will continue to be rich whether we work hard or not

B. We are poor and we are going to stay poor no matter what we do

education comes across as a waste of time and it fails.

With MIT's open courseware if you have the time, the desire and some basic intelligence, I'd say chances are high you can probably educate yourself these days.

"Once those influences were eliminated, the relationship between use of computers and performance in maths and literacy tests was reduced to zero"

Then, "The more access pupils had to computers at home, the lower they scored in tests, partly because they diverted attention from homework"

Well? Which is it, zero or negative correlation?

Read it for yourself.

The gist of it seems to be that they estimated several measures of the affect. Any which were significant were negative.

Humans think of the world using language, but we also think of the world using visual, spatial, temporal, sensorial... reasoning.

Do ask a real psychologist, she will say that there are different kinds of think. Textual is best suited for abstract, logical reasoning. But associative thinking is often better done visually. In the Programmer's Guide to the Mind you have an interesting classification of all these.

A programming environment should take care of all these kinds of thought, not just support the logic abstractions as they do now. A promising field of research is Programming By Example. This programming style tries to build the final program by using concrete reasoning over samples of data, instead of forcing you to think of the general, abstract procedure.

Humans think of the world using language, but we also think of the world using visual, spatial, temporal, sensorial... reasoning.

Do ask a real psychologist, she will say that there are different kinds of think. Textual is best suited for abstract, logical reasoning. But associative thinking is often better done visually. In the Programmer's Guide to the Mind you have an interesting classification of all these.

A programming environment should take care of all these kinds of thought, not just support the logic abstractions as they do now. A promising field of research is Programming By Example. This programming style tries to build the final program by using concrete reasoning over samples of data, instead of forcing you to think of the general, abstract procedure.

Programming By Example

There are some lectures at MIT World http://mitworld.mit.edu/

I downloaded MIT's Structure and Interpretation of Computer Programs (SICP) lectures from here, and converted the audio portion to MP3 so I could play it on my iPod. Outside of that suggestion, I have the same question as you. I recently cancelled an Audible subscription because I had run out of books I was interested in hearing. The local library seems to have mostly fiction in audiobook.

Mach RPC is built on top of Mach ports, but an RPC call is not a single kernel-level operation. So you have the scheduling problem.

Interestingly, although it's very obscure, the kernel call level below mach_msg_send actually does support a combined send/block/receive operation. As the MIT document points out, this "enables certain internal optimizations".

But when you read the GNU HURD tutorial on Mach IPC, you see the suggestion that a program intended to send an integer to another program and wait for a reply should do two separate operations. Somebody doesn't get this.

Also, for no particularly good reason, Mach usually uses the same buffer for both send and receive, so if you use the combined send/receive form, whatever you send gets clobbered by the reply. So using this tends to involve an extra copy.

Mach IPC is incredibly complicated, You have to fill in many data structures just to send something. The implementation is complicated, too. QNX has a nice, simple primitive:

int MsgSend(int connectionid,
const void* sendmsg,
int sendbytes,
void* rcvmsg,
int rcvbytes);

You can probably figure out how to use that just from the definition. Even the reply value is what you think it should be: the receive count if it worked, -1 if it failed, with an error code in errno.

The name of the game here is getting the primitives right.

Because I am a big fan of it, I would like to take a moment to plug MIT's OpenCourseWare, where you have access to MIT's entire course catalog, including assignments, videos, and other materials. Want to learn Japanese? Go for it. Or perhaps Electromagnetic Fields, Forces, and Motion is more to your liking. Have at thee.

MIT has shown their dedication to an open academic atmosphere and the benefits for the public of easy access to knowledge. Their endorsement of free software here is completely consistent with their previous actions.

Good for them.

Because I am a big fan of it, I would like to take a moment to plug MIT's OpenCourseWare, where you have access to MIT's entire course catalog, including assignments, videos, and other materials. Want to learn Japanese? Go for it. Or perhaps Electromagnetic Fields, Forces, and Motion is more to your liking. Have at thee.

MIT has shown their dedication to an open academic atmosphere and the benefits for the public of easy access to knowledge. Their endorsement of free software here is completely consistent with their previous actions.

Good for them.

Because I am a big fan of it, I would like to take a moment to plug MIT's OpenCourseWare, where you have access to MIT's entire course catalog, including assignments, videos, and other materials. Want to learn Japanese? Go for it. Or perhaps Electromagnetic Fields, Forces, and Motion is more to your liking. Have at thee.

MIT has shown their dedication to an open academic atmosphere and the benefits for the public of easy access to knowledge. Their endorsement of free software here is completely consistent with their previous actions.

Good for them.

Maybe they should work with the MIT Fab Lab people (see their overview, or the Wired article by Bruce Sterling).

Maybe they should work with the MIT Fab Lab people (see their overview, or the Wired article by Bruce Sterling).

Do you see anything in the MIT EECS curriculum that looks remotely like that?

Also, there are several pages of tutorials for OpenOffice.org

AbiWord and OpenOffice.org both support the OpenDocument (XML) format as well as their own XML-based formats. OpenDocument is being favored by the EU and developing nations have even more benefit from it as it alllows more choice in software and platforms down the line.

You're thinking of the Lotus 123 case.

The difference was that Lotus didn't have a patent, they only had copyright, which as you rightly point out doesn't cover look and feel.

This time, though, Apple have a patent for the graphical design which means they may well be able to successfully sue those who copy the look and feel of their interface.

Which IMHO just goes to show how dumb patent law is these days, but hey, everyone's doing it so it might be right. Right? :-(

Totota and Hitachi got nuthin' on MIT's Cog and Kismet

I foresee a fight scene ala Anchorman; Cog wielding a switch-blade.

"Como estas, bitches!"

You think the only way to make a non-trivial word processor is to copy the UI style and feature set of Word very closely?

In a world where training is expensive, yes. Imagine if the Sholes (QWERTY) layout of the computer keyboard were copyrightable. Please read this document published by the League for Programming Freedom.

Show me one example where O'Caml beats C++.

ICFP contests:

2004 (ocaml 1st); 2003 (ocaml 1st); 2000 (ocaml 3rd, C++ 5th); >2001 (ocaml tied 3rd place with C);

Others winners: 1998 C derivative, 2002 python (I think C++ actually beat OCaml this year)

Your link doesn't work. Try this instead: segmentation

As far as I'm concernet, I think it was over used in class rooms. Particluarily by teachers who don't understand that it was really only meant as a jumping off point, not the end goal (and a "proof of concept" that childeren could wrap their heads around the concepts of programming a compuer).

So use it that way in your demonstration:

1. Here's the how the turtle moves
2. Here's is how I get the turtle to do this complex task "foo" (whew, that was a lot of work to do every time I want the turtle to do "foo")
3. Here's how I avoid all that work every time (show them the program listing)
4. Run the program serveral time to show them how easy it is now (alla George Jetson - button pusher)
5. Show them a more complex task like that

In reallity, you'll have to use the techniques that a teacher uses to teach their elementary students. Learning on the edge of what they already know.:

1. Have a brain storming session. What do they already know about Software Engineers and Programmers
2. Show them some open-sourced video game running (As a teaser)
3. Show them logo (as above)
4. Let them ask questions
5. Show them the open-sourced video game again, and show them the code behind it
6. And if it's not toooo borring, show them what you are working on right now at work - with your boss's permission of course

Even better: StarLogo.
Runs in Java, you can spawn hundreds of multicolored logo turtles and make them all move in sync (or randomly) using the same 5-line logo program from the console - type in a line and watch the turtles obey.

Plus, they have ready-made projects (Click on Projects from the main page) that are all set to go, simple and super-cool visually, from "Slime-mold cells aggregate into clusters, using a chemical pheromone" to "Diffusion Limited Aggregation. Create fractals with simple rules".

Thats what I used in a similar situation - the kids LOVED it.

This seems obvious to me... Logo Primer

Kids these days.

"html sucks for equations"

Not sure what your standard is for "sucks" but MathML works reasonably well and is supported in Firefox/Mozilla, though you need some particular fonts.

Some of the MIT OpenCourseWare like, Calculus makes fairly extensive use of fairly involved formulas.

I wish it worked on Konqueror, maybe it does, but it didn't last time I tried it.

Probably would be nice if this stuff was supported in a more standard way in Linux/Firefox distributions too, without having to rummage around.

It would be great to see stuff like MathML and the MIT course ware get a lot wider exposure, especially to young people. Maybe it would counter the general bad education system, especially in the U.S.

At least one machine learning journal, JMLR , has already moved to a free online distribution method. There is an interesting story here about how the editors of a previous journal (MLJ) quit en masse due to dissatisfaction with the publisher, and started JMLR.

At least one machine learning journal, JMLR , has already moved to a free online distribution method. There is an interesting story here about how the editors of a previous journal (MLJ) quit en masse due to dissatisfaction with the publisher, and started JMLR.

It's also hfsutils (also GPLd) Compare CherryOS (grep's output): C:\CherryOS\chktarakh.ctq: error reading MacBinary file header C:\CherryOS\chktarakh.ctq: no error C:\CherryOS\chktarakh.ctq: error writing data to http://web.mit.edu/afs/athena/astaff/project/hfsto ols/src/hfsutils-3.2.6/copyin.c

In the field of AI we have at least two highly respected journals which do not have paper editions (even though libraries can buy bound collections of papers on a yearly basis) and which make their content available for free to everyone:
Journal of Artificial Intelligence Research and
Journal of Machine Learning Research

This works because an academic journal does not really have any expenses for peer-review. Academics review for free as part of their job - it gives status to review for a prestigious journal. If you don't have any costs for editing and printing a paper edition, you suddenly have almost no operating expenses at all. Cost of bandwith is negligible. A typical research paper in pdf format is a 100k download, so any one of us could operate one of those servers from our home. Furthermore, the cost of bandwith is continually decreasing.

In sum, I don't understand what the IEEE is whining about. Let those who want a journal on paper pay for the paper, and let the rest of us have it for free!

In the machine learning community, one of the most important journal is JMLR http://jmlr.csail.mit.edu/. From the beginning, this paper decided to go free, online. As a results, the time to publish were very small, and since reviewing was very strict, the paper quickly gained a high Impact Factor. Now it appears that JMLR is also published as paper volumes. I don't know about their economic model, but surely this success-story shows it is feasible to publish scientific free journals.

Furthermore, many authors (like me) do post a copy (called 'draft' for copyright reasons) of their paper on their webpages. Sometimes some googleing avoids having to pay for scientific journals...

You want LUNAR. It's especially cool because it uses an underlay network called selnet (ARP forwarding instead of straight-up IP routing). Also, there are a bunch of normal layer 3 ad-hoc multihop protocols designed especially for highly dynamic/mobile nets that you can install for free (I can verify that they work on 2.4 kernels, anyway):

NIST AODV
unik OLSR
US NAVY labs OLSR
CLICK modular router (contains a DSDV and DSR implementation, provides a framwork for rapid prototyping of stack behaviors)

These all might be nice for a smallish office as a way to extend and enhance the probable coverage area of the network without getting more APs.

If it comes from a real school, a Master of Engineering is a very good degree. For example, MIT offers Master of Engineering. http://cee.mit.edu/index.pl?id=4629&isa=Category&o p=show

You fail to mention your Federally-guaranteed student loans and cheap mortgages. You don't bitch about the highways that whisk you to your suburban home or the miracle drugs discovered thanks to government research.

Are these the federally-guaranteed student loans that American students need only because the average student would otherwise have to pay $40,000 a year? Tuition at Oxford, for an EU citizen, is less than £1,500, which the LibDems want to eliminate. Tuition at McGill, for a Quebecois citizen, is something like CDN$1,700, and for any other Canadian citizen it's $4,012.50. I'd give up my student loans in a second for tuition fees like that.

Cheap mortgages? A three-bedroom house that would CDN$150,000 in Toronto costs well over $500,000 in Boston. The only thing you can get in Boston for less than $200,000 is a studio apartment (which in Toronto would cost CDN$100,000 or so). Whatever the interest rates are, I think it's cheaper to buy in Canada. (Certainly, Idaho is cheaper still, but I'm talking about places where people actually want to live.)

And actually, I bitch about the highways a lot. The automobile is, in my opinion, the most destructive machine that has ever been invented. Rural communities are farming towns, and everything else is urban; suburbs are just really low-density cities, and as such, they're eating up all our land and resources. When cheap oil ends, and people can't afford to drive anymore, the world will be infinitely better off.

Lest you think that I only criticize, I do indeed have solutions. Ban cars from all major cities, and in their place build streetcar tracks. Stop building new cities on grids, and instead do something like this. Make buildings at least four stories taller but not more than six, because anything taller becomes overwhelming, but anything shorter (like most single family homes) is wasteful and discourages community. End single-use zoning, so that businesses and residences are side-by-side, and so that no one is ever more than a five minute walk from some kind of gathering-place or more than a ten minute walk from a grocery store.

But don't listen to me, I'm just a small-minded vapid windbag. Take the highway back to your cheaply-mortgaged suburban home and enjoy your life of isolation.

Rodney Brooks at MIT has done quite a bit of research in the past in this area quite a few years ago. It seems that the links regarding his projects are currently broken, but do a bit of googling, I'm sure you can find his papers on the subject.

Cambrian Intelligence is pretty good book that covers his techniques for AI in robotics. It's essentially a collection of eight early papers by Brooks.

During our discussion, important value and ethical issues arose in the design and use of such "relational robots." These robots are meant to create bonds and simulate "authentic" relationships. They react to voices, track their owner's eyes, respond and project emotions, and so on. Yet, they remain robots - all these actions and reactions are programmed - pre-determined. So, how do the designers decide what emotions to program and which to omit? In an effort to be realistic, My Real Baby gets happy as well as sad. If you bounce her when she's happy, she gets more happy; if you bounce her when she's fussy, her fussiness only increases. How should she react, then, if she is abused? It is not hard to imagine a child (especially one who is herself a victim of abuse) to violently shake, strike or otherwise "abuse" the doll. How should this evocative object respond? Should she show pain? Begin to cry? Eventually "pass out" or even "die" if the abuse continues? How "real" should the robot be in order to create an "authentic" relationship?

[In the end, the designers wanted the doll/robot to react as a child would, with pain and sadness. However, the company's lawyers stepped in and were concerned that any type of response by the doll might encourage further abuse (stimulus-response theory), and they didn't want to be accused of actually encouraging abusive behavior. In the end, the doll simply did not react to abuse.]

Other ethical dilemmas related to the design of such robots included whether they should be capable of deception or betrayal, two common features of human relationships. Or, should they "die." On one hand, the experience of death as part of the life cycle is an important part of psychological development and would add to the "authenticity" of the relationship. On the other hand, one of the benefits of these robots seems to be the avoidance of the emtional damage that can happen when a "real" companion (whether a human friend, or even a companion dog) dies.

During our discussion, important value and ethical issues arose in the design and use of such "relational robots." These robots are meant to create bonds and simulate "authentic" relationships. They react to voices, track their owner's eyes, respond and project emotions, and so on. Yet, they remain robots - all these actions and reactions are programmed - pre-determined. So, how do the designers decide what emotions to program and which to omit? In an effort to be realistic, My Real Baby gets happy as well as sad. If you bounce her when she's happy, she gets more happy; if you bounce her when she's fussy, her fussiness only increases. How should she react, then, if she is abused? It is not hard to imagine a child (especially one who is herself a victim of abuse) to violently shake, strike or otherwise "abuse" the doll. How should this evocative object respond? Should she show pain? Begin to cry? Eventually "pass out" or even "die" if the abuse continues? How "real" should the robot be in order to create an "authentic" relationship?

[In the end, the designers wanted the doll/robot to react as a child would, with pain and sadness. However, the company's lawyers stepped in and were concerned that any type of response by the doll might encourage further abuse (stimulus-response theory), and they didn't want to be accused of actually encouraging abusive behavior. In the end, the doll simply did not react to abuse.]

Other ethical dilemmas related to the design of such robots included whether they should be capable of deception or betrayal, two common features of human relationships. Or, should they "die." On one hand, the experience of death as part of the life cycle is an important part of psychological development and would add to the "authenticity" of the relationship. On the other hand, one of the benefits of these robots seems to be the avoidance of the emtional damage that can happen when a "real" companion (whether a human friend, or even a companion dog) dies.

Created in the shadow of Mt. Hood: http://www.swiss.ai.mit.edu/~bob/hood.jpg

Sigh. Everything was so simple and clean back then.

All these new office suites make me feel depressed, and they make baby Jesus cry.

:(

Created in the shadow of Mt. Hood: http://www.swiss.ai.mit.edu/~bob/hood.jpg

Sigh. Everything was so simple and clean back then.

All these new office suites make me feel depressed, and they make baby Jesus cry.

:(

A buddy of mine and I decided that the ringtones we really really really wanted were the music from isi and Final isi. This is no mean feat, however: our phones (Sanyo SCP-8100s) only play midi and some crappy .wav format, and the latter is clearly unacceptable (both because it's limited in length to 30 seconds and because... well... it's just not cool enough!).

So what did we do? Jim modified DOSBOX's OPL3 emulation code to dump out the opcodes being sent to the FM channels and handed over the output to me.

From there, it was a matter of parsing the various channel setup data into some semblance of notes, deciding which combination of general MIDI patches best emulated the sound of the FM synthesizer given the patchset on my phone, and writing a whole bunch of code.

In the end, we did it: isi.mid and fisi.mid are the full soundtracks to isi and Final isi, respectively. In addition, I made a couple other versions of the Final isi soundtrack to skip to various parts of it that are more interesting and/or make better ringers than starting at the beginning: fisi2.mid, fisi3.mid.

These ringers pretty much rule the roost.

-rsw

A buddy of mine and I decided that the ringtones we really really really wanted were the music from isi and Final isi. This is no mean feat, however: our phones (Sanyo SCP-8100s) only play midi and some crappy .wav format, and the latter is clearly unacceptable (both because it's limited in length to 30 seconds and because... well... it's just not cool enough!).

So what did we do? Jim modified DOSBOX's OPL3 emulation code to dump out the opcodes being sent to the FM channels and handed over the output to me.

From there, it was a matter of parsing the various channel setup data into some semblance of notes, deciding which combination of general MIDI patches best emulated the sound of the FM synthesizer given the patchset on my phone, and writing a whole bunch of code.

In the end, we did it: isi.mid and fisi.mid are the full soundtracks to isi and Final isi, respectively. In addition, I made a couple other versions of the Final isi soundtrack to skip to various parts of it that are more interesting and/or make better ringers than starting at the beginning: fisi2.mid, fisi3.mid.

These ringers pretty much rule the roost.

-rsw

A buddy of mine and I decided that the ringtones we really really really wanted were the music from isi and Final isi. This is no mean feat, however: our phones (Sanyo SCP-8100s) only play midi and some crappy .wav format, and the latter is clearly unacceptable (both because it's limited in length to 30 seconds and because... well... it's just not cool enough!).

So what did we do? Jim modified DOSBOX's OPL3 emulation code to dump out the opcodes being sent to the FM channels and handed over the output to me.

From there, it was a matter of parsing the various channel setup data into some semblance of notes, deciding which combination of general MIDI patches best emulated the sound of the FM synthesizer given the patchset on my phone, and writing a whole bunch of code.

In the end, we did it: isi.mid and fisi.mid are the full soundtracks to isi and Final isi, respectively. In addition, I made a couple other versions of the Final isi soundtrack to skip to various parts of it that are more interesting and/or make better ringers than starting at the beginning: fisi2.mid, fisi3.mid.

These ringers pretty much rule the roost.

-rsw

A buddy of mine and I decided that the ringtones we really really really wanted were the music from isi and Final isi. This is no mean feat, however: our phones (Sanyo SCP-8100s) only play midi and some crappy .wav format, and the latter is clearly unacceptable (both because it's limited in length to 30 seconds and because... well... it's just not cool enough!).

So what did we do? Jim modified DOSBOX's OPL3 emulation code to dump out the opcodes being sent to the FM channels and handed over the output to me.

From there, it was a matter of parsing the various channel setup data into some semblance of notes, deciding which combination of general MIDI patches best emulated the sound of the FM synthesizer given the patchset on my phone, and writing a whole bunch of code.

In the end, we did it: isi.mid and fisi.mid are the full soundtracks to isi and Final isi, respectively. In addition, I made a couple other versions of the Final isi soundtrack to skip to various parts of it that are more interesting and/or make better ringers than starting at the beginning: fisi2.mid, fisi3.mid.

These ringers pretty much rule the roost.

-rsw