Slashdot Mirror

Journaled file systems reduce disk write latency because new data (and metadata) can be written in one sequential write. Non-journaled file systems (like ext2) must move the disk head to multiple locations to update the data and then the metadata. For more info, John Ousterhout (creator of Tcl) has some good papers about journaling file systems, such as "The Design and Implementation of a Log-Structured File System (1991)".

Alasdair Rawsthorne has written a few technical reports on the technology behind this which includes some discussion on performance. See:

Machine Adaptable Binary Translation

The paper's a few years old, but I would definitely say this is not vaporware and I would expect them to have worked out most of their performance limitations since that time. Will it run as seamless as people here seem to expect? That remains to be seen. In the meantime, instead of immediately crying foul, read up on the subject and come to an informed conclusion. It looks promising. Search around, there's more up to date info.

The idea of machine code to machine code translation is not new. For example, Dynamo appeared in 1999, although in its case the source and target instruction sets were the same. Regardless of whether or not transitive is for real (I'm guessing it is), it's clearly just a matter of time before this technology appears in a practical form.

Not quite... Note: I'm about to karma whore here.

About a year ago, right before starting my senior year at UCLA, I was offered an opportunity to work on this P2P project. At the time it was called "Gnucla," and was being developed by the UCLA EE department's Complex Networks Group. I turned it down, because I had already committed to working on a p2p system in the CS department. But since in all honesty their research was more novel than ours (and my friend was in their group), I subscribed to their mailing list and kept informed on what they were doing.

What they've done isn't find a novel way of picking ultrapeers. Let's review what motivated ultrapeers -- in the beginning, there was Gnutella. Gnutella was a power-law based network. What this meant is that there was no real "topology" to it, unlike peer to peer networks that were emerging and based on Distributed Hash Tables (such as Chord, Pastry, Kademlia [on which Coral is based]). It had nice properties: a low diameter, and very resilient to attacks common on p2p networks. (Loads of peers dropping simultaneously could not partition the network, unlike, say, in Pastry -- unless they are high degree nodes.) But the big problem was that to search the network, you had to flood it. And that generated so much traffic that the network eventually tore itself apart under its own load.

So someone thought that maybe if only a few, select, high-capacity nodes participated in the power-law network, it wouldn't tear itself apart because they could handle the load. These would become the ultrapeers. The nodes that couldn't handle the demands of a flooding, power-law network would connect to ultrapeers and let the ultrapeers take note of their shared files, and handle search requests for them. Thus, when a peer searches, no peer connected to an ultrapeer ever sees the search unless they have the file being searched for, because the searching happens at a level above them. Between low-capacity nodes and ultrapeers, it's much like a client-server model. Between ultrapeers, it's still a power-law network.

But the ultrapeer network has problems in itself, so this group sought to find a way to search a power-law based network, such as Gnutella, without flooding. They exploited the fact that, in a power-law network, select nodes have very high degree connectivity. If you take a random walk on a power-law based network (meaning, starting from your own PC, randomly jump to a node connected to you, randomly jump to a node connected to that node, etc...) you'll end up at or passing through a node with very high connectivity. Thus, they were a natrual spot rendezvous point for clients wishing to share files, and clients wishing to download files. Perhaps, in this sense, they are an "ultrapeer," but we haven't separated the network into two different architectures like before. The network is still entirely power-law based, and retains all its wonderful properties.

But that's not the entire story, just the gist of it. There are other neat tricks to it... Trust me, this is really good stuff we're talking about here. They recently won Best Paper Award at the 2004 IEEE International Conference on Peer-to-Peer Computing. (See paper here.)

"Brunet," as they call it, is designed to be a framework for any peer-to-peer application that could exploit the percolation search outlined above. Google-like searching is just one possible approach (and perhaps a little unrealistic...). Right now I can tell you that they have a chat program in the works, and it is working well. The framework should be released when it's ready.

Please don't flood me with questions -- remember, I'm not actually in their research group :)

- sm

Looks like he might like Pastiche.

The XML community seems to be largely devoid of any knowledge of history or computer science depth. I have yet to find a description of XML schema processing in terms of grammars and parsers.

Hedge automata: a formal model for XML schemata

The brain damaged SAX parser has become popular, while few know about the XmlPullParser. Since many of those who use XML parsers don't seem to have ever parsed anything else, they do not seem to find it odd that the scanner should call the parser, rather than the other way around.

What does scanning have to do with XML parsing? The whole point of XML is that the tokens are predefined (elements, attributes, text nodes), so that the application writer does not need to deal with scanning at all. It is an internal detail of the parser.

If you are trying to say that the application should call the parser rather than the other way around then you are making some sense.

But the underlying problem is actually the programming languages we have available. The single-stack languages that are popular force either the parser or the application to NOT keep its local data on the stack. Since the first XML parsers were more complicated than the applications built on top of them, the parsers took the stack and the application had to deal with callbacks. Now parsers are commonplace and it is right that the roles should reverse and the complexity should be pushed into the parser. (and pull parsers are gaining in popularity)

But whose fault was it that we had to make the choice in the first place? The programming languages! Languages with coroutines do not have this problem.

Perhaps this lack of computer science depth is due to the fact that XML grew out of the dot-com bubble, when people felt they had no time to design or think about much.

Let's presume that's true. Then what is the excuse for the commercial DB world that Fabian Pascal claims is just as broken? I guess that things went wrong in the "minicomputer bubble"? Can you remind me of a time when the industry was moving slowly enough that people could be contemplative?

Internet2 p2p is in the works, read about LionShare notice that the vision includes sharing resources, sorta like a Grid.

I did my final year project / dissertation on delta compression and created a java web service & GUI to allow the distribution of delta files that users could download and apply.

It still requires a fair bit of work to make it very usable (hardly the best software engineering development ever), the Swing code is awful because I had to learn it in a week and it could do with some object serialisers on the data it returns. It worked ok though.

If anyone's interested they can read my report [PDF] (2.3MB). The point of doing that project was the reason that it is a technology that was massivly under exploited. It is quite limited for some things however, especially compressed archives and to a certain extent binary compiled files. However if you want to compress tars of source code it's brilliant and massivly improves over zip technology.

The package I used at the time was a java port of the xdelta project, javaxdelta. It had some bugs in it at the time however which meant that it didn't always work, I think from the discussion on their maling list that they've been fixed recently. I don't think it's as fast as the normal C++ xdelta implementation and xdelta as an algorithm isn't as good as some others, notably vcdiff and zdelta (see Suel & Memon, Algorithms for Delta Compression and Remote File Synchronization (2002))

I'm happy though, there may be some money making opertunities for my project =)

Emmm, what about this, for example?

Emmm, what about this, for example?

Emmm, what about this, for example?

The reason private industry isn't in space is that there's no money in it. Until that changes, we'll do nothing but probes - and no one ever comes up with an idea for making it profitable.

Actually, it's not that there's no money in space. It's that there's a chicken and egg issue. Why would anyone invest in space before there's an economic incentive? But who's going to create an economic incentive if there's no money in it?

That's where NASA comes in. NASA should be using large (and comparatively cheap) boosters to send hundreds of metrics tons of space station and workshop equipment into orbit. Once they have the ability to build many things in space (and launch from a decent orbit *grumble*), then the possibility of mining becomes a reality. I don't know of a single investor who wouldn't jump at the chance of bringing back precious metals at a high profit margin.

No, you don't need the level of industry we have here on Earth. (At least not initially.) More than enough can be done with standard workshop equipment. Drills, laser cutters, PCB printers, plastic molding, etc. all exist in fairly small machines. The tradeoff is that your production wouldn't be very high until better factories are built.

Now imagine this. On Earth, 3D printing is a very difficult and expensive process. Gravity ensures that stability is a key problem during the printing process. But what if we were to 3D print in freefall? With no forces acting upon the object, many more materials and shapes could be printed. Imagine if robots could be used to "print" an entire space ship! You just can't do that on Earth.

Two words. Breeder reactor.

Two words: Doesn't matter. A breeder reactor increases the amount of energy you get out of a given amount of materials. It does not, however, magically create materials out of non-nuclear fuels. It really amounts to further enriching existing fuels. Yes, it will extend the time. No, it will not extend it indefinitely.

100% pure pie-in-the-sky. Even if we could magically tap huge amounts of solar power, where do you put the antimatter once you've created it? And once you have it, so what?

There's no "magic" to tapping the sun. You need heat on one side, and cooling on the other. That's how thermal generators work. My original plan was to put a station at about 0.1 au. Speaking with engineers about the issue, they explained that using mylar mirrors would provide you with the same amount of thermal energy at a much greater distance. There's an interesting discussion on this here.

BTW, you may find this, this, and this very interesting. Yes, that's right. NASA plans to launch an antimatter powered space craft!

According to this (pages 35 and 36), around 74 bits for symmetric and 1024 bits for asymmetric key cryptosystems works right now. An old paper, but still the benchmark for sufficient key lengths in cryptosystems.

Moral of the story: You may need to upgrade to 2048-bit PGP keys in the long run.

Well, his resume is NOT extraordinary.

actually, his
CV does look quite amazing.

Is it normal for companies like SGI and Cray to have `oil and gas industry analysists' or a
'Director of Complex Scientific Computing Projects'? The only site referencing the latter term is his, not SGI's.

I'm not saying the CV is a fake, but some of it I just don't understand. For someone that has worked at Cray's Research departement and was the aforementioned director at SGI he hasn't published a lot. Citeseer doesn't know him, neither does DBLP. To be fair, publications in french wouldn't be indexed. Then again, he (allegedly) worked for american companies.

Google has a few pages with his emailadres @cray from '95, but that doesn't say the address is correct. AFAIK Cray has never had a french research departement.

So, is this harmless resume embellishment or more? Hopefully I'm just getting paranoid. Roland, if you're listening, perhaps you can explain (or update your CV).

This is the Stanford spinoff of the Polaris project presented at IEEE InfoVis several times over the past few years. Chris Stolte was the main student involved, and a CiteSeer search on his name will turn up most of the related work. To summarize: The goal of the work was to provide a visual programming environment (using a spreadsheet-like layout) to presenting data in multi-dimensional databases. It uses some sound (a.k.a., proven) results to create initial intiutive mappings of this data. See the papers for more details.

I'm not really that into operating system history, but at least a few of those things you listed aren't innovations.

Async I/O? Can we say UNIX sockets or iostreams?

fully preemptible kernel with fine-grained locking: A quick search turned up this: http://citeseer.ist.psu.edu/eykholt92beyond.html
so solaris definately had it in 92, if not earlier.

support for HT: wtf? now it's innovation to support a chip's features?

Actually, none of the things you have listed are innovations. You sir are a troll.

Indeed, Bloom Filters are the shit.

These days, in my spare time, I'm writing a p2p program -- think of it as a swarm-download system, like BitTorrent, on an overlay network topology, like eMule (only eMule uses Kademlia, and I'm using Pastry). It has been shown, here and here, that Bloom Filters can drastically reduce the traffic generated when searching peer to peer networks. I recently coded a Java implementation of a Bloom Filter for my p2p program, and it works great in testing. (But the p2p program isn't anywhere near done, so don't ask about it ;)

Furthermore, Bloom Filters can be compressed -- see Michael Mitzenmacher's work here. The idea that you can compress a Bloom Filter is a little counter-intuitive, because the size of the bit vector and the number of hash functions are derived using calculus to maximize the compactness of the set, for a given false positive rate -- thus, in this state, it is non-compressable (it is "already compressed" by simply being an optimal Bloom Filter). To compress a bloom filter, you must choose a large bit vector, and a non-optimal number of hash functions, then apply the compression algorithm (typically arithmetic coding). Because the bit vector is so large, it is sparsely populated -- and so compression works.

Often you can save 10% and 20% on the size of your bloom filter, while having a lower false positive rate. Score!

A very nice, very interesting survey of all the applications of Bloom Filters can be found here.

- sm

Indeed, Bloom Filters are the shit.

These days, in my spare time, I'm writing a p2p program -- think of it as a swarm-download system, like BitTorrent, on an overlay network topology, like eMule (only eMule uses Kademlia, and I'm using Pastry). It has been shown, here and here, that Bloom Filters can drastically reduce the traffic generated when searching peer to peer networks. I recently coded a Java implementation of a Bloom Filter for my p2p program, and it works great in testing. (But the p2p program isn't anywhere near done, so don't ask about it ;)

Furthermore, Bloom Filters can be compressed -- see Michael Mitzenmacher's work here. The idea that you can compress a Bloom Filter is a little counter-intuitive, because the size of the bit vector and the number of hash functions are derived using calculus to maximize the compactness of the set, for a given false positive rate -- thus, in this state, it is non-compressable (it is "already compressed" by simply being an optimal Bloom Filter). To compress a bloom filter, you must choose a large bit vector, and a non-optimal number of hash functions, then apply the compression algorithm (typically arithmetic coding). Because the bit vector is so large, it is sparsely populated -- and so compression works.

Often you can save 10% and 20% on the size of your bloom filter, while having a lower false positive rate. Score!

A very nice, very interesting survey of all the applications of Bloom Filters can be found here.

- sm

now we get them, but without messages!

Screenshot here.

Probably everyone can see the ad for themselves, but here you go.

And I have to nit-pick about the "ironic"... this is not ironic. It's just a coincidence. It would be somewhat ironic if ANOTHER ad service was advertised or something like that...

But they are far away of being technically feasable or cheap enough to be adopted.

energy from moon is feasible with current technology and could be implemented within near 50 years.

the nuclear power on fast neutrons is less expensive and less pollutant that current nuclear power and as well is scheduled to appear in coming years ( at least in russia we have a national draft program for nuclear energy which predicts heavy development and then use of fast neutron nuclear power stations up to 2030 -2050) - yes this will require investments - but there is nothing fantastic with that.

as of fusion and space elevators - yes these technologies are still not developed but the probability that they will appear soon is quite high. and they will help to solve a lot (the space elevator will make solar energy from space even more cheap , fusion (on Helium3 which could be gathered on moon produces no waste http://www.personal.psu.edu/users/t/j/tjh233/heliu m3.pdf) )

so there are things which already could be implemented and things which could greatly improve situation

I could agree that more people - more pollution but it is possible to look into ways to reduce such dependency ( again use for cleaning new technologies ;) ) such as nanotech) and then also look into what mostly produces pollution - and this is actually - use of coal, gasoline etc - if we get rid of their use - average pollution will reduce.

but i disagree with

My point is that the only and "easy" solution we have right now, that can be implemented, is population reduction.

while India , China have programs to stop growing of population it takes them quite a bit of resources and the results ( especially as I know in India) are not so great.

so where on earth to look at examples of easy implementation of your ideas?

citeseer has full text available for for most of its articles, and its a free service, so maybe copyright isn't such a big deal for some reason. Maybe it's because most papers in computer science are available from the author's website.

-jim

If you're already solid in technical skills, you might want to looks at an MBA in general Management or some non-technical concentration. With concentrations like Technology Management, you'll have to take some dumbed-down technical survey courses that will offer absolutely nothing new to you. You're better off spending that time pursuing the business skills you're looking for, be that accounting, people management, project management, marketing, etc.

If you're not up for a full-blown MBA program, you might consider a graduate certificate. For example, Penn State offers non-credit certificate programs in very specific areas like HR Management, Project Management, Leadership Development, Supply Management, etc. Your local university might have something similar.

Here's something kinda similar for CS papers, curtosy of the excellent citeseer:

http://citeseer.ist.psu.edu/articles.html

Dave

No, I am not a microbiologist, I am just a lowly student that has to bow down to microbiologist profs.

Minority Report shows a sort of example of this: stop the disease by fully understanding it and stopping before it happens. This may involve a lot of science (which may include the horrid bad science ), but it does look promising and more research into it is very exciting.

There are many spectulated culprits ( randomness, heredities and free radicals ), but I believe I am going to be one of those graduate students doing 1000s of hours of expirments and research on this. Or maybe I will write a bash script for it instead (bash 3.0 released!)

Just do what the Easter Bunny did -- hire Lobo deal with him.

(For those who don't read comic books, see The Lobo Paramilitary Christmas Special. The AFI even did a short film version

Yeah I've heard that, but DNA says they are probably not relatives of ours.

At least one wasn't.

<a href="http://www.engr.psu.edu/antimatter/documents .html">Antimatter Engine</a>

In the academic community, there are quite a few interesting projects going on. I work on a project called LionShare, which is integrating services like authentication, authorization, and directory in to a federated P2P network.

Penn. State University Made roughly the same deal a few months ago with Napster... Even tried a pilot launch of it last semester ... actually was probably a test for wide-scale launch with othre colleges/universities, like this one...

Not really anything new... but nice to see this kinda thing is getting attention... I stoped downloading (as many) songs when I tried Napster... but napster doesn't have all the songs, even from 'main stream' bands...

Just to protect myself though, now that I'm outta college with a job(and money), I just buy the CD's and rip em... saves me legal issues...

I think the RIAA is a bit too up-tight about fining people who don't have money.. then again, CD's arn't cheap, and I only buy them because I get discounts on CD's with my roomate working at Circuit City....

Penn. State University Made roughly the same deal a few months ago with Napster... Even tried a pilot launch of it last semester ... actually was probably a test for wide-scale launch with othre colleges/universities, like this one...

Not really anything new... but nice to see this kinda thing is getting attention... I stoped downloading (as many) songs when I tried Napster... but napster doesn't have all the songs, even from 'main stream' bands...

Just to protect myself though, now that I'm outta college with a job(and money), I just buy the CD's and rip em... saves me legal issues...

I think the RIAA is a bit too up-tight about fining people who don't have money.. then again, CD's arn't cheap, and I only buy them because I get discounts on CD's with my roomate working at Circuit City....

Not so much a formal study but a fun paper to read: On the Directional Correlation of Axial Rotation in Inverted Felines and Planetary Spin: Coriolis Revisited

The author also happens to be a Computer Science professor in data networks. Quite a dry sense of humor -- his classes are a lot of fun!

Mandatory Disclaimer : yes, I'm a starving grad student of his, and yes, I'm pulling for a graduation date this decade!

Just thought I should point out that you are not going to see the Coriolis effect in a toilet flush unless it's a really big toilet.

Who use Macs for their work, btw.

Which reminds me of those commercials:

Toasted bunny

Snail

(QuickTime required)

Who use Macs for their work, btw.

Which reminds me of those commercials:

Toasted bunny

Snail

(QuickTime required)

The executive branch MUST enforce them.

"I was only following orders" is not an excuse, whether those orders come from the legistlature (in the form of laws) or from a superior. If Congress passed a law tomorrow requiring all blond-haired people to sit in the back of buses, the executive branch would be obligated - legally (by the Constitution) and morally - to fail to enforce it.

The Judiciary is supposed to rule on the constitutionality of law. This is based on the Constitution itself and the body of cases that have been decided by the Supreme and lower courts since the founding of the nation. The Supreme Court can't throw out a "bad" law unless it violates the constitution.

The judicial branch has not just judges, but juries, who have the authority to judge laws and "nullify" bad ones.

Two things about your comment:

1. Document summary (text summarization) is nothing like video summary
2. Text summarization is nothing new. Early papers date back to 1958, (more). MS Word has had it (Tools | Autosummarize) since 1997.

given that the page will be smoking in a while, here's
a mirror of the video

I've been considering this for awhile, and I think the best solution is to convert the energy to fuel. Specifically, anti-matter. If we can develop an effective way to extract energy from matter/antimatter annihilation, then we could use Solar Energy to power antimatter fuel plants.

And if we had some ham, we could have ham and eggs - if we had some eggs.

Antimatter might be a very dense way of storing energy, but making it is incredibly inefficient (PDF). The efficiency of current particle accelerators is about 0.0001% (in terms of energy in/energy stored in antimatter out), and the best that the physicsts seem to think we can do in the near term is about 0.01%. You'd probably get better energy efficiency by putting mirrors in orbit, shining extra light on a plantation in Canada, and running a wood-fired turbine on the extra wood grown.

Antimatter is cool, but it's not going to be widely used as the world's ultimate battery in your or my lifetime.

The End of the Universe exists in Houston, TX.

Ever hear of this Lewis Black skit concerning the Starbucks across the street from a Starbucks.

Check it out here:
http://www.personal.psu.edu/users/i/d/idg101/starb ucks/

And for all the nay-sayers, I have proof!

There are a few of those. The most well known is the (Twin?) River Oaks in Houston. I couldn't find the video of it again, but it is the one Lewis Black is famous for complaining about. Here is a link to some of the double starbucks phenom.
(http://www.personal.psu.edu/users/i/d/idg101/st arbucks/)

Arch-ranter Lewis Black has a funny bit about finding a "Starbucks across from a Starbucks" and he declares that to be the end of the universe. They're in Houston, Texas. Also shown in a composite photograph on another website.

This paper has a bunch of examples.

That's not true! (And neither is the person who responded to you...)

The first exoplanets were discovered by Alexander Wolszczan in 1991, around PSR B1257+12.

They're pulsar planets, yes, but they're planets. Give the guy credit. :)

Story here. Curious that the first discovered planets were Earth-sized. Also the planetary system is very much like Earth's. Dead, yes, but still encouraging.

I work for the Penn State University visualization group and we use Chromium (which supports cave-like deformed viewpoints and stereoscopic rendering over them, as well as low-cost linux clustering and high-end myrinet and infiniband clustering) and OpenSG (A general purpose toolkit for writing scenegraph-based applications). Good luck! Visit our website for information on how we've used it: PSU Vis. Group

• 38 inches according to a page at Arkansas State University and another at Microflex Technologies. Well the conversion (1 meter = 38 inches) is mentioned actually by some apparently russian website which is linked on this page at the arkansas state university
• 38.16 inches according to a rounding-happy math teacher at Norfolk Collegiate School in Virginia.
• Couldnt test this one, because the website was down (probably slashdotted)
• 38.37 inches according to Honeywell's Sensotec folks.
• Ok, well, this is indeed incorrect. However, on the same PDF it is mentioned that 1 inch = 2.54 cm, 1m = 1.0936 yards, which are both correct values. So I seriously believe that (1 m = 38.37in) is just a typo and should have actually been 1m=39.37 in.
• 38.8 inches according to some numerological babble
• Well, if it is "babble",then why consider it at all?
• 39 inches according to Fife Products and some folks who sell quilting products.
• That makes sense, doesn't it? Quilt and other such manufacturers would want to save on by "trimming" or low-rounding such conversions wouldnt they? For selling 1000m of their product, they save 37 inches!
• 39.14 inches according to the specifications on a measuring wheel for engineers. (uh-oh!)
• This does look incorrect. I can't think of why they'd equate 1m=39.14 inches.
• 39.15 inches according to an October 30 2002 entry in a blog.
• Why would you be concerned about what's on a blog. People put whatever they want to.
• 39.21 inches according to Richard Bowles.
• Again, who is richard bowles? I've no idea.. do other slashdotters know? Even if he is an authority on metric systems, why would you use an individual's figures as a source of reference? Would you not prefer to look at a metrics standards body or other such resource?
• 39.27 inches according to pages at University of Wisconsin Stevens Point and the National Optical Astronomy Observatory.
• On the same page you'd notice: "Since many of our students travel to Europe or Australia, we've prepared the chart below to show you how to estimate foreign measurements. We hope you find it helpful:"...Did you notice the word "estimate"? Well, if anything, it wasn't at helpful to you I presume :-)
• 39.28 inches according to Jonathan Brooks at Penn State University.
• Again, I think this "Jonathan Brooks" is a user/student at Penn State University, and this URL you posted isnt an authoritative advisory from the University itself.
• 39.3 inches according to some

This evening, I learned that one meter equals 39.3700787 inches. While this may come as no surprise to some people, it was one to me - for years, I had mistakenly believed a meter was 39.77 inches, and now I know it's basically 39.37.

Of course, I'm not alone in my confusion. A bit of research on Google revealed quite a few different conversions from meters to inches. Here are some of them:

• 38 inches according to a page at Arkansas State University and another at Microflex Technologies.
• 38.16 inches according to a rounding-happy math teacher at Norfolk Collegiate School in Virginia.
• 38.37 inches according to Honeywell's Sensotec folks.
• 38.8 inches according to some numerological babble
• 39 inches according to Fife Products and some folks who sell quilting products.
• 39.14 inches according to the specifications on a measuring wheel for engineers. (uh-oh!)
• 39.15 inches according to an October 30 2002 entry in a blog.
• 39.21 inches according to Richard Bowles.
• 39.27 inches according to pages at University of Wisconsin Stevens Point and the National Optical Astronomy Observatory.
• 39.28 inches according to Jonathan Brooks at Penn State University.
• 39.3 inches according to some laser folks.
• 39.34 inches according to a page about photography, and another about a role-playing game. Hey, it's only a game, their meters can be whatever length they want.
• 39.36 inches according to some ham radio sorts and some NASA folks among others. Pretty close... but... shouldn't NASA know better by now?
• 39.38 inches according to people who race 1-meter model yachts, talk about prehistory in California, and, um, other NASA folks. Again, pretty close!
• 39.39 inches according to someone ranting against metric (how ironic), as well as a page about UFOs.
• 39.4 inches according to a list of conversions from a company that makes electric motors and such things, and the Secretary of the Navy.
• 39.45 inches according to a set of math problems f

In my opinion, Google's market dominance has more to do with massive computing and network resources than technology. (Their reputation of being not evil doesn't hurt, either.)

Their architecture is relatively well-understood. If you want to know how google works, read The Anatomy of a Large-Scale Hypertextual Web Search Engine, The PageRank Citation Ranking: Bringing Order to the Web, and some of the other papers from Google. (They probably have some special tricks they haven't revealed to the public, but the basic algorithms aren't that complicated.)

It's hard for anyone to compete with Google without having enough bandwidth to crawl a large portion of the web on a regular basis, and then store the resulting data structures in RAM.

-jim

This means you have to be skeptical with experiments performed just in simulation without testing the same model in reality.