Similar concept: Project Gutenberg has several CD images and a DVD image for free download. We encourage people to make copies and give them away.
We just dropped off about 300 free CDs at the Berkeley Public Library last week (stop by the Info Desk for a copy), during some recent events.
As others have pointed out, libraries don't really want to catalog and manage stuff, nor do they want to worry about broken and scratched CDs. So, give 'em a spindle of 100 burned CDs or DVDs and let these discs walk out the door!
There are a lot of challenges to making this work truly smoothly (like the cost of putting a nice label on the CD, and troubles with competing DVD formats that don't always read correctly, and who's willing to burn them), but if the goal is to get content "out there," why bother with lending when it only costs a few cents to just give away a CD?
At Gutenberg, we're trying to start a volunteer-based effort that will let anyone request one of our CDs or DVDs via a Web form, then we'll send it to them by postal mail -- free! For a few hours of volunteers' time per month, and minimal costs (donated or reimbursed), why not!
Greg
Test your drive, burn ~9400 free eBooks!
on
DVD-Rs go 8x
·
· Score: 1
Project Gutenberg is doing a big DVD & CD giveaway in December (here is event info). There's a 4.13GB DVD image with about 9400 free eBooks on it at this address:
Help yourself, make a bunch of copies, and be sure to give them away! If you'd like to donate a few copies (or a few hundred) for the giveaway, drop me an email
The short answer is, we are also interested in post-1922 literature. The Project Gutenberg Plan for World Domination through Literature;-) is to get it all!!!
Some "growth areas" include:
Translations and other transformations of public domain works (such transformations get their own copyright, but we're seeing more that comes in with CreativeCommons or similar licenses)
Copyrighted works submitted by contemporary authors. We have a standard non-exclusive permission procedure for this, see our HOWTO at ibiblio.org/gutenberg
Works from 1922 through 1964 that were not renewed. While proving non-renewal has historically been very hard, we are working to make this much easier by digitizing renewal records. In the hundreds of years of copyright registration, only about 10% is ever renewed -- that means that literally millions of items that were registered from 1922-1964 are now public domain, but we need proof from the renewal records. More our copyright howto at the site mentioned above.
2018 is nearly upon us. If copyright is not extended further, we'll once again start getting a year's work added to the public domain each year. (That was the case, until the Bono act of 1998 halted this growth of the public domain for 20 years)
Also, we continue to work with the EFF and ACLU to challenge copyright extension activities. You can expect a rigorous challenge, if YACTE (Yet Another Copyright Term Extension) is proposed in Congress (there were 14 extensions during the 20th century!).
This is something I don't understand. I hope someone can explain or point me to an explanation of what I will try to describe. We all know there are very few laws against spam, and jurisdiction problems even if such laws exist.
Explain to me why it's not easy to demonstrate that someone that puts explicit spam in a child's mailbox isn't committing some sort of other crime. I don't mean "get a good mortgage rate," I mean some of the bad porn related stuff we all see, at least periodically.
So, hypothetically, let's say it's against the law in California to send some gang-bang smut ad to young Timmy. What is preventing the district attorneys, Timmy's mom, etc. from getting an injunction against John Doe? From a subpoena being issued?
Forget for now that tracing back the originator is tough. I'm asking, can't they be charged with a crime in, say, California? THEN, if they're discovered, OR if they ever travel to California & get caught (say, for a speeding ticket), they'll be in deep doo-doo.
This costs money, takes time, and doesn't find the spammers, I agree. But it will make a spammer who wants to go to a conference or travel think twice....and maybe open a whole new dot-com business opportunity: bounty hunters for the charged-but-not-yet-caught spammers.
Someone please explain why these people aren't guilty of crimes that are not spam specific, and why they can't be charged in jurisdictions where the spam is received.
In the academic field of information retrieval, this is called "relevance feedback." It's a part of many information retrieval (IR) algorithms, some of which can happen automatically (i.e., unsupervised). There is also overlap with the fields of machine learning and even Bayesian processes (see today's other/. story about spam filters -- spam filtering is actually the same problem, conceptually, as search engines try to solve).
In Yahoo and other search engines (but not Google, that I've seen), you often get a "click-through" that goes to their system before transparently redirecting to the actual URL you clicked. This is relevance feedback. It's true that the system can't determine whether you LIKED the site (aka, whether it was "relevant"), but at least it's some sort of feedback the system can use to tune.
The other most familiar type of system I can think of is Alexa (now owned by Amazon.com, and the brainchild of the Internet Archive's Brewster Kahle). With Alexa, they could count not just that you visited a site, but how long you spent and where else you went. This is at least part of the basis for Amazon's recommendation system for books and other geegaws they sell.
Can this work in a search engine? Yes, certainly. Does it mean that a search engine that implements relevance feedback will instantly be better than Google? Definitely not! There are many other things (about 20, from what I've heard) that go in to the ranking system that Google uses...Pagerank is one of them, but there are many other factors (such as term frequency, document HTML structure, etc.). Some these, notably Pagerank, work poorly on relatively small collections (in the TREC conference, people have almost never found that Pagerank, HITS or similar algorithmns improve performance with "only" a few tens of GB of Web documents -- a few million pages).
Wanna know more about information retrieval? The TREC page above is very good for state-of-the-art research reports (see the Publications area -- it's all online and free). More general texts are mostly in libraries, but one good one online is Managing Gigabytes, which covers the IR aspects thoroughly and also has lots of ideas about how to use compression in an IR system (something that I'm curious whether Google & others do).
Thanks to everyone who has helped contribute eBooks and other support to Project Gutenberg! If you haven't already, please visit Distributed Proofreaders and proof a page today!
Lots of plans for the future:
Post-#10000 formatting changes. We'll be rearranging our directories to make it easier to find things. Likely we'll go with something OAI (OpenArchives.org) compliant
Conversion on the fly to many formats. We'll putting eBooks into XML format (mostly using teixlite.dtd, we think) for conversion on the fly to many other formats.
New ways to donate. "Sponsor a book"
More contemporary content. We receive donations nearly every week from currently published authors who want to make their stuff available to a wider audience (i.e., our Doctorow's Down and Out)
Your ideas! Visit gutenberg.net to sign up for newsletters, find out how to get started producing an eBook, and find eBooks
Thanks especially to our main and backup distribution sites, iBiblio and The Internet Archive. And thanks to the THOUSANDS of volunteers who have brought us nearly to our 10,000th eBook.
Dr. Gregory B. Newby
Chief Executive and Director
Project Gutenberg Literary Archive Foundation http://gutenberg.net
A 501(c)(3) not-for-profit organization with EIN 64-6221541
Cute. For those of you without compilers handy: # g++ a.cc a.cc:9: warning: all member functions in class `DecentDemocracy' are private a.cc:25: cannot declare variable `USD' to be of type `USDemocracy' a.cc:25: because the following virtual functions are abstract: a.cc:10: virtual void DecentDemocracy::SensibleCopyRight(IntellectualMat erial)
For servers based in the US not trying to profit, there is no restriction on mirroring Project Gutenberg. In fact, we'll even list you in our official mirrors list (http://www.gutenberg.net/list.html) if you'd like!
If you're outside of the US, you might be mirroring some stuff that is under copyright in your country. But many mirrors still do this, prefering to mirror the whole collection rather than try to select items based on copyright rules. For commercial redistribution, the "small print" applies (basically, you need to pay a trademark fee -- details are in each eBook).
Here is the skinny:
The Project Gutenberg etext collection is distributed primarily by FTP, although you can have your Web server point to the same directory and distribute by HTTP. For example, these addresses point to the same content:
ftp://ftp.ibiblio.org/pub/docs/books/gutenberg an d http://www.ibiblio.org/gutenberg
(though ftp or rsync is best for mirroring; see below)
The collection is over 16GB (January 2003), and expected to grow another few GB this year. New etexts are added almost every day, so it's best to mirror nightly.
Our experience has been that a static IP address and T1 (~1.5Mb symmetric) or better permanent network connection is desirable for mirroring; DSL and cable modems do not seem to offer the necessary bandwidth and sometimes suffer stability problems.
The best place to mirror from currently is our master download site at ftp://ftp.ibiblio.org/pub/docs/books/gutenberg . Most mirrors use rsync (easiest), wget (easy) or the mirror PERL software (requires some configuration). Here is an overview for each:
1. Rsync (available for all Unix systems; standard on Linux). The last argument is the local directory for the mirror destination:
2. Wget: Freely available from any GNU mirror. With appropriate command-line options, this can be used with either a HTTP or FTP interface, but please use the FTP URL above for Project Gutenberg. The key is to only get updated files, not files you already have. A wget command line that should work with some adjustment for your local needs (run it from wherever you want the mirror to go) is:
The wget homepage is http://www.gnu.org/gnulist/production/wget.html
3. Mirror PERL software: Available from http://sunsite.org.uk/packages/mirror/ (among other places). We can help you set this up for a Unix system. The mirror PERL software has been reported to work with PERL for WinNT, as well as Unix/Linux/BSD. Note that the wu-ftpd software patch supplied with the program must be applied for it to work!
For any mirror method, run a daily job to check for newly updated files. Unix/Linux employs cron for this; Windows systems could use the task scheduler.
I can help you with setting up the mirroring software, or any other details, if you would like.
We don't distribute the Web-based search engine that's available on the main PG page at http://promo.net/pg. However, we'll add your site to the list there, so people can find you. The FTP directories are the only part we offer for mirror, while the central list of mirrors and search capability is centralized at promo.net.
Once you tell us your mirror is active, we'll announce it in our next weekly & monthly newsletters. After a month or so (to confirm stability) we'll add you to the mirror list and download facility at http://promo.net/pg
Let me know how else I can help. If you decide to go ahead with the mirror, email me and/or webmaster@promo.net so we can add you to the mirror list.
Thanks again for getting in touch! And, thanks for your interest in helping Project Gutenberg reach more readers.
-- Greg
Dr. Gregory B. Newby Chief Executive and Director Project Gutenberg Literary Archive Foundation A 501(c)(3) not-for-profit organization with EIN 64-6221541 gbnewby@ils.unc.edu// 919-962-8064
Michael Hart and I are working on a written response that we'll send to Wired and other media, but by then this/. article will be off the front page. So, allow me to make a few comments.
Projecting back to 1971, Project Gutenberg has tracked Moore's Law quite precisely. January 2003 will be our most productive month ever, and we are looking forward to continuing to double our rate of new eBooks every 18 months.
Project Gutenberg has received some big donations, and we're working on grants and other funding. However, when you do the math you realize that there's essentially no hope for paying for content -- it takes thousands and thousands of people. The hope for "someone" to do it is naive -- the only answer is to figure out ways for "everyone" to work on digitization.
While the author makes 6200 books sound like small potatoes, in fact it represents about 1/3 of all eBooks listed in places like the Internet Public Library. Not bad, and it certainly explains why some random book the author wants isn't part of the collection -- there just aren't that many projects working on digitizing literature.
Where did the author figure on $750million, and for what? Over 30 million printed books were registered for copyright in the last 100 years (this doesn't count magazines, recordings, etc.). The notion that $25/book could pay for digitization is not unreasonable. But where do you get the books, and what about copyright? If there's a plan, I'd like to hear it.
One more point, to keep this short: We have just under 7000 eBooks (up about 800 from whenever the author did his research!). We have over 1000 active volunteers. The books are in over 20 languages, dozens of formats and, if printed, would fill a small library. We're on track to reach #10,000 in 2003. Via Distributed Proofreading, as mentioned here and in a previous/. story, we can and frequently do complete digitizing a 300 page book in just a few hours. Mr. DeLong, I don't feel apologetic about these numbers at all.
That's all for now. Thanks to all the supportive comments in this thread, and to all the constructive criticism. And remember,
a page a day is all it takes to contribute!
Greg Newby, Director and CEO
The Project Gutenberg Literary Archive Foundation
www.gutenberg.net
Can't get through? Try ibiblio
on
Just One Page a Day
·
· Score: 3, Informative
The main Gutenberg page is slashdotted right now, but you can get nearly the same access to the books via the main ibiblio page at ibiblio.org/gutenberg,
which is the main distribution site for the collection.
It looks like the texts01.archive.org/dp site is holding up fairly well! If you cannot get through today, though, please check back later. Slashdot effect aside, it's usually quite speedy and has a decent 'net connection. If you want to keep informed of current events, get on one of our mailing lists via (when it's not slashdotted) our subscriptions page.
Dr. Gregory B. Newby
Chief Executive and Director
Project Gutenberg Literary Archive Foundation http://gutenberg.net
A 501(c)(3) not-for-profit organization with EIN 64-6221541
gbnewby@ils.unc.edu// 919-962-8064
Yes, that's exactly what I have. Lots of people already do this with their Linux system -- using ReiserFS, jfs or ext3 for their data partition, likely ext2 (maybe mounted readonly) for/boot, etc. For filesystems storing large files, ext3, xfs, etc. perhaps tuned for larger block sizea or whatever.
You can also use RAID judiciously. Depending you your anticipated needs, different RAID configurations (including software RAID and RAID with IDE drives) can boost read or write performance (sometimes both, depending on the pattern, size of files, etc.).
In days of large drives, avoiding the fsck delay at boot is critical. On the system I described, it used to have ext2 filesystems on all 13 drives (11x73GB + 2x18GB). Booting after a crash could take hours if fsck needed to fix things. With ReiserFS or another of the journaling filesystems, it's nice and quick.
This topic is near and dear to me....truly "news for nerds, stuff that matters."
My application is for information retrieval, I'm using some software that utilizes BerkeleyDB files at the back end. I spent the last week trying to figure out why I wasn't getting better throughput, and eventually figured out it's related to BerkeleyDB's handling of lots of tree duplicate pages. But that's not why I wanted to post.
One thing people didn't mention: The file system. The file system can make a big difference. For larger files, think about ext2 or XFS. For lots of small files, think ReiserFS. ext3 does journaling and is supposed to have comparable throughput. There's a lot of information out there about filesystems, including a filesystem HOWTO at ibiblio.org. Pick the right filesystem for your application.
Here's what I found. I was copying an 8GB file back and forth (this was one of my DB files; yes, it was sparse, I used "cp --sparse=always"). This was on a Dell 530 with dual 1.7Ghz Xeons, 2GB of PC800 RAM, an Adaptec 39160 controller (U160 SCSI) and JBOD (just a bunch of disks=no raid). Linux kernel is 2.4.18-64GB-SMP on a SUSE 8.0 distribition. The experiments were between different drives on separate channels on the same controller. The drives are 73GB 10KRPM Cheetahs.
I copied the 8GB file and a few other multi-gig files, and used "vmstat" to track progress. This is NOT the way to benchmark for files of just a few meg or even a few hundred meg, because it only samples every few seconds. But for long-running processes, I would "vmstat 10 10000" (resample every 10 seconds; 10000 times) and watch as the files copied in the background on a quiescent system. The "bi" colums is blocks in (typically 4KB blocks, but you can tune this on your system); "bi" is blocks out.
I did XFS to ext2 and back again. I also copied off a ReiserFS drive.
Both XFS and ext2 were comparable for reading & writing. They peaked at about 35,000 bi or bo. 35,000 * 4096 bytes per block =~ 143MB/second. In other words, I was getting close to the max transfer rate for the SCSI bus (160MB/sec per channel). Long-term average was closer to 25K blocks or ~100MB/second.
With a ReiserFS, either for reading or writing, the pattern was that it could peak at ~18K blocks bi or bo, but generally was far lower, on the order of 3000-8000 (i.e., sustained rates of about 12-35MB/sec). What seemed to be happening was the other drive (XFS or ext2) would outpace the ReiserFS' ability to read or write, then wait. If you read the ReiserFS info, they admit this is part of the design (ReiserFS is *great* for loads of small files, really really great). For longer files, they end up needing to basically chain it across a lot of blocks in their B-tree.
I know the question was about IDE, not SCSI, but I'm sure that the filesystem matters for IDE as well, especially if very large files are involved.
If you're working with large files and are willing to lose a percentage to block roundoffs, some filesystems let you choose a block size > 4096 (though I think Linux ends up chunking in 4K blocks anyway).
Project Gutenberg isn't the biggest collection at ibiblio, but it might be the biggest one that is actually based there (as opposed to stuff that's mirrored, such as the Linux distros). You're right that many of these decisions were made awhile ago, I'll try to clarify.
First, you should know that we're in the midst of a big Web page redesign. We'll be moving our main pages from http://promo.net/pg to http://ibiblio.org/gutenberg (with virtual domains of course: gutenberg.net and friends). We'll be addressing many of your concerns. You heard it here, first.
Second, I agree our "finding aids" (in library terms) are poor. It's my #1 priority to get this stuff working better, and in fact several people are working right now to put a new database-driven system into place.
Responses to your questions:
ascii over html: We take everything, but also try to make sure we have a plain ASCII file in addition to other formats. Most volunteers give us just text, since that's what comes from their OCR of books. In the near future, we will have automatic conversion on the fly into nearly ANY format, starting with Braille, then adding HTML, XML, PDF and others including PDA eBook formats....text too, of course.
small print: Since November 2001 the small print at the start is only 35 lines or so, including the title, author, pub date, etc. The long annoying legalese is at the end now. The automatic conversion process mentioned above will enable us to put the most recent header (with the short front part) on all the older content. As to "why do we need the legalese," read the small print itself, it's pretty clear.
server indirection: this is one of those finding aids problems we'll overcome. A cookie or other configuration would do the trick here...
bibliogaphic information: All the recent (last year or two) texts include this right up top. Even the older ones include a "release date" or something similar. The improved finding aids will let you search by publication date, by the way.
We're actively discussing this stuff on the Project Gutenberg Volunteer's Discussion List, see mailing list subscription info for how to subscribe.
Dr. Gregory B. Newby Chief Executive and Director Project Gutenberg Literary Archive Foundation http://gutenberg.net A 501(c)(3) not-for-profit organization with EIN 64-6221541
There are probably far fewer slashdotters familiar with Mathematica than PERL, so I wanted to write a few lines about Mathematica. From what he's said, Wolfram wrote Mathematica a lot like Larry Wall wrote PERL: to solve lots of problems by hiding complexity behind the scenes.
As PERL is the swiss army chainsaw of computer programming, Mathematica is the swiss army chainsaw of mathematics. The syntax isn't as forgiving as PERL, but it's not bad. Here's a snippit I use for singular value decomposition:
svout = SingularValues[N[mydata],Tolerance->0];
{u, md, v} = svout;
Print["u is ", u//MatrixForm];
I've done the same thing with LAPACK and CLAPACK (scientific programming libraries) and in 3 lines of FORTRAN, C or C++ you haven't even started to define your data. In Mathematica, you're already done.
Then there's visualization. Running on a PC or via XWindows, Mathematica can do stunning graphics -- including interactive graphics -- with almost no coding. It's not entirely flexible (sort of like using SAS or SPSS' graphics routines), but again you can do astoundingly great things with almost no code.
In short, Mathematica is very close, for mathematics, to what PERL is for programming (or insert your favorite programming language or toolkit - but I think PERL fits best). While in the olden days of CGI everyone would have their own copy of cgi-lib.pl, now PERL has this functionality built in -- we just do stuff like do stuff like "my $query = new CGI;". In Mathematica, the language has evolved similarly so that stuff you needed to write lots of code for previously is now abstracted to a few functions. Like PERL's ability to use modules, you can write your own add-ons for Mathematica. Like PERL's POD, Mathematica can be used for documentation (and *was* used to write the Mathematica Book, and presumably Wolfram's new book).
Just a few words about Mathematica. Give it a try, if you're remotely interested in how this stuff works. You'll probably like it!
Re:The forbidden fruits of radio
on
GNU Radio
·
· Score: 3, Interesting
In the US, radio receivers and scanners are forbidden from receiving in the cellular phone bands (not that hard to get around by buying a radio elsewhere or modifying your own).
Spread-spectrum technology, as used in wireless phones, especially 3G wireless and other communication, is generally inaccessible to your generic home scanner or ham setup. Also, it's often digital, which again means a home scanner or ham is out of luck.
Enter GNU radio. This sounds like it will easily enable receiving the "forbidden" bands, and give a lot of computer power to re-assemble the spread-spectrum signals. I don't know if this is the intent of the developers, but the potential for bypassing existing radio "security" (really, security through obscurity & legalese) seems strong.
The information retrieval technologies involved
on
Teoma Aims To Kill Google
·
· Score: 3, Informative
Their "jobs" link mentions a variety of technologies, including LAPACK. LAPACK is a collection of scientific functions (there's a C version, CLAPACK, but LAPACK is FORTRAN). My guess is they're using, among other things, techniques related to latent semantic indexing (LSI) and vector space models (VSM) for their ranking.
Unless you're an Information Retrieval Wienie (like me), you might not know about LSI and the VSM. The cool thing is that these are methods that work really well in the laboratory, but have scaling problems so are not found much in large-scale systems.
Google, we know, uses Page Rank to rank pages based (partially) on the "authority" of the page. It's not clear whether Teoma uses this or not (it is patented). LSI is also patented (by Bell/Lucent), but VSM is not.
For both Google and Teoma, they seem to use hybrid approaches:
- Word occurrence, with weighting (weight of a term in a document; weight of a term in a collection). This is fundamental to all search engines (it's part of what distinguishes an information retrieval system from a database).
- Statistical relations among words and documents (e.g., VSM and LSI techniques -- there are many variations). These look at either a term by document matrix (where each cell is a term count), or term by term matrices (where each cell a measure of the terms' pairwise relatedness).
- Clustering, to eliminate duplicates and identify groupings (Teoma seems to do this; Google does this in their directory. This is NorthernLight's claim to fame, and is patented)
- Authority ranking (it's not clear whether Teoma does what Google does, but this is probably a part of the mix)
Each search engine has its own recipe for how these and other factors are combined. If only they would share (and stop getting software patents)!
Read the bill, or at least the summary at the top. Unfortunately, the Congress might actually have our (the people's) best interests at heart. Also unfortunately, the telcos and cable company operators just aren't interested in EITHER opening to competition OR giving good service.
What we THOUGHT was that the telecom act of 96 would level the playing field for smaller players. This hasn't happened, for reasons you see in other posts in this thread.
What we THOUGHT was that technology would rapidly get better, yielding higher bandwidth and a greater ability to get beyond the coupla-kilometers limit. There's been progress, but basically we're still stuck with the same technology as in '96 and before.
What we THOUGHT was that other players (power companies, wireless companies and funky stuff like blimps flying around over cities) would provoke telcos & cable companies to do better. But apart from satelite Internet (which is too slow for gaming and most other interactive use), there are not viable alternatives for most people.
Basically, things have moved more slowly than we, the geeks, thought they would, and the cable companies and telcos have been able to have their way: little competition, top price, and little need for good service.
There's still hope for new technologies and other developments (like municipalities' interest in WLANs) that might give hope to competition for xDSL and cable modem service for "broadband" Internet service. But it doesn't look like there's any hope that any sort of regulation will create real improvements for most users (or wannabe users) for today's "broadband" Internet services.
I went ahead and installed this to show my systems administration class on Monday (INLS183). If you're trying to install the software, I included steps in the sample directory at this location. The installation steps I used are in parody-steps.txt.
I made a parody, visit here to see (it probably won't be up too long...). Finally, you can also get the code in the directory mentioned above, if you are having trouble finding a mirror. Retrieve yesiwill-1.0.tar.gz
Good theory, but in most cases this is just a dream.
For those organizations lucky enough to have cheap or free legal advice, lawyers are not going to be interested in doing a lot of work.
When the 'squatting' organization is in another state, it can be very hard (and more expensive) to sue them, as the lawsuit will likely need to be filed in the state where the 'squatter' is.
When the squatter is overseas, fuggetaboutit. Have you tried finding a lawyer that would do this sort of work when there's not big money involved? Most lawyers have no credentials or experience with things like international trademarks and oversees lawsuits, and the ones who do don't work cheaply.
The likely scenario is to find a lawyer who is willing to work on the case. If the case is not reasonably assured of success, he/she will want a retainer or other payment. This isn't like chasing ambulances: it's hard to predict success, and even harder to know whether damages (or even legal fees) will be rewarded.
Take a look at ICANN's approved providers list for domain name dispute resolution. The same page links to their Uniform Dispute Resolution Policy.
The big problem here, which a few other people touched on, is that it costs from $1250-$4000 to petition for a domain name dispute. (The different providers set their own fees.)
This is ridiculously expensive for not-for-profits and individuals, but chump change for big companies. What would make much more sense is a pay scale depending on who you are and who you're going up against...obviously, we don't wnat to make it trivial for every yahoo to claim they have a stake in coke.com for something like $19.95, but it's hardly logical that a broke organization should need to cough up $1250 to fight a porn operation.
This is near and dear to my heart right now, because there's an anti-muslim hate site at projectgutenberg.com (I'm the CEO of Project Gutenberg (the real site), and we really don't have the dough to go through the domain dispute process.
Greg
Temporary mirrors (reply here)
on
GNU Emacs 21
·
· Score: 2
It's tough getting through to ftp.gnu.org, and most of the mirror sites won't update until overnight tonight.
So, I've put emacs-21.1.tar.gz and leim-21.1.tar.gz for a temporary mirror. Visit:
In "The Transparent Society" (highly recommended) sci-fi author and physicist David Brin talks about a future society where cameras are everywhere.
I'm starting to think this part of his vision is inevitable. In the US, where the 4th amendment protects against some forms of surveillance, it's mostly been interpreted to only protect things we do in our homes. The streets, and the workplace, are open game for nearly any type of surveillance.
The important part: Brin wanted ANYONE to be able to tap into the cameras, ANY TIME. He also wanted cameras watching the watchers: we should be able to turn into our local police station, and make sure they're doing their job properly. This is the part that's missing from current proposals in the US and current practice in the UK, yet it would clearly be beneficial:
More watchers, including the public
We would know when the cameras are being abused (e.g., zooming in on the pretty girls in the mall...)
Accountability for law enforcement. If the cameras are there, we want to make sure they're i the right locations, and functional, and being used properly.
Tangential benefits, such as having better network infrastructure for live video multicasting.
In a world where surveillance seems impossible to avoid, I can only wish that Brin's vision had a better chance of becoming reality.
You lease dark fiber when you want to connect your own equipment to it at either end. The fiber is lit, but the U. doesn't need to pay for it (any more than they do to connect any other buildings or locations; that is, it's internal, part of the campus network). The U. lights it itself, but doesn't need to pay the phone company (Bell South) except to lease the fiber.
This is very typical at big universities, school districts, and other organizations that have multiple locations in a metropolitan area. It's used to connect off-campus offices etc. to the campus backbone. I'm sure some companies do it, too, if they have multiple offices in a relatively small area.
BTW, you're right that only phone companies and cable companies (and maybe a few other lucky players) own dark fiber. But, like any good whore, they're happy to sell it to you....for a little while.
Sealand has a line of sight wireless link, I understand (microwave or a similar frequency). They're about 12 miles off the coast, right? They need to pay for their upstream provider(s), of course, but don't need to pay the phone company for the local loop. (They are in the UK, so the telephone industry is quite different, as are regulations for things like ground-based microwave relays). (OK, they're not in the UK, they're an independent country, the Principality of Sealand. But their upstream link is in the UK. Geesh, you people are soooooo picky!)
In the silo, you'd want at least two independent pipes coming in; the specifics depend on what's available, good and cost-effective. But at least one would probably involve a local loop charge by the phone company, and would be a significant monthly expense in addition to whatever else is paid for the upstream Internet link. As an example, the U. where I work pays $2000/month for a single pair of dark fiber totalling less than 1 kilometer.
Anyway, I agree that whoever has $1.5M for the silo can afford the cost of the Internet service. The question is whether they can make any money running a secure ISP or co-lo that way!
In case you haven't noticed, all the big ISP/hosting/co-lo/outsourcing providers seem to be located in or near phone company buildings, NAPs or MAEs. Apart from providing a reasonably secure facility with good network access, these offer the benefit of not needing to pay for distance from the telco building.
I was in the Netherlands for HAL2001 this summer, and got to visit a NAP right across the street from the U. Twente campus (where HAL was held). A NAP is where different networks peer, and in this case is also where at least one ISP provided co-lo space and other ameneties.
The cool part: it was in a retired federal bank. Literally a fortress: fully bulletproof, tempest-shielded, multiple sub-basements, iron gates and fully enclosed by fences or walls, the works.
The ICBM silo gets me thinking about the same thing. They have on-site power generation and battery backup and an obviously pretty damn secure setup. So, why not open a secure hosting facility? It's not HavenCo/Sealand, but it's not bad.
The main problem is it's in the middle of nowhere (Mapquest link ), about 50 miles from Topeka. Paying the local loop charges for dedicated (and redundant) Internet access is probably going to cost a fortune.
Slashdot Mirror
We just dropped off about 300 free CDs at the Berkeley Public Library last week (stop by the Info Desk for a copy), during some recent events. As others have pointed out, libraries don't really want to catalog and manage stuff, nor do they want to worry about broken and scratched CDs. So, give 'em a spindle of 100 burned CDs or DVDs and let these discs walk out the door!
There are a lot of challenges to making this work truly smoothly (like the cost of putting a nice label on the CD, and troubles with competing DVD formats that don't always read correctly, and who's willing to burn them), but if the goal is to get content "out there," why bother with lending when it only costs a few cents to just give away a CD?
At Gutenberg, we're trying to start a volunteer-based effort that will let anyone request one of our CDs or DVDs via a Web form, then we'll send it to them by postal mail -- free! For a few hours of volunteers' time per month, and minimal costs (donated or reimbursed), why not!
ftp.ibiblio.org/pub/docs/books/gutenberg/cdimages/ pgdvd.iso (md5sum 59d8a193874349181122ff52e2e3e114, size 4139646976)
Help yourself, make a bunch of copies, and be sure to give them away! If you'd like to donate a few copies (or a few hundred) for the giveaway, drop me an email
Also, we continue to work with the EFF and ACLU to challenge copyright extension activities. You can expect a rigorous challenge, if YACTE (Yet Another Copyright Term Extension) is proposed in Congress (there were 14 extensions during the 20th century!).
This is something I don't understand. I hope someone can explain or point me to an explanation of what I will try to describe. We all know there are very few laws against spam, and jurisdiction problems even if such laws exist.
Explain to me why it's not easy to demonstrate that someone that puts explicit spam in a child's mailbox isn't committing some sort of other crime. I don't mean "get a good mortgage rate," I mean some of the bad porn related stuff we all see, at least periodically.
So, hypothetically, let's say it's against the law in California to send some gang-bang smut ad to young Timmy. What is preventing the district attorneys, Timmy's mom, etc. from getting an injunction against John Doe? From a subpoena being issued?
Forget for now that tracing back the originator is tough. I'm asking, can't they be charged with a crime in, say, California? THEN, if they're discovered, OR if they ever travel to California & get caught (say, for a speeding ticket), they'll be in deep doo-doo.
This costs money, takes time, and doesn't find the spammers, I agree. But it will make a spammer who wants to go to a conference or travel think twice....and maybe open a whole new dot-com business opportunity: bounty hunters for the charged-but-not-yet-caught spammers.
Someone please explain why these people aren't guilty of crimes that are not spam specific, and why they can't be charged in jurisdictions where the spam is received.
In Yahoo and other search engines (but not Google, that I've seen), you often get a "click-through" that goes to their system before transparently redirecting to the actual URL you clicked. This is relevance feedback. It's true that the system can't determine whether you LIKED the site (aka, whether it was "relevant"), but at least it's some sort of feedback the system can use to tune.
The other most familiar type of system I can think of is Alexa (now owned by Amazon.com, and the brainchild of the Internet Archive's Brewster Kahle). With Alexa, they could count not just that you visited a site, but how long you spent and where else you went. This is at least part of the basis for Amazon's recommendation system for books and other geegaws they sell.
Can this work in a search engine? Yes, certainly. Does it mean that a search engine that implements relevance feedback will instantly be better than Google? Definitely not! There are many other things (about 20, from what I've heard) that go in to the ranking system that Google uses...Pagerank is one of them, but there are many other factors (such as term frequency, document HTML structure, etc.). Some these, notably Pagerank, work poorly on relatively small collections (in the TREC conference, people have almost never found that Pagerank, HITS or similar algorithmns improve performance with "only" a few tens of GB of Web documents -- a few million pages).
Wanna know more about information retrieval? The TREC page above is very good for state-of-the-art research reports (see the Publications area -- it's all online and free). More general texts are mostly in libraries, but one good one online is Managing Gigabytes, which covers the IR aspects thoroughly and also has lots of ideas about how to use compression in an IR system (something that I'm curious whether Google & others do).
Lots of plans for the future:
Thanks especially to our main and backup distribution sites, iBiblio and The Internet Archive. And thanks to the THOUSANDS of volunteers who have brought us nearly to our 10,000th eBook.
Dr. Gregory B. Newby
Chief Executive and Director
Project Gutenberg Literary Archive Foundation
http://gutenberg.net
A 501(c)(3) not-for-profit organization with EIN 64-6221541
gbnewby@pglaf.org
Cute. For those of you without compilers handy:
t erial)
# g++ a.cc
a.cc:9: warning: all member functions in class `DecentDemocracy' are private
a.cc:25: cannot declare variable `USD' to be of type `USDemocracy'
a.cc:25: because the following virtual functions are abstract:
a.cc:10: virtual void DecentDemocracy::SensibleCopyRight(IntellectualMa
For servers based in the US not trying to profit, there is no restriction on mirroring Project Gutenberg. In fact, we'll even list you in our official mirrors list (http://www.gutenberg.net/list.html) if you'd like!
n d http://www.ibiblio.org/gutenberg
g . Most mirrors use
/home/ftp/pub/mirrors/gute
// 919-962-8064
If you're outside of the US, you might be mirroring some stuff that is under copyright in your country. But many mirrors still do this, prefering to mirror the whole collection rather than try to select items based on copyright rules. For commercial redistribution, the "small print" applies (basically, you need to pay a trademark fee -- details are in each eBook).
Here is the skinny:
The Project Gutenberg etext collection is distributed primarily by
FTP, although you can have your Web server point to the same directory
and distribute by HTTP. For example, these addresses point to the
same content:
ftp://ftp.ibiblio.org/pub/docs/books/gutenberg
a
(though ftp or rsync is best for mirroring; see below)
The collection is over 16GB (January 2003), and expected to grow another
few GB this year. New etexts are added almost every day, so it's best
to mirror nightly.
Our experience has been that a static IP address and T1 (~1.5Mb
symmetric) or better permanent network connection is desirable for
mirroring; DSL and cable modems do not seem to offer the necessary
bandwidth and sometimes suffer stability problems.
The best place to mirror from currently is our master download site at
ftp://ftp.ibiblio.org/pub/docs/books/gutenber
rsync (easiest), wget (easy) or the mirror PERL software (requires
some configuration). Here is an overview for each:
1. Rsync (available for all Unix systems; standard on Linux). The last
argument is the local directory for the mirror destination:
rsync -rlHtSv --delete ftp@ftp.ibiblio.org::gutenberg
nberg
2. Wget: Freely available from any GNU mirror. With appropriate
command-line options, this can be used with either a HTTP or FTP
interface, but please use the FTP URL above for Project Gutenberg.
The key is to only get updated files, not files you already have. A
wget command line that should work with some adjustment for your local
needs (run it from wherever you want the mirror to go) is:
wget --mirror --no-host-directories --passive-ftp --no-parent --cut-dirs=4 \
--output-file=/tmp/wget-gutenberg.log \
ftp://ftp.ibiblio.org/pub/docs/books/gutenberg
The wget homepage is http://www.gnu.org/gnulist/production/wget.html
3. Mirror PERL software: Available from
http://sunsite.org.uk/packages/mirror/ (among other places). We can
help you set this up for a Unix system. The mirror PERL software has
been reported to work with PERL for WinNT, as well as Unix/Linux/BSD.
Note that the wu-ftpd software patch supplied with the program must be
applied for it to work!
For any mirror method, run a daily job to check for newly updated
files. Unix/Linux employs cron for this; Windows systems could use
the task scheduler.
I can help you with setting up the mirroring software, or any other
details, if you would like.
We don't distribute the Web-based search engine that's available on
the main PG page at http://promo.net/pg. However, we'll add your site
to the list there, so people can find you. The FTP directories are
the only part we offer for mirror, while the central list of mirrors
and search capability is centralized at promo.net.
Once you tell us your mirror is active, we'll announce it in our next
weekly & monthly newsletters. After a month or so (to confirm
stability) we'll add you to the mirror list and download facility at
http://promo.net/pg
Let me know how else I can help. If you decide to go ahead with the
mirror, email me and/or webmaster@promo.net so we can add you to the
mirror list.
Thanks again for getting in touch! And, thanks for your interest in
helping Project Gutenberg reach more readers.
-- Greg
Dr. Gregory B. Newby
Chief Executive and Director
Project Gutenberg Literary Archive Foundation
A 501(c)(3) not-for-profit organization with EIN 64-6221541
gbnewby@ils.unc.edu
That's all for now. Thanks to all the supportive comments in this thread, and to all the constructive criticism. And remember, a page a day is all it takes to contribute!
Greg Newby, Director and CEO
The Project Gutenberg Literary Archive Foundation
www.gutenberg.net
It looks like the texts01.archive.org/dp site is holding up fairly well! If you cannot get through today, though, please check back later. Slashdot effect aside, it's usually quite speedy and has a decent 'net connection. If you want to keep informed of current events, get on one of our mailing lists via (when it's not slashdotted) our subscriptions page.
Dr. Gregory B. Newby // 919-962-8064
Chief Executive and Director
Project Gutenberg Literary Archive Foundation http://gutenberg.net
A 501(c)(3) not-for-profit organization with EIN 64-6221541
gbnewby@ils.unc.edu
Yes, that's exactly what I have. Lots of people already do this with their Linux system -- using ReiserFS, jfs or ext3 for their data partition, likely ext2 (maybe mounted readonly) for /boot, etc. For filesystems storing large files, ext3, xfs, etc. perhaps tuned for larger block sizea or whatever.
You can also use RAID judiciously. Depending you your anticipated needs, different RAID configurations (including software RAID and RAID with IDE drives) can boost read or write performance (sometimes both, depending on the pattern, size of files, etc.).
In days of large drives, avoiding the fsck delay at boot is critical. On the system I described, it used to have ext2 filesystems on all 13 drives (11x73GB + 2x18GB). Booting after a crash could take hours if fsck needed to fix things. With ReiserFS or another of the journaling filesystems, it's nice and quick.
My application is for information retrieval, I'm using some software that utilizes BerkeleyDB files at the back end. I spent the last week trying to figure out why I wasn't getting better throughput, and eventually figured out it's related to BerkeleyDB's handling of lots of tree duplicate pages. But that's not why I wanted to post.
One thing people didn't mention: The file system. The file system can make a big difference. For larger files, think about ext2 or XFS. For lots of small files, think ReiserFS. ext3 does journaling and is supposed to have comparable throughput. There's a lot of information out there about filesystems, including a filesystem HOWTO at ibiblio.org. Pick the right filesystem for your application.
Here's what I found. I was copying an 8GB file back and forth (this was one of my DB files; yes, it was sparse, I used "cp --sparse=always"). This was on a Dell 530 with dual 1.7Ghz Xeons, 2GB of PC800 RAM, an Adaptec 39160 controller (U160 SCSI) and JBOD (just a bunch of disks=no raid). Linux kernel is 2.4.18-64GB-SMP on a SUSE 8.0 distribition. The experiments were between different drives on separate channels on the same controller. The drives are 73GB 10KRPM Cheetahs.
I copied the 8GB file and a few other multi-gig files, and used "vmstat" to track progress. This is NOT the way to benchmark for files of just a few meg or even a few hundred meg, because it only samples every few seconds. But for long-running processes, I would "vmstat 10 10000" (resample every 10 seconds; 10000 times) and watch as the files copied in the background on a quiescent system. The "bi" colums is blocks in (typically 4KB blocks, but you can tune this on your system); "bi" is blocks out.
I did XFS to ext2 and back again. I also copied off a ReiserFS drive.
Both XFS and ext2 were comparable for reading & writing. They peaked at about 35,000 bi or bo. 35,000 * 4096 bytes per block =~ 143MB/second. In other words, I was getting close to the max transfer rate for the SCSI bus (160MB/sec per channel). Long-term average was closer to 25K blocks or ~100MB/second.
With a ReiserFS, either for reading or writing, the pattern was that it could peak at ~18K blocks bi or bo, but generally was far lower, on the order of 3000-8000 (i.e., sustained rates of about 12-35MB/sec). What seemed to be happening was the other drive (XFS or ext2) would outpace the ReiserFS' ability to read or write, then wait. If you read the ReiserFS info, they admit this is part of the design (ReiserFS is *great* for loads of small files, really really great). For longer files, they end up needing to basically chain it across a lot of blocks in their B-tree.
I know the question was about IDE, not SCSI, but I'm sure that the filesystem matters for IDE as well, especially if very large files are involved. If you're working with large files and are willing to lose a percentage to block roundoffs, some filesystems let you choose a block size > 4096 (though I think Linux ends up chunking in 4K blocks anyway).
First, you should know that we're in the midst of a big Web page redesign. We'll be moving our main pages from http://promo.net/pg to http://ibiblio.org/gutenberg (with virtual domains of course: gutenberg.net and friends). We'll be addressing many of your concerns. You heard it here, first.
Second, I agree our "finding aids" (in library terms) are poor. It's my #1 priority to get this stuff working better, and in fact several people are working right now to put a new database-driven system into place.
Responses to your questions: ascii over html: We take everything, but also try to make sure we have a plain ASCII file in addition to other formats. Most volunteers give us just text, since that's what comes from their OCR of books. In the near future, we will have automatic conversion on the fly into nearly ANY format, starting with Braille, then adding HTML, XML, PDF and others including PDA eBook formats....text too, of course.
small print: Since November 2001 the small print at the start is only 35 lines or so, including the title, author, pub date, etc. The long annoying legalese is at the end now. The automatic conversion process mentioned above will enable us to put the most recent header (with the short front part) on all the older content. As to "why do we need the legalese," read the small print itself, it's pretty clear.
server indirection: this is one of those finding aids problems we'll overcome. A cookie or other configuration would do the trick here...
bibliogaphic information: All the recent (last year or two) texts include this right up top. Even the older ones include a "release date" or something similar. The improved finding aids will let you search by publication date, by the way.
We're actively discussing this stuff on the Project Gutenberg Volunteer's Discussion List, see mailing list subscription info for how to subscribe.
Dr. Gregory B. Newby
Chief Executive and Director
Project Gutenberg Literary Archive Foundation
http://gutenberg.net
A 501(c)(3) not-for-profit organization with EIN 64-6221541
As PERL is the swiss army chainsaw of computer programming, Mathematica is the swiss army chainsaw of mathematics. The syntax isn't as forgiving as PERL, but it's not bad. Here's a snippit I use for singular value decomposition:
{u, md, v} = svout;
Print["u is ", u//MatrixForm];
I've done the same thing with LAPACK and CLAPACK (scientific programming libraries) and in 3 lines of FORTRAN, C or C++ you haven't even started to define your data. In Mathematica, you're already done.
Then there's visualization. Running on a PC or via XWindows, Mathematica can do stunning graphics -- including interactive graphics -- with almost no coding. It's not entirely flexible (sort of like using SAS or SPSS' graphics routines), but again you can do astoundingly great things with almost no code.
In short, Mathematica is very close, for mathematics, to what PERL is for programming (or insert your favorite programming language or toolkit - but I think PERL fits best). While in the olden days of CGI everyone would have their own copy of cgi-lib.pl, now PERL has this functionality built in -- we just do stuff like do stuff like "my $query = new CGI;". In Mathematica, the language has evolved similarly so that stuff you needed to write lots of code for previously is now abstracted to a few functions. Like PERL's ability to use modules, you can write your own add-ons for Mathematica. Like PERL's POD, Mathematica can be used for documentation (and *was* used to write the Mathematica Book, and presumably Wolfram's new book).
Just a few words about Mathematica. Give it a try, if you're remotely interested in how this stuff works. You'll probably like it!
In the US, radio receivers and scanners are forbidden from receiving in the cellular phone bands (not that hard to get around by buying a radio elsewhere or modifying your own).
Spread-spectrum technology, as used in wireless phones, especially 3G wireless and other communication, is generally inaccessible to your generic home scanner or ham setup. Also, it's often digital, which again means a home scanner or ham is out of luck.
Enter GNU radio. This sounds like it will easily enable receiving the "forbidden" bands, and give a lot of computer power to re-assemble the spread-spectrum signals. I don't know if this is the intent of the developers, but the potential for bypassing existing radio "security" (really, security through obscurity & legalese) seems strong.
Their "jobs" link mentions a variety of technologies, including LAPACK. LAPACK is a collection of scientific functions (there's a C version, CLAPACK, but LAPACK is FORTRAN). My guess is they're using, among other things, techniques related to latent semantic indexing (LSI) and vector space models (VSM) for their ranking.
Unless you're an Information Retrieval Wienie (like me), you might not know about LSI and the VSM. The cool thing is that these are methods that work really well in the laboratory, but have scaling problems so are not found much in large-scale systems.
Google, we know, uses Page Rank to rank pages based (partially) on the "authority" of the page. It's not clear whether Teoma uses this or not (it is patented). LSI is also patented (by Bell/Lucent), but VSM is not.
For both Google and Teoma, they seem to use hybrid approaches:
- Word occurrence, with weighting (weight of a term in a document; weight of a term in a collection). This is fundamental to all search engines (it's part of what distinguishes an information retrieval system from a database).
- Statistical relations among words and documents (e.g., VSM and LSI techniques -- there are many variations). These look at either a term by document matrix (where each cell is a term count), or term by term matrices (where each cell a measure of the terms' pairwise relatedness).
- Clustering, to eliminate duplicates and identify groupings (Teoma seems to do this; Google does this in their directory. This is NorthernLight's claim to fame, and is patented)
- Authority ranking (it's not clear whether Teoma does what Google does, but this is probably a part of the mix)
Each search engine has its own recipe for how these and other factors are combined. If only they would share (and stop getting software patents)!
...Greg
Read the bill, or at least the summary at the top. Unfortunately, the Congress might actually have our (the people's) best interests at heart. Also unfortunately, the telcos and cable company operators just aren't interested in EITHER opening to competition OR giving good service.
What we THOUGHT was that the telecom act of 96 would level the playing field for smaller players. This hasn't happened, for reasons you see in other posts in this thread.
What we THOUGHT was that technology would rapidly get better, yielding higher bandwidth and a greater ability to get beyond the coupla-kilometers limit. There's been progress, but basically we're still stuck with the same technology as in '96 and before.
What we THOUGHT was that other players (power companies, wireless companies and funky stuff like blimps flying around over cities) would provoke telcos & cable companies to do better. But apart from satelite Internet (which is too slow for gaming and most other interactive use), there are not viable alternatives for most people.
Basically, things have moved more slowly than we, the geeks, thought they would, and the cable companies and telcos have been able to have their way: little competition, top price, and little need for good service.
There's still hope for new technologies and other developments (like municipalities' interest in WLANs) that might give hope to competition for xDSL and cable modem service for "broadband" Internet service. But it doesn't look like there's any hope that any sort of regulation will create real improvements for most users (or wannabe users) for today's "broadband" Internet services.
I made a parody, visit here to see (it probably won't be up too long...). Finally, you can also get the code in the directory mentioned above, if you are having trouble finding a mirror. Retrieve yesiwill-1.0.tar.gz
The likely scenario is to find a lawyer who is willing to work on the case. If the case is not reasonably assured of success, he/she will want a retainer or other payment. This isn't like chasing ambulances: it's hard to predict success, and even harder to know whether damages (or even legal fees) will be rewarded.
The big problem here, which a few other people touched on, is that it costs from $1250-$4000 to petition for a domain name dispute. (The different providers set their own fees.)
This is ridiculously expensive for not-for-profits and individuals, but chump change for big companies. What would make much more sense is a pay scale depending on who you are and who you're going up against...obviously, we don't wnat to make it trivial for every yahoo to claim they have a stake in coke.com for something like $19.95, but it's hardly logical that a broke organization should need to cough up $1250 to fight a porn operation.
This is near and dear to my heart right now, because there's an anti-muslim hate site at projectgutenberg.com (I'm the CEO of Project Gutenberg (the real site), and we really don't have the dough to go through the domain dispute process.
So, I've put emacs-21.1.tar.gz and leim-21.1.tar.gz for a temporary mirror. Visit:
If you make a temporary mirror, perhaps you could respond to this post. ... Greg
The important part: Brin wanted ANYONE to be able to tap into the cameras, ANY TIME. He also wanted cameras watching the watchers: we should be able to turn into our local police station, and make sure they're doing their job properly. This is the part that's missing from current proposals in the US and current practice in the UK, yet it would clearly be beneficial:
In a world where surveillance seems impossible to avoid, I can only wish that Brin's vision had a better chance of becoming reality.
"You're a pussy, pussy." (E.Cartman).
You lease dark fiber when you want to connect your own equipment to it at either end. The fiber is lit, but the U. doesn't need to pay for it (any more than they do to connect any other buildings or locations; that is, it's internal, part of the campus network). The U. lights it itself, but doesn't need to pay the phone company (Bell South) except to lease the fiber.
This is very typical at big universities, school districts, and other organizations that have multiple locations in a metropolitan area. It's used to connect off-campus offices etc. to the campus backbone. I'm sure some companies do it, too, if they have multiple offices in a relatively small area.
BTW, you're right that only phone companies and cable companies (and maybe a few other lucky players) own dark fiber. But, like any good whore, they're happy to sell it to you....for a little while.
Sealand has a line of sight wireless link, I understand (microwave or a similar frequency). They're about 12 miles off the coast, right? They need to pay for their upstream provider(s), of course, but don't need to pay the phone company for the local loop. (They are in the UK, so the telephone industry is quite different, as are regulations for things like ground-based microwave relays). (OK, they're not in the UK, they're an independent country, the Principality of Sealand. But their upstream link is in the UK. Geesh, you people are soooooo picky!)
In the silo, you'd want at least two independent pipes coming in; the specifics depend on what's available, good and cost-effective. But at least one would probably involve a local loop charge by the phone company, and would be a significant monthly expense in addition to whatever else is paid for the upstream Internet link. As an example, the U. where I work pays $2000/month for a single pair of dark fiber totalling less than 1 kilometer.
Anyway, I agree that whoever has $1.5M for the silo can afford the cost of the Internet service. The question is whether they can make any money running a secure ISP or co-lo that way!
In case you haven't noticed, all the big ISP/hosting/co-lo/outsourcing providers seem to be located in or near phone company buildings, NAPs or MAEs. Apart from providing a reasonably secure facility with good network access, these offer the benefit of not needing to pay for distance from the telco building.
The cool part: it was in a retired federal bank. Literally a fortress: fully bulletproof, tempest-shielded, multiple sub-basements, iron gates and fully enclosed by fences or walls, the works.
The ICBM silo gets me thinking about the same thing. They have on-site power generation and battery backup and an obviously pretty damn secure setup. So, why not open a secure hosting facility? It's not HavenCo/Sealand, but it's not bad.
The main problem is it's in the middle of nowhere (Mapquest link ), about 50 miles from Topeka. Paying the local loop charges for dedicated (and redundant) Internet access is probably going to cost a fortune.