Back when the DNS was first coming out, some of us UUCP-geezers weren't convinced that a central registration system would be accepted by the community - this was before trademarks were an issue, and it was ok to name your PDPs and Vaxen after colors like red, green, and blue, or common office equipment like xerox, coke, and mrcoffee, and there were 17+ machines named mozart and 30+ named bilbo or frodo. Local naming was the only real alternative (remember bang-routed email addresses?). Needless to say, we were wrong, and DNS was a big success for the first decade and a half, but there's still the problem that if there's only One Root To Rule Them All, somebody's got to run the thing and there will be naming conflicts. One of the main reasons DNS worked as well as it did is that most machines belonged to organizations with well-identified names, and they could fight it out internally for whose machine got to be mozart.foobar.com.
There have been several proposals for adding more TLDs - the IAHC International Ad-Hoc Committee was relatively reasonable, ICANN was a bit less so, thoguh that was partly because it was a year or two later so there was more commercial conflict, and Esther (bless her heart) knew it would be a dirty job when she took it. As far as I know, the only proposals for new TLDs that have actually succeeded have been a few new country codes (because there's an existing bureaucracy for that, plus of course the countries who've made a quick buck by renting out their namespace), and Brad Templeton's proposal for.invalid, which is declared to be syntactically correct, so you can use it in books and demoware, but doesn't point to anything real.
There have also been the disorganized proposals, from people like Kaspureff at Alternic, and the orange.net folks - start an alternative root, and try to convince people to use your root instead of the Big Roots, but they're fighting a losing game. It's partly a losing game because they've been losing (:-), and partly because it doesn't solve the fundamental problem, it just trashes any efficiencies you gain by shoving conflicting names down a layer in the tree so you don't see them if you're not looking for them.
The people who've been successful at pushing new namespaces have taken different approaches - ICQ numbers are a global namespace, and nobody minds because they don't spell anything and the server can cope with the scale. Realnames sells namespace, and people who want it can use it. And all of these things can easily be patched under the DNS tree, e.g. 1234567678.icq.net or mycompanyname.realnames.com.
(Unless I'm mixing up names,) Joe does, however, get Extra Slack points for having been the guy who tied up various Canadian provincial and federal government organizations for a while by constantly faxing them his requests for fair treatment, better laws and regulations, etc.
AT&T Wireless PocketNet service (www.att.ws.com) offers free or flat-rate ($7/mo, $15/mo) wireless internet browsing - the free offer is a limited set of web sites with translation to phone formats, and the two non-free offers provide email, fax, and access to more kinds of Internet sites. data offers; demo thing
CDPD is Cellular Digital Packet Data, which crams data packets around the TDMA digital cellphone space, giving 19.2kbps always-on IP data service. AT&T offers a flat-rate service for about $55/mo, and there are various other service providers that offer per-packet pricing.
Metricom Ricochet radio modems are cool - depending on the model, they range from about 28kbps-equivalent to 128kbps performance for radio-based Internet access. They're mainly located in high-tech areas and big airports, but they've gotten recent investment from MCI, so they're starting to grow a lot. It's a microcell system with pole-top radio pods connected to the network either by radioing to each other to reach wired pods. I'm not sure about the new service - the older modems could switch cells easily at walking speed, but not very well at driving or train speed. Hang one on your laptop and you can work wherever you feel like, or at least stay connected when you head out for coffee, meetings, work in the park, etc.
A long time ago, in a phone company far, far away, though after the Death Star exploded, I was working at Bell Labs in a building with mainframes in the basement and Vaxen and PDP11s scattered all over. A friend of mine was in (or maybe running, by then) the Unix System V porting group, and they did a port of System V Unix to mainframes, and installed it on the Amdahl in my building.
It wasn't perfect - getting backspace-echo to work well on that sort of I/O controller just wasn't going to happen - but it was pretty close, and you could at least use vi. I was taking a compiler course at the time, doing a lot of compilation, and the choice of timesharing a Vax with ~40 people or using the Amdahl with ~2 people was pretty obvious:-)
Why would you port Unix to Big Iron? Well, not only could you use the blazingly fast 10+ MIPS of CPU (when Vaxen were the canonical 1 MIPS), but more importantly, the distributed I/O architecture lets you do immense quantities of disk I/O to run databases. Not only is this Entertaining Research, but it was valuable for phone company billing and equipment-configuration-management applications, allowing more flexible Unix development environments, and it was a much better development environment that Vax-sized machines for the 5ESS phone switch development folks, who needed to compile and build programs that were huge then and large even today.
On the other hand, fsck took a *long* time to run, since the machines had a lot of disks, and this was back when Unix file systems really did need to be checked every time you booted:-) That was one of the things that prompted the development of multi-threaded fsck programs, since checking one at a time was immensely annoying.
David never has better than a 90% chance of beating GoliathCo if he sues them when he's dead right, between luck of the draw and the quality of the lawyers he can afford vs. the quality they can afford. So in a loser-pays court, it's worth GoliathCo's time to spend 10X what David spends on lawyers, partly because they've got positive expected return, and partly because they chase away most of the potential Davids by looking big and scary and by making it clear that not only is losing a lawsuit against them a bad deal, but winning a lawsuit against them takes a much bigger upfront investment to counteract GoliathCo's horde of lawyers.
The good part about this is that it also chases off most of the anklebiters from bringing less-winning lawsuits. Also, when they lose, they've usually chased off the low-value lawsuits, so the cost of David's lawyer is usually much less than the value of the case itself.
If Mean Nasty GoliathCo unfairly sues David, Loser Pays means he's got to come up with an amazingly strong defense or settle right away - because he's got to pay for GoliathCo's Horde Of Lawyers, so even if his attorney is Pro Bono or defending him free-up-front in return for a chance at legal costs if he wins, he's got a good chance of losing big.
On the other hand, if Unfairly Ripped Off GoliathCo legitimately sues Nasty Little David, Loser Pays means they need to keep their costs low because they may not recover much from the case if David settles, which he's got even more incentive to do if he's the Bad Guy. So they're likely to start off with a Threatening Form Letter rather than a well-structured case even when they're right.
So why haven't the Brits descended into pre-Neanderthal chaos? They haven't? (:-) More precisely, it's tough to sue or even criticize misbehaving big companies in the UK, because of Loser Pays. It's especially bad because British libel law doesn't have the American development that truth is an adequate defense against libel, so you can get sued, and lose, for saying true but unfriendly things about them in public. There was a celebrated case a few years ago where somebody's bank card glitched his money, and he tried to get it reimbursed, and they sued him for libeling their Totally Bug Free bank card system, because after all they could lose immense amounts of money if the public thought it could possibly be their fault, therefore it *couldn't* have been their fault, therefore he's liable for immense amounts of money plus legal costs. (I forget the details; I think it may have been Midland Bank, but I could be wrong, and yes, the cash machine system *was* wrong.)
By the way, people also propose "reforming" the tort law system by forbidding lawyers from taking cases on contingency (e.g. no money if they lose, 33% of the take if they win.) This also would make it hard for the little guy to afford to sue the big guy, even if he's willing to risk the Loser Pays costs.
It's one thing to let experts shoot back. It's another thing to make it a widely distributed capability, especially an automated one. Currently a bad guy who wants to run a DDOS needs to crack a few hundred poorly-run machines and then fire up his scripts to abuse them. But if "shoot-back" tools are widely distributed, all he needs to do is find how to forge an attack in a way that will convince a particular shootback tool to attack some victim, and then spam out as many attacks as necessary to get the shootbacks to overwhelm the victim. (Obviously it's still worth doing this from a cracked machine, but you don't have to own a lot of cracked machines to obfuscate yourself.)
This is different from mostly-passive traps like teergrube (FAQ; jargon) or Deception Toolkit or spider traps which sit around waiting for Bad Guys to attack them and react unexpectedly when attacked (e.g....res.p...o...n...d....v...e...r...y....s..l..o.. o...ooo...w...l...y.... while logging stuff or sending back odd replies). ("mostly passive" doesn't exclude leaving lots of inviting copies of your address around for harvesters or script kiddies to find.)
Raph's an inventor - he developed a bunch of cool graphics algorithms useful to the printing industry, patented them, and sold them to the printing companies, and made money off it. Raph also develops Free Software. So if you want to use Raph's algorithms to make money putting out proprietary software, you can license the patents from him. But if you're writing stuff to give away Free(TM) (using the GPL definitions of Free(TM), which make it inconvenient to make money off derivatives of it), you can use his algorithms free.
I'm not quite sure how the Trippy Sunglasses patent can be implemented in software, or whether you can do it in hardware with the Gnu Public License small print on the earpieces:-)
A couple more reasons why universities are often used for attacks
#3 - Insiders vs. Firewalls - Attacks by students. Firewalls are usually designed to keep unauthorised outsiders out. But universities have lots of bright kids with time and computer resources on their hands, who know a lot more about computers than they did in junior high school, know a lot more people who know a lot more about computers, and have a lot more computing resources than when they were using their Mom's AOL account and 486 Win3.1 box. One of the standard computer security problems is "How do you know you're talking to the server you think you're talking to and not to some grad student at Berkeley?" Well, if you're the sysadmin at Berkeley, that's a tough question:-) It's harder than the corporate "disgruntled employee" situation, except that most of your security problem students aren't malicious - they're just more creative than you are....
#4 - Newbies with lots of bandwidth - Most college students aren't experienced computer security experts - they're English Majors, and Chemical Engineers, and MBA-seekers, and pre-law or pre-meds, and Freshman CS Students who aren't all experienced yet, and most of them are running Windows versions that are fundamentally insecure even when administered well. And all these attractive targets are in one place with lots more bandwidth than dialup users and relatively stable IP addresses - so if you crack one of them, you can use it to search for more targets, and it's a lot easier on a campus LAN than in a dialup network. Once you've got your suckers, they can output a lot more bandwidth than AOL newbies you've suckered with a new game program like "Attack On Troy", though networked games are a fun attack at colleges as well - especially high-pressure high-tech schools where students do their recreation intensely as well.
#5 - Not every school is MIT. Podunk Community College may not have quite the same resources to abuse, but it doesn't have the same level of defenses, either, and it may have more resources than half the small ISPs on the market.
#6 - Early Adopters of Networked applications - Universities are great places to distribute things like napster://horse_with_no_name.mp3 and IRCfreefone and Quake 6.2: Mass Destruction and CryptoStealthGnuTella and UsenetPornHider and that eXcellent rave-support tool XFinder. Bad Guys don't need to infect everybody - just enough people to reach critical mass.
It's a target-rich environment out there. We've been lucky so far.
A simple way to get rid of major ad banner sites is to use DNS to map them into nonexistence./etc/hosts files are the easiest approach, though if you run a DNS server you can also become your own SOA for them (which is cleaner on 127.0.0.1, of course.)
makes them point to the nonexistent machine next to you. Alternatively, if you've got a real machine that wasn't already running a web server, you can point it there, especially if you've got a handy 404-returner application to make it fail faster (or to return a blank dummy banner.) Or you can point it to a Class E IP address or a nonexistent machine on your LAN.
If you're running Windows, you'll need to find a way to convince the OS to check your C:\WINDOWS\HOSTS file before checking DNS - or get your DNS admininstrator to help you (or find a DNS server for Windows and run it on 127.0.0.1.)
Another simple method for Windows, if you're inside a corporate firewall and use an explicit proxy server (as opposed to auto-proxy or a transparent proxy) is to list the annoying banner domains in the "Don't Proxy" box in Netscape along with your corporate network. That way it tries to reach annoying-banner.com directly and gets blocked by the firewall instead of proxied. It's not quite as fast as not looking for the banner at all, but usually as fast as or faster than actually retrieving it.
Compare this with a $150 game console or a $400 PC which has a $100 disk drive, some PCI slots, legacy interfaces, $50 power supply, $50-100 Windows license, etc. It's well within the range of practical manufacturing and production, as long as there's enough sales volume to amortize the costs. From a hardware perspective, it really needs USB or Ethernet or another expansion mechanism, but that's not tough to include in that price range.
The question is whether the applications are well designed (and of course whether the marketing is done well.) What software support does one of these things need?
The software platform needs to be open for developers - either something widely supported like WinXX or Linux or *BSD, so existing applications work, or at least a higher level such as JVM where everything's supposed to be easily portable.
Obviously the browser's built in, and relatively persistent bookmarking.
There needs to be some sort of word processor and spreadsheet, and either they need to be small enough that downloading's quick, or else pre-loaded. Sounds like a job for Java?
There needs to be a good mechanism for caching downloaded applications.
There needs to be a good way to use an Internet file storage provider - probably easiest to piggyback off XDrive/IDrive/Driveway/50Megs/etc. - but is there another mechanism to use instead? Is non-free storage part of the business model?
Security is essential - do any storage providers support SSL?
Games - If you're going for the home-user market, and supporting television display interfaces and modem network connections, that's an important market growth mechanism. You' re not just selling them to schools and businesses, you're selling them to Nintendo/Sega/Sony consumers, so you need some standalone games and some networked games.
Getting approval to use open-source software for internal use should be easy, though it's more trouble if your company makes commercial software because of different open-source licenses (espe ially GPL) affecting the software you sell, if you're not careful about interfaces. But the more difficult part is getting approval to publish software you've written as open-source. Other than companies in the business of providing and supporting open-source software (e.g. RH, Cygnus, *linux, *bsd), I'd be interested in hearing people's experience getting open-source out of their companies.
The standard boss-convincement mechanism is to write a business case - estimate the benefits and costs in a one-pager, with maybe several pages of backup data to accompany it. It's not just PHBs that need that - you're making effective use of your boss's time.
Unix systems being what they are, it's often pretty straightforward to put together a prototype in a short time, or point to URLs for a couple of similar projects that other people have done. This lets your business case say "Expand the demo that we did in an afternoon which did 80% of the functionality to a full system", which is a much stronger postition than "Start something unknown and untrusted from scratch", though of course the last 20% of the work takes much longer than the first 80%.
A decade or less ago, there was a computer out there called "Brick", which was a slightly larger version of this. No monitor, no batteries, but easy to carry around between places where you've got a monitor and keyboard, so you can do demos, read email, etc.
I commute by train, so I want something I can read my email with, and this doesn't quite cut it for me, but it's still intriguing. (And of course, if it were $300 instead of $900, I might very well buy it in spite of its limitations:-)
The URL in the announcement was a moving-target pointer to the White House Press Releases, so today's 0th press release is something about www.americasteens.gov, a Federal program to prevent the corruption of our kids' precious bodily fluids or something. If you dredge the pointers to previous days, you get a probably-moving-target pointer http://www.whitehouse.gov/library/PressReleases. cgi?date=2&briefing=5 , which at least tonight points to the real site www.igeb.gov
So the picture is pretty, and it appears to have a screen, modem, and browser, and costs a small amount of money in return for lots of marketing tracking data. But what can you *do* with it? Is there a disk or other file system, or can you add one? Can you install software? Does the browser support Java? Can you turn Javascript/ActiveX/Dangerware off? Can you use real email, or only webemail?
How long can you stay dialed up and inactive - does "10 hours per month" mean you need to tap the keyboard every few minutes, or can you dial up one evening a month and let it stay on all night?
Can you dial into other ISPs? (e.g. dial into your home Linux machine's modem?) If you do all your browsing through The Anonymizer, can they track anything?
HTML, XML, and their parent SGML are content description languages - they describe document content entities such as paragraphs, headings, lists and tables, but don't describe how to make black marks on paper or RGB marks on CRTs.
They're well-suited to automated layout programs , letting the document reader determine how to lay out the visuals, which may be different depending on whether the target reading environment is hi-res dead tree phototypesetters, medium-res CRTs, text-to-speech readers, braillewriters, cell-phone microscreens, PDA mini-screens, dumb terminals, browsers with images turned off for speed or font set really large for low-sight people or really small for pocket-sized printouts, etc. SGML was the original flexible metalanguage; HTML is a simplified static instantiation, as are the cell-phone variants, and XML is a newer SGML variant that's learned from 15 years of real-world experience.
They're also well-suited for automated content handling, such as the XML developments that are replacing EDI for applications like purchase orders. Editing CDL documents is easy, as long as you stick to the defined structures.
Page Description Languages
PDLs let authors tell the computer how to make documents look the way they want them to, and programs that process them make various compromises to support different presentation media, such as paper or CRTs. They range from things like Postscript, which forces the display to do the best job it can rendering an image that the authoring program specifies, to things like MSWord, which let the display device determine layout, and reprocess the entire input document any time you change printers. They also range from higher-level systems which know a lot about the document structure to lower-level systems that know about output but know next to nothing about structure - "systems" includes the application programs as well as the representation language.
All of them have the problem that if you want to edit the contents, the output looks different so they have to cope with what you've done. Depending on the document structure and app design, they may have to repaginate the entire document, or only up to a chapter/section break, or they may be crude and only patch the current page and force you to renumber the rest if you want.
Hybrids
Lots of authors want to specify the output appearance, regardless of whether this constrains the readers' choices - you can see hybridization like this crunched into HTML, with commands for fonts, font sizes, colors, and the newer cascading style sheet stuff. It's possible to do this in ways that preserve content - the language represents that this is a "Heading Type 2", and instructs that "Heading Type 2" be represented in Palatino Bold Blinking with a full line-break after the heading text. It's also depressingly common to lose content structure information, especially during translations, either because the target language doesn't have a mechanism for representing the content tags, or because the translator writer JUST DOESN'T GET IT. An example of the former problem is rendering for constant-width ASCII or for GIFs or Faxes. A depressingly stupid example of the latter is saving MSWord documents in HTML - MSWord knows about objects like headings and paragraphs, and knows that the current user's settings for a "Heading 2" object and "Normal Paragraph" object are 14-point Arial Bold followed by a single blank line and 10-point Times Roman followed by a single blank line - but instead of outputting an H2 heading object, a P paragraph marker, the text, and another P, the depressingly stupid program outputs a request for 14-point Arial Bold font, the header text, a couple of BR spaces, a request for 10-point Times Roman font, the paragraph text, and some more spaces.
Application Program Dependence
Application Programs can do lots of different things with PDLs or CDLs. For instance, you can use comments to put non-printed document structure information into PDLs, or to put layout information into CDLs, and programs that know about it will use it, while programs that don't know about it will ignore it. That doesn't mean there are any industry standards about doing this, so of course one editor may stomp all over another's markings, or may leave them in place while adding things the first program doesn't notice because the comments weren't updated. Postscript is an egregious contributor to this - it's an extremely general-purpose programming language, and there are lots of different ways to get the same set of black marks on paper, ranging from bitmaps to format-annotated CDLs, and almost no two applications can read each other's Postscript.
Production Software
Framemaker
While PDFs are both liked and disliked because they are designed not to be editable, I'm surprised your customer couldn't accept FrameMaker. It's one of the best WYSIWYG large-document production systems I've seen over the last decade, and if the customer wants to export pieces into MSFoo, they can, but if edit the entire document, they probably should have enough control over the process that they can buy a few copies of Frame for what's basically a trivial addition to the cost they've already paid for producing 3000 pages of documentation. Also, to do big documents, you need tools that can cope with multiple authors working simultaneously, and Word isn't really designed for that.
Alternatives
If you're doing 3000 pages of documentation, or for that matter 300, and you're not a graphic arts shop or something, it's probably going to be mostly text, or text with user-interface illustrations, and you're going to use a uniform formatting style for the whole thing, modulo a few tweaks for illustrations that need to be placed on a page to fit together with an occasional tweak. I'd think the best approaches these days are either to build the thing in hypertext to start with, or else use some of the tools from the GNU / Emacs world, or else use a batch production system like LaTeX or troff with an appropriate macro set. Learning HTML was soeasy, since it looked a lot like the Troff -mm Macros:-)
It's been a long time since I've been part of anything over a hundred pages or so; the last time I was on a very large RFP response project, the boss had some kinky troff macros and basic shell mungers that let us keep the entire document in a database, so we could track which of the N thousand requirements were being handled by which authors, in what files, patch up the figure and page-number references (really a two-pass process) , and build the indexes to the document and the crossreferences to the RFP requirements document it was a response to.
Subject to heavy regulations on what you must carry and the terms you must carry it under.
Somebody else pointed out that Common Carrier rules, if applied to the email world, wouldn't let an ISP refuse to carry spam. They also wouldn't be able to make whatever policies they wanted about what kinds of traffic they carry, probably couldn't offer censored services for the parents that want it, probably couldn't do free-service deals with some partners and not others, probably couldn't have a special "clean-up-the-mess" fee for spammers, etc.
Common carrier status also might affect the ability to offer various privacy services - can you support anonymous users? Must you treat all your users anonymously? Do you need to collect ID information and communication logs on all users so the Fedz can track people they don't like? This is especially an issue for free internet access services, where collecting user information is a major marketing opportunity, but verifying it is a major cost, and for ISPs that want to provide access for kids, where there are special rules about handling information collected on them that may contradict other proposed rules mandating information. Do you allow anybody to claim to be under 13 (and hence non-loggable), or do you insist that they get a "parent"'s signature saying they're a kid?
Common carriers traditionally have to make all their pricing policies public, open to anybody, filed in advance and subject to regulatory approval. Do you want this in your business?
Much nicer to avoid the whole regulatory game, which exists largely to help monopolies and near-monopolies use political influence to restrict their competition anyway:-)
Actually the proposals were to use the cell system for location, rather than GPS, for a 125-meter (125-foot?) radius. Most digital cell technology can get the accuracy if you wrap enough coordination and processing betwen the cell sites (you might locate _them_ with DGPS, for instance.) GPS is a power hog, so you don't want to use it if you can avoid it; most GPS receivers get about 24 hours per set of batteries, while digital cell phones can last a week.
This doesn't mean that *you*, the cell-phone user can locate where you are - it just means that the phone company can, so that 911 can locate you (if you're not paranoid, and for some reason believe the official explanations), or so anybody with a badge can locate you (if you *are* paranoid), or so any 2600-script-kiddie or at least any good social engineer can locate you.
If you do want to check out the paranoia options, spend a while thinking about the requirements that the 911 center be able to locate you any time your phone is on without you acknowledging it, and the lack of requirements for a standard locatee user interface.....
ANonymous Coward just posted a good explanation in #81 I'm not a moderator today, so I'll quote it, but you can go moderate the original up if you'd like. -------------- from Trimble Navigation, "Differential GPS Explained":
Summary of GPS Error Sources:
Per Satellite Accuracy Standard GPS Satellite Clocks 1.5 m Orbit Errors 2.5 m Ionosphere 5.0 m Troposphere 0.5 m Receiver Noise 0.3 m Multipath 0.6 m SA 30 m
That was Spamford, a few years back, after he'd been blasted off every other ISP. He failed miserably. Being an "ISP" is no good is you can't connect to any other ISP and all your customers are spammers, not spammees. AGIS tolerated him longer than most of his service providers did, and nearly died from the ensuing boycotts and flame wars. (They're gone now.) He and the Cantor&Siegel Green Card Spam Lawyers pretty much invented spamming as an internet industry; he'd also been a major player in fax spam before that.
The recent edition of Wired has a set of articles on machine translation. Certainly Babelfish is a helpful start, and there's a large economic incentive for at least basic translation capabilities. (In practice I tend to use it as a crutch for my bad German and worse French, but it'll get better and handle more languages.)
An alternative to machine translation is access to real human translators. Language Line (started as an AT&T business) provides telephone access to translators for a large number of languages. The Internet will make it easier to access translators for a much wider set of languages, both for real-time translation and non-realtime. This is especially useful for finding native speakers of non-English languages to translate into those languages, which generally produces higher quality than non-native speakers who understand the source language well. Of course, for translating technical documentation on complex things, you need to find translators who understand the subject matter as well, and the ability of the Internet to access a large number of people makes this more convenient.
Hopefully the days of anyone ruling are over. (Fukuyama's The End Of History argues that this sort of thing is obsolete.) English's prominence is largely from British Empire conquests as well as US trade, and of course from the decline of French colonial and economic power, though French is still the language to use in much of Africa. The primary alternative will be machine translation, probably with English as an intermediate for most pairs of non-European languages.
How else do languages spread widely, other than conquest and trade? Religion is one way, though that's unfortunately related to conquest. Until Vatican II, Latin was pretty universally findable - most people might not know more than the Latin they used in church, but you could at least find a local priest if you had to get communication done. Arabic is also widespread, not only in the Middle East and North Africa, but also in Islamic Southeast Asian island areas such as Indonesia, and becoming wider spread in the US. I've heard it argued that Hebrew fills some of the same function, though I'm not sure how true this is outside of Europe, North Africa, and North America.
One of the widely used methods of communications between North America Indian tribes was sign language. IIRC, the Kiowa believe it originally came from their ancestors and spread out to the other tribes. It's obviously not a direct map to what we do on the Internet today, though some icon things fill a similar function.
Slashdot Mirror
There have been several proposals for adding more TLDs - the IAHC International Ad-Hoc Committee was relatively reasonable, ICANN was a bit less so, thoguh that was partly because it was a year or two later so there was more commercial conflict, and Esther (bless her heart) knew it would be a dirty job when she took it. As far as I know, the only proposals for new TLDs that have actually succeeded have been a few new country codes (because there's an existing bureaucracy for that, plus of course the countries who've made a quick buck by renting out their namespace), and Brad Templeton's proposal for
There have also been the disorganized proposals, from people like Kaspureff at Alternic, and the orange.net folks - start an alternative root, and try to convince people to use your root instead of the Big Roots, but they're fighting a losing game. It's partly a losing game because they've been losing (:-), and partly because it doesn't solve the fundamental problem, it just trashes any efficiencies you gain by shoving conflicting names down a layer in the tree so you don't see them if you're not looking for them.
The people who've been successful at pushing new namespaces have taken different approaches - ICQ numbers are a global namespace, and nobody minds because they don't spell anything and the server can cope with the scale. Realnames sells namespace, and people who want it can use it. And all of these things can easily be patched under the DNS tree, e.g. 1234567678.icq.net or mycompanyname.realnames.com.
(Unless I'm mixing up names,) Joe does, however, get Extra Slack points for having been the guy who tied up various Canadian provincial and federal government organizations for a while by constantly faxing them his requests for fair treatment, better laws and regulations, etc.
The Crusoe-powered S3 internet appliance on Transmeta.com looks suspiciously like an Etch-A-Sketch.
data offers;
demo thing
CDPD is Cellular Digital Packet Data, which crams data packets around the TDMA digital cellphone space, giving 19.2kbps always-on IP data service. AT&T offers a flat-rate service for about $55/mo, and there are various other service providers that offer per-packet pricing.
Metricom Ricochet radio modems are cool - depending on the model, they range from about 28kbps-equivalent to 128kbps performance for radio-based Internet access. They're mainly located in high-tech areas and big airports, but they've gotten recent investment from MCI, so they're starting to grow a lot. It's a microcell system with pole-top radio pods connected to the network either by radioing to each other to reach wired pods. I'm not sure about the new service - the older modems could switch cells easily at walking speed, but not very well at driving or train speed. Hang one on your laptop and you can work wherever you feel like, or at least stay connected when you head out for coffee, meetings, work in the park, etc.
It wasn't perfect - getting backspace-echo to work well on that sort of I/O controller just wasn't going to happen - but it was pretty close, and you could at least use vi. I was taking a compiler course at the time, doing a lot of compilation, and the choice of timesharing a Vax with ~40 people or using the Amdahl with ~2 people was pretty obvious
Why would you port Unix to Big Iron? Well, not only could you use the blazingly fast 10+ MIPS of CPU (when Vaxen were the canonical 1 MIPS), but more importantly, the distributed I/O architecture lets you do immense quantities of disk I/O to run databases. Not only is this Entertaining Research, but it was valuable for phone company billing and equipment-configuration-management applications, allowing more flexible Unix development environments, and it was a much better development environment that Vax-sized machines for the 5ESS phone switch development folks, who needed to compile and build programs that were huge then and large even today.
On the other hand, fsck took a *long* time to run, since the machines had a lot of disks, and this was back when Unix file systems really did need to be checked every time you booted
The good part about this is that it also chases off most of the anklebiters from bringing less-winning lawsuits. Also, when they lose, they've usually chased off the low-value lawsuits, so the cost of David's lawyer is usually much less than the value of the case itself.
If Mean Nasty GoliathCo unfairly sues David, Loser Pays means he's got to come up with an amazingly strong defense or settle right away - because he's got to pay for GoliathCo's Horde Of Lawyers, so even if his attorney is Pro Bono or defending him free-up-front in return for a chance at legal costs if he wins, he's got a good chance of losing big.
On the other hand, if Unfairly Ripped Off GoliathCo legitimately sues Nasty Little David, Loser Pays means they need to keep their costs low because they may not recover much from the case if David settles, which he's got even more incentive to do if he's the Bad Guy. So they're likely to start off with a Threatening Form Letter rather than a well-structured case even when they're right.
So why haven't the Brits descended into pre-Neanderthal chaos? They haven't? (:-) More precisely, it's tough to sue or even criticize misbehaving big companies in the UK, because of Loser Pays. It's especially bad because British libel law doesn't have the American development that truth is an adequate defense against libel, so you can get sued, and lose, for saying true but unfriendly things about them in public. There was a celebrated case a few years ago where somebody's bank card glitched his money, and he tried to get it reimbursed, and they sued him for libeling their Totally Bug Free bank card system, because after all they could lose immense amounts of money if the public thought it could possibly be their fault, therefore it *couldn't* have been their fault, therefore he's liable for immense amounts of money plus legal costs. (I forget the details; I think it may have been Midland Bank, but I could be wrong, and yes, the cash machine system *was* wrong.)
By the way, people also propose "reforming" the tort law system by forbidding lawyers from taking cases on contingency (e.g. no money if they lose, 33% of the take if they win.) This also would make it hard for the little guy to afford to sue the big guy, even if he's willing to risk the Loser Pays costs.
This is different from mostly-passive traps like teergrube (FAQ; jargon) or Deception Toolkit or spider traps which sit around waiting for Bad Guys to attack them and react unexpectedly when attacked (e.g.
I'm not quite sure how the Trippy Sunglasses patent can be implemented in software, or whether you can do it in hardware with the Gnu Public License small print on the earpieces
- #3 - Insiders vs. Firewalls - Attacks by students. Firewalls are usually designed to keep unauthorised outsiders out. But universities have lots of bright kids with time and computer resources on their hands, who know a lot more about computers than they did in junior high school, know a lot more people who know a lot more about computers, and have a lot more computing resources than when they were using their Mom's AOL account and 486 Win3.1 box. One of the standard computer security problems is "How do you know you're talking to the server you think you're talking to and not to some grad student at Berkeley?" Well, if you're the sysadmin at Berkeley, that's a tough question
:-) It's harder than the corporate "disgruntled employee" situation, except that most of your security problem students aren't malicious - they're just more creative than you are....
- #4 - Newbies with lots of bandwidth - Most college students aren't experienced computer security experts - they're English Majors, and Chemical Engineers, and MBA-seekers, and pre-law or pre-meds, and Freshman CS Students who aren't all experienced yet, and most of them are running Windows versions that are fundamentally insecure even when administered well. And all these attractive targets are in one place with lots more bandwidth than dialup users and relatively stable IP addresses - so if you crack one of them, you can use it to search for more targets, and it's a lot easier on a campus LAN than in a dialup network. Once you've got your suckers, they can output a lot more bandwidth than AOL newbies you've suckered with a new game program like "Attack On Troy", though networked games are a fun attack at colleges as well - especially high-pressure high-tech schools where students do their recreation intensely as well.
- #5 - Not every school is MIT. Podunk Community College may not have quite the same resources to abuse, but it doesn't have the same level of defenses, either, and it may have more resources than half the small ISPs on the market.
- #6 - Early Adopters of Networked applications - Universities are great places to distribute things like napster://horse_with_no_name.mp3 and IRCfreefone and Quake 6.2: Mass Destruction and CryptoStealthGnuTella and UsenetPornHider and that eXcellent rave-support tool XFinder. Bad Guys don't need to infect everybody - just enough people to reach critical mass.
It's a target-rich environment out there. We've been lucky so far.makes them point to the nonexistent machine next to you. Alternatively, if you've got a real machine that wasn't already running a web server, you can point it there, especially if you've got a handy 404-returner application to make it fail faster (or to return a blank dummy banner.) Or you can point it to a Class E IP address or a nonexistent machine on your LAN.
If you're running Windows, you'll need to find a way to convince the OS to check your C:\WINDOWS\HOSTS file before checking DNS - or get your DNS admininstrator to help you (or find a DNS server for Windows and run it on 127.0.0.1.)
Another simple method for Windows, if you're inside a corporate firewall and use an explicit proxy server (as opposed to auto-proxy or a transparent proxy) is to list the annoying banner domains in the "Don't Proxy" box in Netscape along with your corporate network. That way it tries to reach annoying-banner.com directly and gets blocked by the firewall instead of proxied. It's not quite as fast as not looking for the banner at all, but usually as fast as or faster than actually retrieving it.
The question is whether the applications are well designed (and of course whether the marketing is done well.) What software support does one of these things need?
re not just selling them to schools and businesses, you're selling them to Nintendo/Sega/Sony consumers, so you need some standalone games and some networked games.
Also, it'd be interesting after you're done to get your opinions on the experience - hosting centers are a major part of the Internet these days.
Unix systems being what they are, it's often pretty straightforward to put together a prototype in a short time, or point to URLs for a couple of similar projects that other people have done. This lets your business case say "Expand the demo that we did in an afternoon which did 80% of the functionality to a full system", which is a much stronger postition than "Start something unknown and untrusted from scratch", though of course the last 20% of the work takes much longer than the first 80%.
I commute by train, so I want something I can read my email with, and this doesn't quite cut it for me, but it's still intriguing. (And of course, if it were $300 instead of $900, I might very well buy it in spite of its limitations
The URL in the announcement was a moving-target pointer to the White House Press Releases,. cgi?date=2&briefing=5 , which at least tonight points to the real site
so today's 0th press release is something about www.americasteens.gov, a Federal program to prevent the corruption of our kids' precious bodily fluids or something. If you dredge the pointers to previous days, you get a probably-moving-target pointer
http://www.whitehouse.gov/library/PressReleases
www.igeb.gov
How long can you stay dialed up and inactive - does "10 hours per month" mean you need to tap the keyboard every few minutes, or can you dial up one evening a month and let it stay on all night?
Can you dial into other ISPs? (e.g. dial into your home Linux machine's modem?) If you do all your browsing through The Anonymizer , can they track anything?
Content Description Languages (CDLs)
HTML, XML, and their parent SGML are content description languages - they describe document content entities such as paragraphs, headings, lists and tables, but don't describe how to make black marks on paper or RGB marks on CRTs.
They're well-suited to automated layout programs , letting the document reader determine how to lay out the visuals, which may be different depending on whether the target reading environment is hi-res dead tree phototypesetters, medium-res CRTs, text-to-speech readers, braillewriters, cell-phone microscreens, PDA mini-screens, dumb terminals, browsers with images turned off for speed or font set really large for low-sight people or really small for pocket-sized printouts, etc. SGML was the original flexible metalanguage; HTML is a simplified static instantiation, as are the cell-phone variants, and XML is a newer SGML variant that's learned from 15 years of real-world experience.
They're also well-suited for automated content handling, such as the XML developments that are replacing EDI for applications like purchase orders. Editing CDL documents is easy, as long as you stick to the defined structures.
Page Description Languages
PDLs let authors tell the computer how to make documents look the way they want them to, and programs that process them make various compromises to support different presentation media, such as paper or CRTs. They range from things like Postscript, which forces the display to do the best job it can rendering an image that the authoring program specifies, to things like MSWord, which let the display device determine layout, and reprocess the entire input document any time you change printers. They also range from higher-level systems which know a lot about the document structure to lower-level systems that know about output but know next to nothing about structure - "systems" includes the application programs as well as the representation language.
All of them have the problem that if you want to edit the contents, the output looks different so they have to cope with what you've done. Depending on the document structure and app design, they may have to repaginate the entire document, or only up to a chapter/section break, or they may be crude and only patch the current page and force you to renumber the rest if you want.
Hybrids
Lots of authors want to specify the output appearance, regardless of whether this constrains the readers' choices - you can see hybridization like this crunched into HTML, with commands for fonts, font sizes, colors, and the newer cascading style sheet stuff. It's possible to do this in ways that preserve content - the language represents that this is a "Heading Type 2", and instructs that "Heading Type 2" be represented in Palatino Bold Blinking with a full line-break after the heading text. It's also depressingly common to lose content structure information, especially during translations, either because the target language doesn't have a mechanism for representing the content tags, or because the translator writer JUST DOESN'T GET IT. An example of the former problem is rendering for constant-width ASCII or for GIFs or Faxes. A depressingly stupid example of the latter is saving MSWord documents in HTML - MSWord knows about objects like headings and paragraphs, and knows that the current user's settings for a "Heading 2" object and "Normal Paragraph" object are 14-point Arial Bold followed by a single blank line and 10-point Times Roman followed by a single blank line - but instead of outputting an H2 heading object, a P paragraph marker, the text, and another P, the depressingly stupid program outputs a request for 14-point Arial Bold font, the header text, a couple of BR spaces, a request for 10-point Times Roman font, the paragraph text, and some more spaces.
Application Programs can do lots of different things with PDLs or CDLs. For instance, you can use comments to put non-printed document structure information into PDLs, or to put layout information into CDLs, and programs that know about it will use it, while programs that don't know about it will ignore it. That doesn't mean there are any industry standards about doing this, so of course one editor may stomp all over another's markings, or may leave them in place while adding things the first program doesn't notice because the comments weren't updated. Postscript is an egregious contributor to this - it's an extremely general-purpose programming language, and there are lots of different ways to get the same set of black marks on paper, ranging from bitmaps to format-annotated CDLs, and almost no two applications can read each other's Postscript.
Production Software
While PDFs are both liked and disliked because they are designed not to be editable, I'm surprised your customer couldn't accept FrameMaker. It's one of the best WYSIWYG large-document production systems I've seen over the last decade, and if the customer wants to export pieces into MSFoo, they can, but if edit the entire document, they probably should have enough control over the process that they can buy a few copies of Frame for what's basically a trivial addition to the cost they've already paid for producing 3000 pages of documentation. Also, to do big documents, you need tools that can cope with multiple authors working simultaneously, and Word isn't really designed for that.
If you're doing 3000 pages of documentation, or for that matter 300, and you're not a graphic arts shop or something, it's probably going to be mostly text, or text with user-interface illustrations, and you're going to use a uniform formatting style for the whole thing, modulo a few tweaks for illustrations that need to be placed on a page to fit together with an occasional tweak. I'd think the best approaches these days are either to build the thing in hypertext to start with, or else use some of the tools from the GNU / Emacs world, or else use a batch production system like LaTeX or troff with an appropriate macro set.
Learning HTML was soeasy, since it looked a lot like the Troff -mm Macros
It's been a long time since I've been part of anything over a hundred pages or so; the last time I was on a very large RFP response project, the boss had some kinky troff macros and basic shell mungers that let us keep the entire document in a database, so we could track which of the N thousand requirements were being handled by which authors, in what files, patch up the figure and page-number references (really a two-pass process) , and build the indexes to the document and the crossreferences to the RFP requirements document it was a response to.
Arrrgh - Slashdot doesn't let me use H1 and H2
Somebody else pointed out that Common Carrier rules, if applied to the email world, wouldn't let an ISP refuse to carry spam. They also wouldn't be able to make whatever policies they wanted about what kinds of traffic they carry, probably couldn't offer censored services for the parents that want it, probably couldn't do free-service deals with some partners and not others, probably couldn't have a special "clean-up-the-mess" fee for spammers, etc.
Common carrier status also might affect the ability to offer various privacy services - can you support anonymous users? Must you treat all your users anonymously? Do you need to collect ID information and communication logs on all users so the Fedz can track people they don't like? This is especially an issue for free internet access services, where collecting user information is a major marketing opportunity, but verifying it is a major cost, and for ISPs that want to provide access for kids, where there are special rules about handling information collected on them that may contradict other proposed rules mandating information. Do you allow anybody to claim to be under 13 (and hence non-loggable), or do you insist that they get a "parent"'s signature saying they're a kid?
Common carriers traditionally have to make all their pricing policies public, open to anybody, filed in advance and subject to regulatory approval. Do you want this in your business?
Much nicer to avoid the whole regulatory game, which exists largely to help monopolies and near-monopolies use political influence to restrict their competition anyway :-)
Where have you been travelling lately?
This doesn't mean that *you*, the cell-phone user can locate where you are - it just means that the phone company can, so that 911 can locate you (if you're not paranoid, and for some reason believe the official explanations), or so anybody with a badge can locate you (if you *are* paranoid), or so any 2600-script-kiddie or at least any good social engineer can locate you.
If you do want to check out the paranoia options, spend a while thinking about the requirements that the 911 center be able to locate you any time your phone is on without you acknowledging it, and the lack of requirements for a standard locatee user interface.....
ANonymous Coward just posted a good explanation in #81
I'm not a moderator today, so I'll quote it, but you can go moderate the original up if you'd like.
--------------
from Trimble Navigation, "Differential GPS Explained":
Summary of GPS Error Sources:
Per Satellite Accuracy Standard GPS
Satellite Clocks 1.5 m
Orbit Errors 2.5 m
Ionosphere 5.0 m
Troposphere 0.5 m
Receiver Noise 0.3 m
Multipath 0.6 m
SA 30 m
Typical Accuracy 50 m
That was Spamford, a few years back, after he'd been blasted off every other ISP. He failed miserably. Being an "ISP" is no good is you can't connect to any other ISP and all your customers are spammers, not spammees. AGIS tolerated him longer than most of his service providers did, and nearly died from the ensuing boycotts and flame wars. (They're gone now.) He and the Cantor&Siegel Green Card Spam Lawyers pretty much invented spamming as an internet industry; he'd also been a major player in fax spam before that.
An alternative to machine translation is access to real human translators. Language Line (started as an AT&T business) provides telephone access to translators for a large number of languages. The Internet will make it easier to access translators for a much wider set of languages, both for real-time translation and non-realtime. This is especially useful for finding native speakers of non-English languages to translate into those languages, which generally produces higher quality than non-native speakers who understand the source language well. Of course, for translating technical documentation on complex things, you need to find translators who understand the subject matter as well, and the ability of the Internet to access a large number of people makes this more convenient.
English's prominence is largely from British Empire conquests as well as US trade, and of course from the decline of French colonial and economic power, though French is still the language to use in much of Africa.
The primary alternative will be machine translation, probably with English as an intermediate for most pairs of non-European languages.
How else do languages spread widely, other than conquest and trade? Religion is one way, though that's unfortunately related to conquest.
Until Vatican II, Latin was pretty universally findable - most people might not know more than the Latin they used in church, but you could at least find a local priest if you had to get communication done. Arabic is also widespread,
not only in the Middle East and North Africa, but also in Islamic Southeast Asian island areas such as Indonesia, and becoming wider spread in the US.
I've heard it argued that Hebrew fills some of the same function, though I'm not sure how true this is outside of Europe, North Africa, and North America.