The Internet is not e-mail! It is completely inappropriate to base the DNS name of your organization on what is effectively a content label specific to one particular service. This is the same reason.kids and.xxx are bad.
Heck, let's say I run a porn service, and want to take advantage of this mail feature. I now have to use two different DNS domains? That's stupid.
Just as PICS can give you digitally-signed content ratings for the web, some other service can give you digitally-signed ratings/labels for e-mail. Extend SMTP to, perhaps, operate over TLS or SSL, or at least perform some sort of mutual check that both sides have a SpamHaus certificate that says they're not a spammer, and you can possibly "secure" the connection.
Or just digitally sign your e-mail messages and only accept digitally-signed e-mail. Tweak your trust relationships (for PGP-style signatures) or drop your trust from any roots that are seen to sponsor spammers, and you're all set.
DNS is not a content label. The Internet is not the web. Why in the world should an organization on the Internet be required to parent themselves within DNS under.xxx or.sex just because one server somewhere in their organization happens to provide an HTTP service that contains content that happens to satisfy someone's definition of porn?
This extends to Internet providers. Should "Example Internet" (example.net) need to relocate under "example-internet.xxx" because one of their customers happens to have a nudie pic on his web site (http://users.example.net/~joe-user)?
Use PICS labels for labelling content. Use DNS domains to provide a logical, hierarchical name for your organization and network resources within your organization.
There's no law anywhere that says I have to have a web site on my DNS domain (WIPO/ICANN judgements not withstanding), so why apply HTTP content labels to your DNS domain?
Even so, I could have a fairly extensive, mature set of non-HTTP services running under my DNS domain and one "nothing here yet" document served over HTTP. What does it matter?
Making DNS flat isn't the answer. You're pushing requests and resources towards the root of the DNS tree, at a tremendous expense of the root name server operators. You can't just delegate a piece of the root to the companies that own that TLD. Flattening the namespace means flattening the control structure of the DNS as a whole; moving the distributed infrastructure to one freakin' mammoth root infrastructure.
This is not what DNS is suited for. It will not make this change gracefully.
Ten or fifty new TLDs won't help things: Microsoft will buy a domain under every one, and people will begin to wonder why their "Internet name" starts with their own name and ends with this arbitrary, cryptic label.
DNS simply is not suited for use as a Intellectual Property label, which is precisely why so much litigation goes on.
However, if Ford didn't have a website, and I registered Ford.com, and just had a parked page there, it would be pretty clear that I was doing it in the hope of extorting money from Ford.
It's not that easy, though. The Internet is not the web!
I could just as easily registered ford.com for my organization but not provided any HTTP services for that domain. Maybe I wanted it for e-mail only, or for some other services like VoIP. There's no law that says I have to have a web page at www.ford.com.
Unfortunately entities like WIPO seem to agree with your perspective, though. Internet domains are for web pages, and if you don't have a really good web page at your Internet domain, then you must be a cybersquatter, and you should be forced to give the other guy your domain so that he can put a nice web page there. It disgusts me that Corporate and Legal America not only doesn't get this, but that they are able to receive judgements based on this distortion.
if my last name was Ford, and I registered ford.com, and on it I had a huge website, with pages for every member of my family, and a history of the family, and all my relatives had @ford.com e-mail addresses, and then Ford suddenly decides they want to cash in on this e-commerce thing and sues me, well that's a sticky situation.
Exactly: in order to have a good claim on a DNS domain, you've gotta have a fancy web page there.
I might suggest that one other thing would have prevented this problem far earlier than your two important things: the creation of a directory that proposed to map legally-protected real-world names to Internet entities (DNS domains, etc.). DNS is far too simplistic to be warped into a label deserving of Intellectual Property protection. Another layer is needed to translate the world of Marketing and Legal to the world of IT.
Most of the companies providing Internet service are in it to make money. They are in turn driven by their shareholders to make even more money. Their marketing department and executives believe that they have some ideas and the weight to make it work and tell their IT department to figure out how to do it.
IT people work for companies. Companies work for their shareholders. They are not obligated to follow RFCs or to respect anyone else in the process (at least to the point where their Internet peers stop talking to them).
Maybe what we need is an organization like the IETF to provide annual "certifications of compliance" so that companies can get a digital certificate saying that they are adhering to RFCs and aren't acting in a fashion that's destructive to the Internet. Border routers and gateways between Internet peers would be required to have a valid certificate for the entity on each interface before it would route traffic coming from that interface.
But then, the companies making routers are in it to make money as well...
The Internet will adapt to market forces. It's going to do that badly, through resistence, or peacefully, by coming up with a proper solution. So far we've all been letting megacorps twist and litigate DNS into a bloody pulp, and the problem will persist until we stop crying and give the megacorps a better solution to the problem of a human-friendly name/servicemark/trademark lookup service.
This is a very good point that can't be overstated.
The definition of what a "generic" name is, or the geographic or marketing reach of a real-world name like Apple is a fuzzy concept.
Internet naming is a concrete concept, and a given DNS name needs to be persistent. We can't have people hitting the "Apple" TLD (or apple.com) in one region and hitting their local Apple Supermarket but in another region hit Apple Computers. A single DNS domain needs to always take you to the same logical entity.
So how do you map a fuzzy concept like that onto a concrete one? Another layer. An authoritative search engine, to be precise. You tell it your search terms "Apple", maybe provide it some hints about your location or the type of Apple you want to see, and it returns the DNS domain of the entity you're trying to reach. This is a very simple application of X.500 or LDAP. We just need someone (ICANN?) to step up and provide this service.
Heck, the registrars already have 90% of this information in the form of WHOIS records.
DNS is a hierarchy. It does lookups with a given key (FQDN), not searches. Bind (and other DNS servers) performs its function efficiently and effectively. There's no problem with DNS.
The problem is that the Marketing Departments want to pretend that DNS is "Internet Keywords", and because there's no alternative standard in place for doing the types of name-to-web-site lookups that they want, they are successfully lobbying to warp DNS into Internet Keywords.
Now, I do think it's worth investigating an alternative or supplement to DNS, but even then it doesn't need to be relational. X.500 or LDAP would do the job very nicely. We need an authoritative lookup service to map these real-world names that Marketing and Legal want to use/defend to a DNS domain (or directly to an IP address and service if you want to replace DNS entirely).
But don't try and graft this onto DNS, as it's completely inappropriate. DNS is a naming service, not a locator or search service.
The problem with this is that most businesses, by virtue of putting a shopping cart on a web site, automatically go global. They may be based in some city somewhere, but through their web site, they're doing business everywhere in the world.
The reality is that the gTLDs have no semantic value anymore. They might as well be "balloon", "cat" and "fast". Only the whole "dot com" phrase gives.com more value than the others.
Even.us has become what corporate America wants: a flat namespace. It used to be geographic (and I still have an old geographically-based DNS domain within it), but now anyone can get example.us with no problem. So it's as useless as the gTLDs.
DNS needs to either be replaced, or supplemented with something that's more appropriate to be a legally-enforced (intellectual property-friendly) label.
Also, what you're observing here is a web-only service. The Internet is not the web! We don't need a mapping between keyword and URL, we need a mapping between a real-world legal name and an Internet *domain*. Nothing about this is application-specific, unlike Internet Keywords. Other lookup services (like DNS SRV) would then map the DNS domain to a service (like a web site).
When I type it in Mozilla, I get an error saying the URL is invalid. If I search via Google, the very first thing on the page is an advertisement for CompUSA. Not quite what we need, I'm afraid.
Though you do have a point: "Internet Keywords" is the technology working in IE (and probably your Mozilla, if that's how you've configured it), but these are proprietary, subscription services that operate on keywords, not necessarily real-world names. I can sign up for whatever keyword I want.
What we need is an authoritative directory, not a commercial "pay for listing" service. We need a whitepages (with business listings), not a yellow pages.
Who pays for the WHOIS databases? The registrars + you.
We're really talking about the same thing here. When an organization registers for its DNS domain, it provides (in an authenticated fashion) its real-world identity. Instead of just dumping it into a WHOIS database, we get a directory giving lookups in the other direction.
Or, if we really want to get fancy about it, ICANN would maintain a root (as they do the root DNS servers), and delegate via referrals to a company-owned X.500 or LDAP directory to provide that mapping. Sorta like how DNS delegation works.
Of course, searching such a mammoth distributed thing probably wouldn't be practical so it's likely something would need to be centralized.
Maybe local governments could manage it? They already give us the ability to look up corporate names to find their business locations, along with the ability to look up trade marks, service marks and the like, and to identify the organization that's registered those marks. These things are just in obscure proprietary databases with only an HTTP interface exposed to the public. There's no reason this information can't be exposed by way of a proper directory so that the terms can be mapped in a machine-readable fashion.
DNS is completely inappropriate for use as a "marketing space" to begin with. This is why we have all of these idiotic lawsuits (and squatting) under the existing TLDs: the domains themselves have been *given* intellectual property status when they do not deserve it. Who deserves apple.com? Apple Supermarkets or Apple Computers? Why?
New TLDs isn't the answer, it's just going to flatten the namespace and give an order of magnitude more traffic to the root servers. Who's going to pay for that? You want to charge new TLD owners $500 a year to register? Who's going to manage that namespace? Is ICANN going to become a registrar, or are we going to start having independent registrars managing the root namespace? Nothing about this looks like a good idea. It might be technically feasible, but it's stretching DNS further than it was intended to go.
A proper solution needs to involve a *proper* directory service. DNS is not a search engine. I shouldn't have to know or guess that apple.com is Apple Computers. Today's search engines search on content and only the quality of their algorithm, the user's ability to research and a bit of luck allow it to point you to authoritative places.
It seems like an X.500 or LDAP directory service does exactly what you'd need here (and conveniently integrates with X.509-style SSL certificates), but it isn't the only solution either. Give users the ability to do a real-world name lookup through a proper directory service, and DNS domains lose their IP value entirely and can end up doing what they were originally intended to do: provide a hierarchial namespace for hosts and services. SSL can be used to start validating this real-world identity instead of just connecting the session with a DNS hostname (which is also part of the problem).
I could query this new directory for "Apple", get back a few matches including the obvious one I wanted, Apple Computers, get a mapping to their DNS domain apple.com, do an SRV lookup against apple.com for an HTTP service, and boom, I have Apple Computer's home page. I don't have to guess the DNS domain and my browser doesn't need to correct my invalid URL.
I'm inclined to agree with this. The other posters do have valid points, in that if you walk up to a psychologist and tell him that this person has Asperger's syndrome, he'll probably think the kid is way worse off than he really is, but that's the point of adjectives like "mild". Some syndromes are really just regular personality traits taken to an (unhealthy?) extreme. If someone has that trait, what else are you going to call it?
I have "mild" ADD. It's not serious enough to regularly affect my daily work, but it's there. Most people don't notice. But it is a trait of mine, and that's what it's called.
You're looking at decades when I'm talking about centuries. One might even suggest that the rising prices in Mexico are an example of this "evening out" trend in action.
As it becomes cheaper to move things across borders (cars, knowledge), borders will cease to be an obstacle to trade (e.g. in cars and knowledge). This trend is only going to continue, it's not going to reverse. Take it to its logical conclusion and you find a world where all things that don't require proximity are fair game for those that aren't proximate.
If we farm all of our jobs out to India, who will be left to buy anything?
Indians, of course.
Globalization will balance everything out in the long run, but the first few hundred years are going to piss a lot of people off.
The USA is increasingly catering to companies and those that own them, at the expense of the individual. Taken to an extreme perspective, the USA might be seen as a land of corporations surrounded by a sea of poverty, an extreme polarization of wealth.
Fortunately there are a few things that can't really be moved overseas (today, at least). Things like person-to-person service, sales, government, construction. Well, and lawyers. And crime. As other jobs dry up and move to less wealthy nations, these industries will probably boom. But in the end salaries will balance out just about everywhere. The only way you might outperform local salary averages is if your position requires physical proximity, and many don't, nowadays.
Some mailer configurations allow you to specify a "mailbox" in the e-mail address, for example username+mailbox@example.com. This is delivered to username@example.com. Your mail client could then be configured to examine the recipient address and do something meaningful with it.
There are actually a number of freedom of speech issues here. A Slashdot-style moderation system is meant to moderate based upon the wills of the majority. Odds are, a comment will have a score that reflects the majority's view. A minority of moderators may vote something +1, but if they're in the minority, the comment can gravitate towards -1.
There's no system to protect the viewpoint of the minority, which makes it unsuitable as a true representation of public views and opinions. All you hear about are the things the majority wants you to hear about. And for Slashdot, that's probably OK, since I think the moderation system was initially intended mainly to separate the annoying posts from the non-annoying ones (non-annoying != interesting). And in fact, you rarely see meaningful comments with a low score on Slashdot. But this isn't so much a success of the technique, but the ability of the participants to do a good job in the long run. The thing is, with a government site, you're going to get a lot of people coming out of the woodwork not to help make the site work, but to suppress the opinions of those that disagree. If that's all people are doing, Slashdot-style moderation may not work as you expect.
You'd almost need two "branches" of moderation. The first would toss out things that are inappropriate, off-topic or irrelevant. The second would accept appeals of those decisions and would be independent in all respects from the first branch. Finding people to do both jobs would be another matter.
On top of that you could still apply a Slashdot-style moderation system if you wanted, only without negative moderation points. All of the truly inappropriate stuff was eliminated already. All you're doing now is picking out those comments that you think are a good read. People could either choose to see those comments first (Slashdot-style), or perhaps see those comments marked up by other moderators that have marked up things you've also marked up (a type of predictive selection; you see the comments that your "crowd" tends to find interesting).
While I don't disagree with you, suggesting that UHF was chosen because it's "old" and proven isn't really accurate. There are only so many ways you can communicate with a lander on another planet (that we know of), and things like whiz-bang 802.11 aren't really appropriate.
Incidentally, the real "technology" decisions here would probably revolve around the data protocols themselves, not necessarily the choice of the radio band. Lots of new technologies use the same radio bands we've used for years. Higher isn't always better.
While true, this is also out of the scope of the design of the markup language. It's an issue you'd need to take up with the HTML editor developer(s) and/or the users (in the form of education). Requiring an 'alt' attribute is an attempt through standards to force people to make things meaningful and explicit. That's pretty much all you can do (from a standards-setting perspective).
(I'm not the original poster, but I figured I could chime in meaningfully.)
From a strictly theoretical perspective, if I were inventing a new language called "HTML" and could be reasonably sure implementers would do it properly, I might personally accept all three, making #2 and #3 equivalent.
The problem is that 99% of the HTML on the Internet is bad, and non-graphical user agents are widely deployed and know that the markup is bad. As a result, they are obligated to not consider #2 and #3 as equivalent. Images without an 'alt' attribute may or may not be significant. It's ambiguous and assuming the image isn't significant will be a bad assumption for the bulk of the HTML content out there.
The solution was to require the 'alt' attribute. Tools that validate HTML, as a result, will fail a document if there's an 'img' element without an 'alt' attribute. This gives HTML authors an opportunity to stop and think about what's going on and to make a meaningful decision about what to put into it.
So in theory, #2 and #3 ought to be treated similarly. That's simply not practical, however, and the newer standards were written with that in mind.
If someone is going through the trouble of trying to format a page with images to get the effect of a
tag, they should probably just use a
tag? If the author needs special consideration for how it's "rendered" (either visually or audibly), they should tweak the presentation with style sheets.
The point is, every bit of content should be meaningful. ASCII art isn't really appropriate anywhere (IMO).
They don't necessarily have images from all filters. In many pictures it's more valuable from a scientific point of view to use the infrared filter instead of the red filter. They may only command the rover to take those three pictures.
In most cases, the infrared filter is close enough to red that a composite still gives you a good image. They do occasionally take a picture with the red filter instead of the infrared, as the article states, but these aren't as useful for scientific purposes.
If the public's interest can be satisfied with a composite using the IR channel, and you get a lot more science done with it, doesn't it make sense to use it? Their mistake was in releasing color photographs without noting that the color might not be right.
Slashdot Mirror
The Internet is not e-mail! It is completely inappropriate to base the DNS name of your organization on what is effectively a content label specific to one particular service. This is the same reason .kids and .xxx are bad.
Heck, let's say I run a porn service, and want to take advantage of this mail feature. I now have to use two different DNS domains? That's stupid.
Just as PICS can give you digitally-signed content ratings for the web, some other service can give you digitally-signed ratings/labels for e-mail. Extend SMTP to, perhaps, operate over TLS or SSL, or at least perform some sort of mutual check that both sides have a SpamHaus certificate that says they're not a spammer, and you can possibly "secure" the connection.
Or just digitally sign your e-mail messages and only accept digitally-signed e-mail. Tweak your trust relationships (for PGP-style signatures) or drop your trust from any roots that are seen to sponsor spammers, and you're all set.
I can't tell if this is a troll or not.
.xxx or .sex just because one server somewhere in their organization happens to provide an HTTP service that contains content that happens to satisfy someone's definition of porn?
DNS is not a content label. The Internet is not the web. Why in the world should an organization on the Internet be required to parent themselves within DNS under
This extends to Internet providers. Should "Example Internet" (example.net) need to relocate under "example-internet.xxx" because one of their customers happens to have a nudie pic on his web site (http://users.example.net/~joe-user)?
Use PICS labels for labelling content. Use DNS domains to provide a logical, hierarchical name for your organization and network resources within your organization.
There's no law anywhere that says I have to have a web site on my DNS domain (WIPO/ICANN judgements not withstanding), so why apply HTTP content labels to your DNS domain?
Or, for that matter:
slashdot.org
computers.geek.opennic.org
Then we don't even have to create new TLDs.
Why invent a new syntax at all?
e k.opennic/
n nic.org/
http://slashdot.org.icann/
http://computers.ge
or
http://slashdot.org/
http://computers.geek.ope
Even so, I could have a fairly extensive, mature set of non-HTTP services running under my DNS domain and one "nothing here yet" document served over HTTP. What does it matter?
Making DNS flat isn't the answer. You're pushing requests and resources towards the root of the DNS tree, at a tremendous expense of the root name server operators. You can't just delegate a piece of the root to the companies that own that TLD. Flattening the namespace means flattening the control structure of the DNS as a whole; moving the distributed infrastructure to one freakin' mammoth root infrastructure.
This is not what DNS is suited for. It will not make this change gracefully.
Ten or fifty new TLDs won't help things: Microsoft will buy a domain under every one, and people will begin to wonder why their "Internet name" starts with their own name and ends with this arbitrary, cryptic label.
DNS simply is not suited for use as a Intellectual Property label, which is precisely why so much litigation goes on.
However, if Ford didn't have a website, and I registered Ford.com, and just had a parked page there, it would be pretty clear that I was doing it in the hope of extorting money from Ford.
It's not that easy, though. The Internet is not the web!
I could just as easily registered ford.com for my organization but not provided any HTTP services for that domain. Maybe I wanted it for e-mail only, or for some other services like VoIP. There's no law that says I have to have a web page at www.ford.com.
Unfortunately entities like WIPO seem to agree with your perspective, though. Internet domains are for web pages, and if you don't have a really good web page at your Internet domain, then you must be a cybersquatter, and you should be forced to give the other guy your domain so that he can put a nice web page there. It disgusts me that Corporate and Legal America not only doesn't get this, but that they are able to receive judgements based on this distortion.
if my last name was Ford, and I registered ford.com, and on it I had a huge website, with pages for every member of my family, and a history of the family, and all my relatives had @ford.com e-mail addresses, and then Ford suddenly decides they want to cash in on this e-commerce thing and sues me, well that's a sticky situation.
Exactly: in order to have a good claim on a DNS domain, you've gotta have a fancy web page there.
I might suggest that one other thing would have prevented this problem far earlier than your two important things: the creation of a directory that proposed to map legally-protected real-world names to Internet entities (DNS domains, etc.). DNS is far too simplistic to be warped into a label deserving of Intellectual Property protection. Another layer is needed to translate the world of Marketing and Legal to the world of IT.
Most of the companies providing Internet service are in it to make money. They are in turn driven by their shareholders to make even more money. Their marketing department and executives believe that they have some ideas and the weight to make it work and tell their IT department to figure out how to do it.
IT people work for companies. Companies work for their shareholders. They are not obligated to follow RFCs or to respect anyone else in the process (at least to the point where their Internet peers stop talking to them).
Maybe what we need is an organization like the IETF to provide annual "certifications of compliance" so that companies can get a digital certificate saying that they are adhering to RFCs and aren't acting in a fashion that's destructive to the Internet. Border routers and gateways between Internet peers would be required to have a valid certificate for the entity on each interface before it would route traffic coming from that interface.
But then, the companies making routers are in it to make money as well...
The Internet will adapt to market forces. It's going to do that badly, through resistence, or peacefully, by coming up with a proper solution. So far we've all been letting megacorps twist and litigate DNS into a bloody pulp, and the problem will persist until we stop crying and give the megacorps a better solution to the problem of a human-friendly name/servicemark/trademark lookup service.
This is a very good point that can't be overstated.
The definition of what a "generic" name is, or the geographic or marketing reach of a real-world name like Apple is a fuzzy concept.
Internet naming is a concrete concept, and a given DNS name needs to be persistent. We can't have people hitting the "Apple" TLD (or apple.com) in one region and hitting their local Apple Supermarket but in another region hit Apple Computers. A single DNS domain needs to always take you to the same logical entity.
So how do you map a fuzzy concept like that onto a concrete one? Another layer. An authoritative search engine, to be precise. You tell it your search terms "Apple", maybe provide it some hints about your location or the type of Apple you want to see, and it returns the DNS domain of the entity you're trying to reach. This is a very simple application of X.500 or LDAP. We just need someone (ICANN?) to step up and provide this service.
Heck, the registrars already have 90% of this information in the form of WHOIS records.
DNS isn't a search engine!
DNS is a hierarchy. It does lookups with a given key (FQDN), not searches. Bind (and other DNS servers) performs its function efficiently and effectively. There's no problem with DNS.
The problem is that the Marketing Departments want to pretend that DNS is "Internet Keywords", and because there's no alternative standard in place for doing the types of name-to-web-site lookups that they want, they are successfully lobbying to warp DNS into Internet Keywords.
Now, I do think it's worth investigating an alternative or supplement to DNS, but even then it doesn't need to be relational. X.500 or LDAP would do the job very nicely. We need an authoritative lookup service to map these real-world names that Marketing and Legal want to use/defend to a DNS domain (or directly to an IP address and service if you want to replace DNS entirely).
But don't try and graft this onto DNS, as it's completely inappropriate. DNS is a naming service, not a locator or search service.
The problem with this is that most businesses, by virtue of putting a shopping cart on a web site, automatically go global. They may be based in some city somewhere, but through their web site, they're doing business everywhere in the world.
.com more value than the others.
.us has become what corporate America wants: a flat namespace. It used to be geographic (and I still have an old geographically-based DNS domain within it), but now anyone can get example.us with no problem. So it's as useless as the gTLDs.
The reality is that the gTLDs have no semantic value anymore. They might as well be "balloon", "cat" and "fast". Only the whole "dot com" phrase gives
Even
DNS needs to either be replaced, or supplemented with something that's more appropriate to be a legally-enforced (intellectual property-friendly) label.
Also, what you're observing here is a web-only service. The Internet is not the web! We don't need a mapping between keyword and URL, we need a mapping between a real-world legal name and an Internet *domain*. Nothing about this is application-specific, unlike Internet Keywords. Other lookup services (like DNS SRV) would then map the DNS domain to a service (like a web site).
When I type it in Mozilla, I get an error saying the URL is invalid. If I search via Google, the very first thing on the page is an advertisement for CompUSA. Not quite what we need, I'm afraid.
Though you do have a point: "Internet Keywords" is the technology working in IE (and probably your Mozilla, if that's how you've configured it), but these are proprietary, subscription services that operate on keywords, not necessarily real-world names. I can sign up for whatever keyword I want.
What we need is an authoritative directory, not a commercial "pay for listing" service. We need a whitepages (with business listings), not a yellow pages.
Who pays for the WHOIS databases? The registrars + you.
We're really talking about the same thing here. When an organization registers for its DNS domain, it provides (in an authenticated fashion) its real-world identity. Instead of just dumping it into a WHOIS database, we get a directory giving lookups in the other direction.
Or, if we really want to get fancy about it, ICANN would maintain a root (as they do the root DNS servers), and delegate via referrals to a company-owned X.500 or LDAP directory to provide that mapping. Sorta like how DNS delegation works.
Of course, searching such a mammoth distributed thing probably wouldn't be practical so it's likely something would need to be centralized.
Maybe local governments could manage it? They already give us the ability to look up corporate names to find their business locations, along with the ability to look up trade marks, service marks and the like, and to identify the organization that's registered those marks. These things are just in obscure proprietary databases with only an HTTP interface exposed to the public. There's no reason this information can't be exposed by way of a proper directory so that the terms can be mapped in a machine-readable fashion.
DNS is completely inappropriate for use as a "marketing space" to begin with. This is why we have all of these idiotic lawsuits (and squatting) under the existing TLDs: the domains themselves have been *given* intellectual property status when they do not deserve it. Who deserves apple.com? Apple Supermarkets or Apple Computers? Why?
New TLDs isn't the answer, it's just going to flatten the namespace and give an order of magnitude more traffic to the root servers. Who's going to pay for that? You want to charge new TLD owners $500 a year to register? Who's going to manage that namespace? Is ICANN going to become a registrar, or are we going to start having independent registrars managing the root namespace? Nothing about this looks like a good idea. It might be technically feasible, but it's stretching DNS further than it was intended to go.
A proper solution needs to involve a *proper* directory service. DNS is not a search engine. I shouldn't have to know or guess that apple.com is Apple Computers. Today's search engines search on content and only the quality of their algorithm, the user's ability to research and a bit of luck allow it to point you to authoritative places.
It seems like an X.500 or LDAP directory service does exactly what you'd need here (and conveniently integrates with X.509-style SSL certificates), but it isn't the only solution either. Give users the ability to do a real-world name lookup through a proper directory service, and DNS domains lose their IP value entirely and can end up doing what they were originally intended to do: provide a hierarchial namespace for hosts and services. SSL can be used to start validating this real-world identity instead of just connecting the session with a DNS hostname (which is also part of the problem).
I could query this new directory for "Apple", get back a few matches including the obvious one I wanted, Apple Computers, get a mapping to their DNS domain apple.com, do an SRV lookup against apple.com for an HTTP service, and boom, I have Apple Computer's home page. I don't have to guess the DNS domain and my browser doesn't need to correct my invalid URL.
I'm inclined to agree with this. The other posters do have valid points, in that if you walk up to a psychologist and tell him that this person has Asperger's syndrome, he'll probably think the kid is way worse off than he really is, but that's the point of adjectives like "mild". Some syndromes are really just regular personality traits taken to an (unhealthy?) extreme. If someone has that trait, what else are you going to call it?
I have "mild" ADD. It's not serious enough to regularly affect my daily work, but it's there. Most people don't notice. But it is a trait of mine, and that's what it's called.
You're looking at decades when I'm talking about centuries. One might even suggest that the rising prices in Mexico are an example of this "evening out" trend in action.
As it becomes cheaper to move things across borders (cars, knowledge), borders will cease to be an obstacle to trade (e.g. in cars and knowledge). This trend is only going to continue, it's not going to reverse. Take it to its logical conclusion and you find a world where all things that don't require proximity are fair game for those that aren't proximate.
If we farm all of our jobs out to India, who will be left to buy anything?
Indians, of course.
Globalization will balance everything out in the long run, but the first few hundred years are going to piss a lot of people off.
The USA is increasingly catering to companies and those that own them, at the expense of the individual. Taken to an extreme perspective, the USA might be seen as a land of corporations surrounded by a sea of poverty, an extreme polarization of wealth.
Fortunately there are a few things that can't really be moved overseas (today, at least). Things like person-to-person service, sales, government, construction. Well, and lawyers. And crime. As other jobs dry up and move to less wealthy nations, these industries will probably boom. But in the end salaries will balance out just about everywhere. The only way you might outperform local salary averages is if your position requires physical proximity, and many don't, nowadays.
What can you do? Buy some stock.
Some mailer configurations allow you to specify a "mailbox" in the e-mail address, for example username+mailbox@example.com. This is delivered to username@example.com. Your mail client could then be configured to examine the recipient address and do something meaningful with it.
There are actually a number of freedom of speech issues here. A Slashdot-style moderation system is meant to moderate based upon the wills of the majority. Odds are, a comment will have a score that reflects the majority's view. A minority of moderators may vote something +1, but if they're in the minority, the comment can gravitate towards -1.
There's no system to protect the viewpoint of the minority, which makes it unsuitable as a true representation of public views and opinions. All you hear about are the things the majority wants you to hear about. And for Slashdot, that's probably OK, since I think the moderation system was initially intended mainly to separate the annoying posts from the non-annoying ones (non-annoying != interesting). And in fact, you rarely see meaningful comments with a low score on Slashdot. But this isn't so much a success of the technique, but the ability of the participants to do a good job in the long run. The thing is, with a government site, you're going to get a lot of people coming out of the woodwork not to help make the site work, but to suppress the opinions of those that disagree. If that's all people are doing, Slashdot-style moderation may not work as you expect.
You'd almost need two "branches" of moderation. The first would toss out things that are inappropriate, off-topic or irrelevant. The second would accept appeals of those decisions and would be independent in all respects from the first branch. Finding people to do both jobs would be another matter.
On top of that you could still apply a Slashdot-style moderation system if you wanted, only without negative moderation points. All of the truly inappropriate stuff was eliminated already. All you're doing now is picking out those comments that you think are a good read. People could either choose to see those comments first (Slashdot-style), or perhaps see those comments marked up by other moderators that have marked up things you've also marked up (a type of predictive selection; you see the comments that your "crowd" tends to find interesting).
While I don't disagree with you, suggesting that UHF was chosen because it's "old" and proven isn't really accurate. There are only so many ways you can communicate with a lander on another planet (that we know of), and things like whiz-bang 802.11 aren't really appropriate.
Incidentally, the real "technology" decisions here would probably revolve around the data protocols themselves, not necessarily the choice of the radio band. Lots of new technologies use the same radio bands we've used for years. Higher isn't always better.
While true, this is also out of the scope of the design of the markup language. It's an issue you'd need to take up with the HTML editor developer(s) and/or the users (in the form of education). Requiring an 'alt' attribute is an attempt through standards to force people to make things meaningful and explicit. That's pretty much all you can do (from a standards-setting perspective).
(I'm not the original poster, but I figured I could chime in meaningfully.)
From a strictly theoretical perspective, if I were inventing a new language called "HTML" and could be reasonably sure implementers would do it properly, I might personally accept all three, making #2 and #3 equivalent.
The problem is that 99% of the HTML on the Internet is bad, and non-graphical user agents are widely deployed and know that the markup is bad. As a result, they are obligated to not consider #2 and #3 as equivalent. Images without an 'alt' attribute may or may not be significant. It's ambiguous and assuming the image isn't significant will be a bad assumption for the bulk of the HTML content out there.
The solution was to require the 'alt' attribute. Tools that validate HTML, as a result, will fail a document if there's an 'img' element without an 'alt' attribute. This gives HTML authors an opportunity to stop and think about what's going on and to make a meaningful decision about what to put into it.
So in theory, #2 and #3 ought to be treated similarly. That's simply not practical, however, and the newer standards were written with that in mind.
tag, they should probably just use a
tag? If the author needs special consideration for how it's "rendered" (either visually or audibly), they should tweak the presentation with style sheets.
The point is, every bit of content should be meaningful. ASCII art isn't really appropriate anywhere (IMO).
They don't necessarily have images from all filters. In many pictures it's more valuable from a scientific point of view to use the infrared filter instead of the red filter. They may only command the rover to take those three pictures.
In most cases, the infrared filter is close enough to red that a composite still gives you a good image. They do occasionally take a picture with the red filter instead of the infrared, as the article states, but these aren't as useful for scientific purposes.
If the public's interest can be satisfied with a composite using the IR channel, and you get a lot more science done with it, doesn't it make sense to use it? Their mistake was in releasing color photographs without noting that the color might not be right.
Incidentally, all of the raw images are available on the NASA web site. Instructions for a do-it-yourself composite are available from the previous Slashdot article discussing the color of the images on Mars.