There's a problem with preselling DVD sets and fans paying for a season.
Unless massive numbers of people are writing Paramount checks and trusting that Paramount, in their kindness, will do "the right thing" with the money, what's really happening is that people are buying some sort of security -- a share, a bond, etc.
The SEC takes a dim view of such movie-related securities. And trying to do things "right" takes a lot of money to set up properly.
Not only does this prevent Firefox from freezing up obnoxiously, but it also means that you don't see the file until it's actually done loading. Progressive PDF's suck.
Well, the thing is, they really need to re-interperet the franchise for a new generation. Remember, TNG and TOS were completely different serieses for completely different generations.
See, the Russians are perfectly capable of making safe, well-engineered stuff. It's just that we don't always recognize it.
A F-16 has a jet intake under the cockpit. Thus, it's awfully easy for it to suck up any debris on the ground while taxing or taking off. Therefore, debris control is important. They need to scout the airport every morning. Our jets need a whole mobile maintenence facility to keep them flying.
A Mig-29? It's got a screen that deploys in front of the engines and auxiliary upward-facing intakes. So they don't need to wory about operating from poorly-prepared fields. They make it such that everything needed right now for an aircraft fits on a single truck. If it's more important than that, you make sure it won't need to be replaced in the middle of your campaign.
The Soyuz has primitive components, yes. But they've got stuff that won't stop working. Like a primitive optical periscope that gives you enough margin to do a re-entry without guidance. They make sure that the systems that are important just won't fail.
I think it was also there because the Russians didn't want to let the US be able to do stuff that they weren't able to.
Like space war. The shuttle would have been able to fly up, retrieve a Russian spysat and then land again, after 1 orbit.
But, after the US wasn't plotting to do such things to Russia, they didn't necessarily need to do those sort of things to us. Therefore, because Soyuz still worked just fine, there was no point in having a shuttle-equivelent because it would be even less useful than the US's shuttle.
*one* RDF-based technology (And only for part of the market) in the past several years means that we'll start having real semantic applications in 2105 or so.
Remember, there's nothing intelligent about syndication. It fits just as well into the "well formed web" as it does the "semantic web". All RDF does is make things much more verbose than it is otherwise. The whole point of the semantic web was so that I could view a RDF Site Summery file and have my web browser automatically figure it out and link to other places with it.
No, I don't have any facts to back up that your average web designer's eyes will glaze over. But then, your average web designer isn't thinking about RDF, so the message is either getting lost or ignored somewhere.
My theses are: 1) RDF, without a dramatic reduction of complexity, coupled with real applications that can only be done with RDF, not merely well-formed XML, is doomed for failure. 2) Trustworth metainformation is valuable. However, it's also extremely rare. Many of the "cool things" that the semantic web was supposed to enable require trusted metainformation. 3) Until there are real applications for Semantic Web technologies, people won't Semantic Web enable their software in any substantial way. Therefore, if you want the Semantic Web to happen you either need to find a way to make existing metadata RDF-accessible or you need to make one with existing technology.
Right, but the history of "let's do better than a standard rocket by.... because we've got $x billion" hasn't been so good.
Case in point, space shuttle.
The big thing to remember is that the Falcon boosters should be signifigantly cheaper than the current crop of launchers and at least partially reusable. So, even though it's not revolutionary, there's much jumpstarting of the launch biz with what he's got.
The problem is that most of the time, you don't need a revolution, just a little evolution.
The problem with RDF is that they are putting really technical terms on really simple ideas and nobody's done an especially good job of distilling it down to the most basic level in such a way that anybody can program it. It's not just that the format's ugly, it's that, as far as I can tell, the vast majority of folks who actually are in a position to output semantic information suffer eye-glaze-over when they try to understand RDF.
Furthermore, the problem is not just that you need to have the tools output the semantics, but you have to get people to put them in the document. When's the last time you've seen somebody fill in the various meta-information fields on a word document? So, sure Dreamweaver could put Dublin Core metadata there, but people won't use it. In fact, they'll probably complain if it's there because it'll either bloat the size of their documents or include information that they'd rather not have included -- like EXIF thumbnails and revision notes and things like that. Dublin Core won't get you anywhere on any of the search engines because the meta tags got too much search-engine-spam crap stuffed into them. Likewise, if there was an image gallery that gave an RDF representation of the EXIF tags, it would slow down load times, so unless there was a good advantage to using it that way, people would complain that crap they don't want is in the document. Remember, people are *surprised* that folks who use flickr and del.icio.us are actually adding metainformation of any sort to their personal stuff.
I don't implictly trust EBay or Amazon.com. However, I do trust that there's at least some modicum of crap-removal at play. Remember, the problem with the web isn't just finding stuff. It's finding stuff without finding crap.
The problem is that there's a log of different tools for building sites. There's quite a few weblogging systems (Slashcode, Drupal, Radio UserLand, Movable Type, etc) out there. There's tons of shopping carts. There's a variety of commercial CMS systems. There's frontpage and dreamweaver. There's stuff that nobody in their right mind would use anymore, except that they are too lazy to use something else. There's hand-coded stuff. Simply put, there's a lot of stuff generating pages. Most of the time, they can't even be bothered to make their pages work on Mozilla or be compliant to any sort of standard. About the only thing that can be said about the task of changing how most of the authoring and content management software in use outputs stuff is that it's easier than moving to IPv6.
No, folks have been moaning for several years now that we need to RDF enable stuff. It's not going to happen. The only road forward is to accept that and figure out how to have a semantic web without requiring huge buy-in before you've got useful apps.
And if there start to be some actual applications of semantic web technology, other than tools for generating the data and viewers that claim to be revolutionary but always seem to just display things as a directed graph, then people will start thinking about outputting RDF.
I think part of why TNG seasons 3 till the end were actually good is because Berman wasn't completely running the show.
In that, it was origionally Gene Roddenberry's show with a lot of vision harkening back to the sixties. So you had a mix between the sixties hippie utopianism with some attempts to put the grittieness of later Trek in, too. He couldn't do a 180, but he could add just the right amount of "the future will suck, too" to balance out the "problems of today will be better in the future" that gets too obnoxious if left unchecked.
The problem is, that doesn't require the semantic web or any sort of semantic technologies.
A simple well-formed XML document will suffice and be simpler to write. And if you *really* want to make it fit into the semantic web, you can provide an XSLT file that translates it to RDF. The problem is that RDF is wounded by having an incredibly ugly syntax.
Furthermore, the simple model of posting XML or RDF documents and google not only magically finding them but "putting the data back in your hands" is flawed. Part of the reason why eBay, Craigslist, etc. all exist is because there is just enough checks and balances in place that makes them not useless. Craigslist has distributed flagging and requires email addresses, eBay has accounts and a reputation system. Sure you can create reputation systems, but in order to make that work, you need to be able to pick one easily and preferably automatically. So you need.. ehrm.. a reputation system for reputation systems...
The problem is, not only do you need to be able to do an "elevator pitch" 2 sentences of a useful app, you also need to make it not fall over due to abuse, scaling laws, etc.
I think the biggest problem is that you can't trust metadata blindly. And most of the big "semantic web" stuff assumes that you can, or that figuring out trust can be "solved".
And there is metadata that's trusted. EXIF tags are trusted, simply because there's no benefit in lying. RSS is provisionally trusted simply because the user picks if they want to syndicate or not -- again, you are doing it to draw people back and lying doesn't help.
The problem is the entire semantic web movement is doing damage to its possibilities for the future. Because RDF and OWL and RDF Schema and such are all such thick and hard-to-grasp items and also because it's a buzzworded hyped technology, stuff that could be written to enable semantic technologies isn't.
See, I have my doubts that you can really write applications that, through the magic of ontologies and other stuff, can make intelligent relations between data in such a way that the total time spent in making an ontology, a schema, and whatever else necessary to make useful things happen is less than it would have taken to write a simple application that does what you want it to.
But, really, if it can be made to work, it needs to scale. There needs to be data. Sure FOAF and RSS 1 use RDF. But OPML, RSS 2, and Atom don't.
The problem is, the Semantic web folk need to quit complaining about there not being any semantic-accessible data or applications and start figuring out how to get around that.
My current thesis is that if you really want to turn the Internet into a tuple space, the fact that there isn't much current data in RDF isn't insurmountable, it's just nobody's willing to accept that you can't just force people to do everything around RDF at this point in the game.
Longwinded rant, but if you go back up to the top of the post, there's the rest of my argument. If you don't have RDF-format data, can't send out jack-booted thugs to force people to make RDF-format data, and need some of it to make things work properly, you need to figure out how to generate it.
Which there's plenty of stuff to draw from. Convert Atom, RSS 2, OPML, and other data to RDF representations and you've increased your space. You can generate more from webpages by parsing HTML for meta tags, links, and stuff. You can query google for "What's Related" and such.
The problem is, nobody's bothered to work on a tool to make tuple-spidering code to generate tuples for RDF. They aren't even trying to come up with halfway-point guidlines about making existing and new documents more able to eventually be converted to RDF.
Instead, semantic web folks are just complaining about how there's nothing to work with.
No, I think the problem with biosphere 2 is that they tried to do *everything* in one fell swoop, using no mechanical systems, etc. And then tried to cover things up when they cheated on their own challenge.
The problem is that it's about as useful as building the world's largest coffeemaker.
And we don't need it for rationally-long space missions, either. The biggest problems with traveling to Mars are problems that need to be investigated in space -- the effects of zero gravity, how to shield humans from radiation, how to not be caught in martian orbit with the critical systems broken and no spares left, etc. By comparison, life support is easy -- pack more than enough preserved food and water (Which isn't actually that heavy by comparison), exhaust CO2 into space, collect water from persperation and make oxygen from that, rinse, repeat, etc.
And if you really want to be more efficent and stop throwing out the CO2 and organic waste, you can prod that over time. Keep a hydroponic garden to take in some of the CO2, organic waste, and evaporated water to provide food. Evolve that to the point where you can keep fish and guinnea pigs (which are eaten as food in Peru) to further reduce stored food requirements, etc. Evolve things, so that there's no surprises and you can measure the exact influence of a single value, not everything at the same time.
But, personally, if I'm traveling to Mars, I'd much rather have more than enough food packed and leave the farming as a source of occasional fresh veggies than rely on a relatively small hydroponic garden.
See, I think you are placing too much faith in the government. Now, I'm a political moderate, not a liberatarian who's going to tell you that all government besides the bedrock requirements is bad.
The problem is, almost inevitably, private industry can do "things" more efficently than the government. This is the same reason why a monopoly is bad -- because there's no competition, people stop improving stuff.
Thus, one of the goals of a good government is to provide structure, where necessary, to grow industries, but not to replace or unduly restrict private industry.
The thing is that the current system doesn't work all that well. It works well enough that people are not demanding its immediate replacement. ATC was born in the days before computers and the current features of the ATC system are often as useful as an apendix. Similarly, the FAA and product liability lawsuits have made airliner travel astonishingly safe but has saddled light aircraft with undue fiscal burden. The only way to buy a new light aircraft that doesn't cost as much as a house is to build it yourself under the Experimental Aircraft rules.
Incidentally, Space Ship One is build with the benefits of these experimental aircraft rules, which do tend to limit what Burt's allowed to do with it.
Why is this the case? Well, partially because the established rules for aircraft product liability are geared for airliners with seasoned pilots, and partially because the FAA requires exacting flight testing and documentation that can be viewed as excessive.
The problem comes in when an inexperienced private pilot augers in. The heirs may not accept that their cousin/child/parent/spouse/etc. is really a alchaholic who was high and not able to deal with flying. Or the lovely borderline case where a important-but-not-critical flight system fails and an impaired pilot is not able to compensate for it.
Burt Rutan got sued. Apparently, the short form is that the heirs of a woman who sat in the back of a Burt Rutan-designed, built-by-somebody-else aircraft, piloted by a very drunk pilot, that crashed through no fault of the aircraft, decided that somehow this was his fault. This is the big reason why you can't buy a kit for any of his aircraft anymore.
Similarly, there's a wide assortment of handheld aircraft instruments for aircraft, like GPSes, altitude warning devices, etc. They are all handheld for the simple reason that if they were installed in the aircraft, they'd need to dramatically increase the price. These would increase safety for light aircraft when they were functional and wouldn't hurt safety if they failed in flight.
Thusly, I've argued that both excess legal liability and excess government requirements together can impair progress.
See, the problem that I think you are confusing two different common corporate behaviours.
See, your average corporation is going to be as unfriendly as they can get away with to their customers. They will send out collection agencies after you for their screwup. They will not let you return things because they think you were trying to pull one over them.
But there's a difference between ruining somebody's life and killing them.
The problem is that as soon as there's the whif that you might be killing people, you get lawsuit after lawsuit. Think about how much lawyers are salivating over Vioxx. There are very few cases where a company has "gotten away" with murder. The only one that I've ever heard of is the Bhopal disaster.
My personal suspicion is that the people spending the money have *always* been distributed between the US and the rest of the world in roughly the same way. It's just a game of what's the best way to get your spam in people's mailboxes. It started out that the best way was open STMP relays, then it changed the chinese rackspace, now it's a constantly shifting collection of zombie machines because the chinese rackspace is too blackholable and the open relays have been closed.
It's just hard to track things back to the source. Which is half the problem of spam laws....
See, I think that there's a difference and a gift in disguise here.
The reason why we don't have a system to deflect asteroids right now is because asteroids are one of those things that "could happen" in the far off future.
It's like smoking. It's not guaranteed to kill you, and nobody drops dead after a single puff. Some smokers live really long lives. So smoking is viewed as something that's "bad for you", not an instant death sentence.
Therefore, we've got a lot of people who smoke in the world.
However, pulling out a shotgun, pointing it at your face, and pulling the trigger is unquestionably lethal.
Therefore, the only people who do that are people who really want to die.
The difference is that we don't always think about things that *might* cause harm, but we always think about things that *will* cause harm.
This is just one of the many ways that the human brain is a little screwed up about risk management. It worked when we were on the plains of Africa and needed to evade predators and manage to survive, but it doesn't necessarily hold up now.
Now, the blessing in disguize is that a quarter century is very much long enough to figure out what to do. Remember, we've got more than enough knowlege to do it -- computers to plot trajectories, a variety of tested and untested propulsion and power systems, techniques, etc. In the quarter of a century timespan, we may just need to paint one side white to provide the push. So, in some sense, it's even easier than trying to go from nothing to the moon landing.
But what we lack, like most things in space, is a feeling of urgency to really do something about it. Thus, this is a blessing in disguize. If they give it a decent possibility of really hitting Earth, 25 years in the future, we've got time to do something about it in ways that if they said that it'll probably hit tomorrow we won't.
Same way I know several folk who, when their doctors told them that they were going to be dead in x years if they didn't quit smoking, were able to go cold turkey.
No, it's more the case that the cost, primarily in weight, of adding the required fan, wiper, or other cleaning system outweighs the value added with more scientific instruments. And given that the panels are relatively fragile (remember, every gram of mass has a dollar amount attatched to it) you'd need to be awfully careful -- wipers are out.
Also, it's something else that can fail. Sure, it sounds like a good idea, but if you ruin the solar panels halfway into the base mission because it doesn't work, people start looking really dumb. Or if the shape of grains of martian soil is not quite the same as earth soil and it ends up not working. Or there's something else that might fail, you leave a backup for it out, and then look really stupid when that part fails and you've still got plenty of solar energy.
The biggest problem, of course, is that the designers of the probe are hamstrung by rather unreasonable launch costs that are showing little signs of getting better and are prevented by vast armies of rather stupid anti-nuclear-power whackos from using a 5 year power source. Oh yeah, and most of the NASA budget is reserved for a space shuttle that is far too expensive and has not been able to be retired and replaced due to a variety of issues.
But, in general, it's much better to get a different assortment of tools on a different probe in a completely different location every 2 years, with a chance to have design improvements, instead of having two massive probes that last for 5 years and can only be launched every 10 years.
Because you don't need wings starting around Mach 2-3. After that point, they become dead weight and add drag.
A lot of folks think that the mass penalty of carying extra fuel for landing is less than the mass penalty of carying wings (a penalty which includes extra fuel and engine mass to compensate for the increased drag).
If you are doing SSTO, you can have much less sophisticated heat shielding because the requirements of heat shielding decrease as you get less dense. At reentry, a SSTO is not very dense at all, so it's easier. Also, there's some arguments about reentering tail-first and using the engines to reduce the heat loading, which hasn't yet been tested.
Furthermore, range safety is simpler with VTOL. You have to assume that, at any point, your spacecraft could explode, raining parts down on populated land. Less gliding means less area to wory about. Airliners don't need to wory about such things, but airliners also have a good track record of not blowing up. Spacecraft don't have that record yet.
Ejection seats and escape capsules aren't very heavy, if they are included in the design early (They are now saying that, given that both the Challenger and Columbia's crew cabin survived the explosion intact, that they really could have made it removable for a minimum weight penalty. However, it's too late to do that now.)
The biggest problem is that NASA spent all of their time between the 1980s and today designing a bunch of different concepts for spacecraft, none of which have actually flown enough to be able to contribute factual data about all of this except for a few low-altitude hops made by the DC-X that made the VTOL model seem rather reasonable.
Very true. I made the connection that a wide-angle prime would be a good 50mm replacement, but I didn't make the connection that a cheap 50mm prime becomes a portrait lens.;)
They do make prime lenses still. Canon makes a 50mm/f1.8 for the EOS mount that's cheap, but still gets a lot of good work done. And there's a few other prime lenses throughout the range that are similarly useful, depending on what type of camera work you are doing. I looked in to it a while back before it was very clear that DSLRS were almost "good enough" so I should reconsider my purchase decisions, so I'm waiting for another year or three before making the big jump away from my trusty old manual-everything Cannon.
I wouldn't say that the worst lenses now are better than the average lenses made 20-30 years ago. Sure, we now have the ability to model the exact light paths and make the optics a little better, but the real problem lies elsewhere.
The biggest problem with a cheap zoom lens is that the range is too wide. If the image quality is only good in the middle of the aperature and zoom ranges, then why did they bother unlocking those ranges? I've got a 50mm/f1.8 lens on my manual-everything that's too cheap to actually pay for and it has good image quality, even stopped down all of the way. You might as well accept that your 28-105/f3.5 lens is really a 50-85/f4.5.
Aperature is *everything*. If you have a 50mm/f1.8 lens and you are in a club taking pictures of the band and you can't use a flash, you just walk closer to the band and/or get the whole band in one shot. If you have a 28-105mm/f3.5 lens, you resign yourself to a blurry picture because there's no way you are going to get the shot without some illumination. If you want to take a portrait, you want to have your lens stopped down as much as possible so that you have a narrow depth of field and the background is left blurry, because the background isn't important anyway.
Slashdot Mirror
That is a common theme in the BOFH chronicles.
Revolting, eh?
There's a problem with preselling DVD sets and fans paying for a season.
Unless massive numbers of people are writing Paramount checks and trusting that Paramount, in their kindness, will do "the right thing" with the money, what's really happening is that people are buying some sort of security -- a share, a bond, etc.
The SEC takes a dim view of such movie-related securities. And trying to do things "right" takes a lot of money to set up properly.
Disable the acrobat plugin.
Not only does this prevent Firefox from freezing up obnoxiously, but it also means that you don't see the file until it's actually done loading. Progressive PDF's suck.
Well, the thing is, they really need to re-interperet the franchise for a new generation. Remember, TNG and TOS were completely different serieses for completely different generations.
See, the Russians are perfectly capable of making safe, well-engineered stuff. It's just that we don't always recognize it.
A F-16 has a jet intake under the cockpit. Thus, it's awfully easy for it to suck up any debris on the ground while taxing or taking off. Therefore, debris control is important. They need to scout the airport every morning. Our jets need a whole mobile maintenence facility to keep them flying.
A Mig-29? It's got a screen that deploys in front of the engines and auxiliary upward-facing intakes. So they don't need to wory about operating from poorly-prepared fields. They make it such that everything needed right now for an aircraft fits on a single truck. If it's more important than that, you make sure it won't need to be replaced in the middle of your campaign.
The Soyuz has primitive components, yes. But they've got stuff that won't stop working. Like a primitive optical periscope that gives you enough margin to do a re-entry without guidance. They make sure that the systems that are important just won't fail.
I think it was also there because the Russians didn't want to let the US be able to do stuff that they weren't able to.
Like space war. The shuttle would have been able to fly up, retrieve a Russian spysat and then land again, after 1 orbit.
But, after the US wasn't plotting to do such things to Russia, they didn't necessarily need to do those sort of things to us. Therefore, because Soyuz still worked just fine, there was no point in having a shuttle-equivelent because it would be even less useful than the US's shuttle.
*one* RDF-based technology (And only for part of the market) in the past several years means that we'll start having real semantic applications in 2105 or so.
Remember, there's nothing intelligent about syndication. It fits just as well into the "well formed web" as it does the "semantic web". All RDF does is make things much more verbose than it is otherwise. The whole point of the semantic web was so that I could view a RDF Site Summery file and have my web browser automatically figure it out and link to other places with it.
No, I don't have any facts to back up that your average web designer's eyes will glaze over. But then, your average web designer isn't thinking about RDF, so the message is either getting lost or ignored somewhere.
My theses are:
1) RDF, without a dramatic reduction of complexity, coupled with real applications that can only be done with RDF, not merely well-formed XML, is doomed for failure.
2) Trustworth metainformation is valuable. However, it's also extremely rare. Many of the "cool things" that the semantic web was supposed to enable require trusted metainformation.
3) Until there are real applications for Semantic Web technologies, people won't Semantic Web enable their software in any substantial way. Therefore, if you want the Semantic Web to happen you either need to find a way to make existing metadata RDF-accessible or you need to make one with existing technology.
Right, but the history of "let's do better than a standard rocket by .... because we've got $x billion" hasn't been so good.
Case in point, space shuttle.
The big thing to remember is that the Falcon boosters should be signifigantly cheaper than the current crop of launchers and at least partially reusable. So, even though it's not revolutionary, there's much jumpstarting of the launch biz with what he's got.
The problem is that most of the time, you don't need a revolution, just a little evolution.
Yeah, but he fought for a black woman mostly because he was sleeping with her at the time.
The problem with RDF is that they are putting really technical terms on really simple ideas and nobody's done an especially good job of distilling it down to the most basic level in such a way that anybody can program it. It's not just that the format's ugly, it's that, as far as I can tell, the vast majority of folks who actually are in a position to output semantic information suffer eye-glaze-over when they try to understand RDF.
Furthermore, the problem is not just that you need to have the tools output the semantics, but you have to get people to put them in the document. When's the last time you've seen somebody fill in the various meta-information fields on a word document? So, sure Dreamweaver could put Dublin Core metadata there, but people won't use it. In fact, they'll probably complain if it's there because it'll either bloat the size of their documents or include information that they'd rather not have included -- like EXIF thumbnails and revision notes and things like that. Dublin Core won't get you anywhere on any of the search engines because the meta tags got too much search-engine-spam crap stuffed into them. Likewise, if there was an image gallery that gave an RDF representation of the EXIF tags, it would slow down load times, so unless there was a good advantage to using it that way, people would complain that crap they don't want is in the document. Remember, people are *surprised* that folks who use flickr and del.icio.us are actually adding metainformation of any sort to their personal stuff.
I don't implictly trust EBay or Amazon.com. However, I do trust that there's at least some modicum of crap-removal at play. Remember, the problem with the web isn't just finding stuff. It's finding stuff without finding crap.
The problem is that there's a log of different tools for building sites. There's quite a few weblogging systems (Slashcode, Drupal, Radio UserLand, Movable Type, etc) out there. There's tons of shopping carts. There's a variety of commercial CMS systems. There's frontpage and dreamweaver. There's stuff that nobody in their right mind would use anymore, except that they are too lazy to use something else. There's hand-coded stuff. Simply put, there's a lot of stuff generating pages. Most of the time, they can't even be bothered to make their pages work on Mozilla or be compliant to any sort of standard. About the only thing that can be said about the task of changing how most of the authoring and content management software in use outputs stuff is that it's easier than moving to IPv6.
No, folks have been moaning for several years now that we need to RDF enable stuff. It's not going to happen. The only road forward is to accept that and figure out how to have a semantic web without requiring huge buy-in before you've got useful apps.
And if there start to be some actual applications of semantic web technology, other than tools for generating the data and viewers that claim to be revolutionary but always seem to just display things as a directed graph, then people will start thinking about outputting RDF.
I think part of why TNG seasons 3 till the end were actually good is because Berman wasn't completely running the show.
In that, it was origionally Gene Roddenberry's show with a lot of vision harkening back to the sixties. So you had a mix between the sixties hippie utopianism with some attempts to put the grittieness of later Trek in, too. He couldn't do a 180, but he could add just the right amount of "the future will suck, too" to balance out the "problems of today will be better in the future" that gets too obnoxious if left unchecked.
RSS and Atom don't have the concept of a meta tag.
;) ). Tags are one metadata format.
Meta tags are there to hold any type of metainformation (but mostly there for people who view document source
There's, of course, nothing preventing you from adding tagging as a meta tag.
The problem is, that doesn't require the semantic web or any sort of semantic technologies.
A simple well-formed XML document will suffice and be simpler to write. And if you *really* want to make it fit into the semantic web, you can provide an XSLT file that translates it to RDF. The problem is that RDF is wounded by having an incredibly ugly syntax.
Furthermore, the simple model of posting XML or RDF documents and google not only magically finding them but "putting the data back in your hands" is flawed. Part of the reason why eBay, Craigslist, etc. all exist is because there is just enough checks and balances in place that makes them not useless. Craigslist has distributed flagging and requires email addresses, eBay has accounts and a reputation system. Sure you can create reputation systems, but in order to make that work, you need to be able to pick one easily and preferably automatically. So you need.. ehrm.. a reputation system for reputation systems...
The problem is, not only do you need to be able to do an "elevator pitch" 2 sentences of a useful app, you also need to make it not fall over due to abuse, scaling laws, etc.
I think the biggest problem is that you can't trust metadata blindly. And most of the big "semantic web" stuff assumes that you can, or that figuring out trust can be "solved".
And there is metadata that's trusted. EXIF tags are trusted, simply because there's no benefit in lying. RSS is provisionally trusted simply because the user picks if they want to syndicate or not -- again, you are doing it to draw people back and lying doesn't help.
The problem is the entire semantic web movement is doing damage to its possibilities for the future. Because RDF and OWL and RDF Schema and such are all such thick and hard-to-grasp items and also because it's a buzzworded hyped technology, stuff that could be written to enable semantic technologies isn't.
See, I have my doubts that you can really write applications that, through the magic of ontologies and other stuff, can make intelligent relations between data in such a way that the total time spent in making an ontology, a schema, and whatever else necessary to make useful things happen is less than it would have taken to write a simple application that does what you want it to.
But, really, if it can be made to work, it needs to scale. There needs to be data. Sure FOAF and RSS 1 use RDF. But OPML, RSS 2, and Atom don't.
The problem is, the Semantic web folk need to quit complaining about there not being any semantic-accessible data or applications and start figuring out how to get around that.
My current thesis is that if you really want to turn the Internet into a tuple space, the fact that there isn't much current data in RDF isn't insurmountable, it's just nobody's willing to accept that you can't just force people to do everything around RDF at this point in the game.
Longwinded rant, but if you go back up to the top of the post, there's the rest of my argument. If you don't have RDF-format data, can't send out jack-booted thugs to force people to make RDF-format data, and need some of it to make things work properly, you need to figure out how to generate it.
Which there's plenty of stuff to draw from. Convert Atom, RSS 2, OPML, and other data to RDF representations and you've increased your space. You can generate more from webpages by parsing HTML for meta tags, links, and stuff. You can query google for "What's Related" and such.
The problem is, nobody's bothered to work on a tool to make tuple-spidering code to generate tuples for RDF. They aren't even trying to come up with halfway-point guidlines about making existing and new documents more able to eventually be converted to RDF.
Instead, semantic web folks are just complaining about how there's nothing to work with.
No, I think the problem with biosphere 2 is that they tried to do *everything* in one fell swoop, using no mechanical systems, etc. And then tried to cover things up when they cheated on their own challenge.
The problem is that it's about as useful as building the world's largest coffeemaker.
And we don't need it for rationally-long space missions, either. The biggest problems with traveling to Mars are problems that need to be investigated in space -- the effects of zero gravity, how to shield humans from radiation, how to not be caught in martian orbit with the critical systems broken and no spares left, etc. By comparison, life support is easy -- pack more than enough preserved food and water (Which isn't actually that heavy by comparison), exhaust CO2 into space, collect water from persperation and make oxygen from that, rinse, repeat, etc.
And if you really want to be more efficent and stop throwing out the CO2 and organic waste, you can prod that over time. Keep a hydroponic garden to take in some of the CO2, organic waste, and evaporated water to provide food. Evolve that to the point where you can keep fish and guinnea pigs (which are eaten as food in Peru) to further reduce stored food requirements, etc. Evolve things, so that there's no surprises and you can measure the exact influence of a single value, not everything at the same time.
But, personally, if I'm traveling to Mars, I'd much rather have more than enough food packed and leave the farming as a source of occasional fresh veggies than rely on a relatively small hydroponic garden.
See, I think you are placing too much faith in the government. Now, I'm a political moderate, not a liberatarian who's going to tell you that all government besides the bedrock requirements is bad.
The problem is, almost inevitably, private industry can do "things" more efficently than the government. This is the same reason why a monopoly is bad -- because there's no competition, people stop improving stuff.
Thus, one of the goals of a good government is to provide structure, where necessary, to grow industries, but not to replace or unduly restrict private industry.
The thing is that the current system doesn't work all that well. It works well enough that people are not demanding its immediate replacement. ATC was born in the days before computers and the current features of the ATC system are often as useful as an apendix. Similarly, the FAA and product liability lawsuits have made airliner travel astonishingly safe but has saddled light aircraft with undue fiscal burden. The only way to buy a new light aircraft that doesn't cost as much as a house is to build it yourself under the Experimental Aircraft rules.
Incidentally, Space Ship One is build with the benefits of these experimental aircraft rules, which do tend to limit what Burt's allowed to do with it.
Why is this the case? Well, partially because the established rules for aircraft product liability are geared for airliners with seasoned pilots, and partially because the FAA requires exacting flight testing and documentation that can be viewed as excessive.
The problem comes in when an inexperienced private pilot augers in. The heirs may not accept that their cousin/child/parent/spouse/etc. is really a alchaholic who was high and not able to deal with flying. Or the lovely borderline case where a important-but-not-critical flight system fails and an impaired pilot is not able to compensate for it.
Burt Rutan got sued. Apparently, the short form is that the heirs of a woman who sat in the back of a Burt Rutan-designed, built-by-somebody-else aircraft, piloted by a very drunk pilot, that crashed through no fault of the aircraft, decided that somehow this was his fault. This is the big reason why you can't buy a kit for any of his aircraft anymore.
Similarly, there's a wide assortment of handheld aircraft instruments for aircraft, like GPSes, altitude warning devices, etc. They are all handheld for the simple reason that if they were installed in the aircraft, they'd need to dramatically increase the price. These would increase safety for light aircraft when they were functional and wouldn't hurt safety if they failed in flight.
Thusly, I've argued that both excess legal liability and excess government requirements together can impair progress.
See, the problem that I think you are confusing two different common corporate behaviours.
See, your average corporation is going to be as unfriendly as they can get away with to their customers. They will send out collection agencies after you for their screwup. They will not let you return things because they think you were trying to pull one over them.
But there's a difference between ruining somebody's life and killing them.
The problem is that as soon as there's the whif that you might be killing people, you get lawsuit after lawsuit. Think about how much lawyers are salivating over Vioxx. There are very few cases where a company has "gotten away" with murder. The only one that I've ever heard of is the Bhopal disaster.
My personal suspicion is that the people spending the money have *always* been distributed between the US and the rest of the world in roughly the same way. It's just a game of what's the best way to get your spam in people's mailboxes. It started out that the best way was open STMP relays, then it changed the chinese rackspace, now it's a constantly shifting collection of zombie machines because the chinese rackspace is too blackholable and the open relays have been closed.
It's just hard to track things back to the source. Which is half the problem of spam laws....
See, I think that there's a difference and a gift in disguise here.
The reason why we don't have a system to deflect asteroids right now is because asteroids are one of those things that "could happen" in the far off future.
It's like smoking. It's not guaranteed to kill you, and nobody drops dead after a single puff. Some smokers live really long lives. So smoking is viewed as something that's "bad for you", not an instant death sentence.
Therefore, we've got a lot of people who smoke in the world.
However, pulling out a shotgun, pointing it at your face, and pulling the trigger is unquestionably lethal.
Therefore, the only people who do that are people who really want to die.
The difference is that we don't always think about things that *might* cause harm, but we always think about things that *will* cause harm.
This is just one of the many ways that the human brain is a little screwed up about risk management. It worked when we were on the plains of Africa and needed to evade predators and manage to survive, but it doesn't necessarily hold up now.
Now, the blessing in disguize is that a quarter century is very much long enough to figure out what to do. Remember, we've got more than enough knowlege to do it -- computers to plot trajectories, a variety of tested and untested propulsion and power systems, techniques, etc. In the quarter of a century timespan, we may just need to paint one side white to provide the push. So, in some sense, it's even easier than trying to go from nothing to the moon landing.
But what we lack, like most things in space, is a feeling of urgency to really do something about it. Thus, this is a blessing in disguize. If they give it a decent possibility of really hitting Earth, 25 years in the future, we've got time to do something about it in ways that if they said that it'll probably hit tomorrow we won't.
Same way I know several folk who, when their doctors told them that they were going to be dead in x years if they didn't quit smoking, were able to go cold turkey.
There's no question that there's water ice on Mars, folks have mostly been making noises about liquid water.
But CO2 makes good frost, too.
No, it's more the case that the cost, primarily in weight, of adding the required fan, wiper, or other cleaning system outweighs the value added with more scientific instruments. And given that the panels are relatively fragile (remember, every gram of mass has a dollar amount attatched to it) you'd need to be awfully careful -- wipers are out.
Also, it's something else that can fail. Sure, it sounds like a good idea, but if you ruin the solar panels halfway into the base mission because it doesn't work, people start looking really dumb. Or if the shape of grains of martian soil is not quite the same as earth soil and it ends up not working. Or there's something else that might fail, you leave a backup for it out, and then look really stupid when that part fails and you've still got plenty of solar energy.
The biggest problem, of course, is that the designers of the probe are hamstrung by rather unreasonable launch costs that are showing little signs of getting better and are prevented by vast armies of rather stupid anti-nuclear-power whackos from using a 5 year power source. Oh yeah, and most of the NASA budget is reserved for a space shuttle that is far too expensive and has not been able to be retired and replaced due to a variety of issues.
But, in general, it's much better to get a different assortment of tools on a different probe in a completely different location every 2 years, with a chance to have design improvements, instead of having two massive probes that last for 5 years and can only be launched every 10 years.
No, it sounds like trusting a private corporation to get me in one piece from one place to another using aircraft.
Are you afraid of airliners, too?
Because you don't need wings starting around Mach 2-3. After that point, they become dead weight and add drag.
A lot of folks think that the mass penalty of carying extra fuel for landing is less than the mass penalty of carying wings (a penalty which includes extra fuel and engine mass to compensate for the increased drag).
If you are doing SSTO, you can have much less sophisticated heat shielding because the requirements of heat shielding decrease as you get less dense. At reentry, a SSTO is not very dense at all, so it's easier. Also, there's some arguments about reentering tail-first and using the engines to reduce the heat loading, which hasn't yet been tested.
Furthermore, range safety is simpler with VTOL. You have to assume that, at any point, your spacecraft could explode, raining parts down on populated land. Less gliding means less area to wory about. Airliners don't need to wory about such things, but airliners also have a good track record of not blowing up. Spacecraft don't have that record yet.
Ejection seats and escape capsules aren't very heavy, if they are included in the design early (They are now saying that, given that both the Challenger and Columbia's crew cabin survived the explosion intact, that they really could have made it removable for a minimum weight penalty. However, it's too late to do that now.)
The biggest problem is that NASA spent all of their time between the 1980s and today designing a bunch of different concepts for spacecraft, none of which have actually flown enough to be able to contribute factual data about all of this except for a few low-altitude hops made by the DC-X that made the VTOL model seem rather reasonable.
No, first it was the H1B visa holders, then it was outsourcing.
Very true. I made the connection that a wide-angle prime would be a good 50mm replacement, but I didn't make the connection that a cheap 50mm prime becomes a portrait lens. ;)
They do make prime lenses still. Canon makes a 50mm /f1.8 for the EOS mount that's cheap, but still gets a lot of good work done. And there's a few other prime lenses throughout the range that are similarly useful, depending on what type of camera work you are doing. I looked in to it a while back before it was very clear that DSLRS were almost "good enough" so I should reconsider my purchase decisions, so I'm waiting for another year or three before making the big jump away from my trusty old manual-everything Cannon.
I wouldn't say that the worst lenses now are better than the average lenses made 20-30 years ago. Sure, we now have the ability to model the exact light paths and make the optics a little better, but the real problem lies elsewhere.
The biggest problem with a cheap zoom lens is that the range is too wide. If the image quality is only good in the middle of the aperature and zoom ranges, then why did they bother unlocking those ranges? I've got a 50mm/f1.8 lens on my manual-everything that's too cheap to actually pay for and it has good image quality, even stopped down all of the way. You might as well accept that your 28-105/f3.5 lens is really a 50-85/f4.5.
Aperature is *everything*. If you have a 50mm/f1.8 lens and you are in a club taking pictures of the band and you can't use a flash, you just walk closer to the band and/or get the whole band in one shot. If you have a 28-105mm/f3.5 lens, you resign yourself to a blurry picture because there's no way you are going to get the shot without some illumination. If you want to take a portrait, you want to have your lens stopped down as much as possible so that you have a narrow depth of field and the background is left blurry, because the background isn't important anyway.
And, furthermore, cheap telephoto lenses can't gather enough light in low light conditions, exacerbating the problem...