Windows users make a good point that they don't want to relearn how to do what they already can.
I dunno. This isn't a good point in my opinion - it just shows they aren't computer "users" - they're Windows users (so they don't know what a computer is really capable of, only what Windows is capable of), or that they're lazy.
Here's the point, I guess. How much time a day does an average person who uses a computer spend typing? Let's say four hours, just for fun. Now let's say that that person learned how to type poorly, probably self-taught. And that person types at, say, 30 wpm.
If you asked that person "why don't you learn to type properly?" they might say "I don't want to relearn what I already know how to do" - ignoring the fact that people who type properly can type over 100 wpm, and that would save them many times the investment on time in a few days.
The real reason is that the person is either a) doubtful that the effort will produce an improvement, which is silly considering you can see there's obvious room for improvement, or b) lazy.
To me, I'm just completely confused. If I buy something, or am going to work on something, I figure out exactly what it's capable of and how to best maximize the time that I spend using it. That used to mean reading the manual (gasp!) but manuals are so pathetically bad nowadays that I usually end up just poking around on the Web trying to find other people's experiences with it. If you aren't willing to spend time maximizing your ability to use a piece of equipment, you're lazy. It's just that simple.
I hate to say it, but someone coming from Fedora is still not a user - they're a Red Hat user, and Red Hat has done quite a bit to make the system "like" Windows. If you expect Debian to act "like" Red Hat, you'll be disappointed. If you expect Debian to act "like" Windows, you'll be disappointed.
If you don't expect anything of Debian, and examine it for what it is, you'll be very impressed.
I don't want to describe all technical problems I had
No operating system works. There are technical problems with every operating system in existence. I can't use plenty of hardware on a Windows XP box. There's no chance in hell of me using my old LANding Gear network adapter on Windows XP. It's a tremendous effort to get the Acer NeWeb WarpLink wireless adapter to work, too - and both of those took only a tiny amount of work in Linux. The question is not "how easy does this appear at first glance?" but "how easy can it become?"
Debian has a huge userbase, and a huge community. Did you ask for help? Did you search for help? File a bug?
It takes exactly 45 minutes and all work perfectly.
Amazing. A former Fedora user thinks Fedora works perfectly. Never would have guessed.
Or maybe it is if you have only debian systems and if you are ready to lose your stability if you don't follow the Debian Way to configure something, I don't know.
In other words, "this operating system sucks if you don't learn how to use it!" Good call. I've said this elsewhere, and I'll say it again. A high learning curve does not make something not user-friendly. Someone who isn't willing to climb a learning curve isn't avoiding software because it's not user-friendly - they're avoiding software because they're lazy.
A system that install itself without problem and *just work* is friendly.
Nothing like this will ever exist for all users and all configurations. What you want is a system that "just works" for you. If you're someone who works "exactly" like Red Hat wants you to work, then Red Hat will probably "just work".
Note: Don't reply to this if you just want to defend Red Hat/Fedora. Fedora's a good distribution, and it aims to be a lot like Debian. What I'm trying to point out is that if you put no effort into getting a new operating system/distribution to work - and only two days for a complete newbie is not effort - you can't claim "it's not user-friendly!" The truth is, you were lazy. Everyone has to put in effort to learn an operating system - even Windows (ever wonder why they sell those "learn Windows now!" CDs?). What makes a userfriendly distribution is one that is friendly to the user - that allows the user to do whatever he wants, and helps as much as possible.)
Wow, so from what you are saying, Debian is the Alpha and Omega, the beginning and the end, of all operating systems.
Click on a parent and read its replies before you reply - I accidentally forgot to add a bit afterwards (Submit and Preview are too close!). OS X is very much the "ideal", though the lack of completely open software is a slight drawback. It's better than Debian, though for people who are more technically inclined, I think the open source portion of Debian makes it more user-friendly.
(For instance, my SiPix Blink2 that I just bought works in gphoto2 perfectly fine. However, Debian's libgphoto2 package accidentally wasn't compiled with libjpeg support, so it didn't work. Whoops. However, a bit of poking around and I figured this out - asking on the gphoto2 Sourceforge site helped! - and a simple apt-get source libgphoto2-2, apt-get builddeps libgphoto2-2, dpkg-buildpackage, and everything's good. If a system library accidentally was forgotten on Mac OS X, you'd have to wait for Apple to come out with a fix. Which would happen, sure, but not in the less than 1 day that I fixed it under Linux. And by the way - it works better under Linux than Windows. The Windows driver postprocesses the image in an absolutely horrendous way. This isn't 100% true, of course, since I think gphoto2 works under OS X, but you get the general idea)
The mantra for any operating system, in terms of usability, should be "Attempt to do what is sensible, but in the end, always let the user decide."
Disagreeing with a comment, but not giving any reasons.
Here's a hint: that's the stupidest thing a person can say. I'm glad your brilliance outshines the rest of humanity so much that any comments you have would be lost on the rest of us.
For us silly mortals, we have to rely on intelligent discussion in order to actually learn.
(p.s.: with a UID of 200586, you either a) haven't been here very long, or b) haven't bothered to contribute to discussions at all. If a), then you lied, and if b) then you've no right to talk about what's dumb and what's not, as you've been silent the entire time.)
I didn't say that Debian should steal software from OS X, just implementation ideas (i.e., make a pretty GUI for editing all of the conffiles, AND allow people to edit the conffiles as well).
If you have software that's not open, and not free, fundamentally, it's not as user-friendly as software that is open source, because then, the user can change it, and the user is the only person who knows what his or her needs are.
It's the one limitation of OS X, but, honestly, Apple spends a crapload of time with usability focus groups, and most people's needs aren't *that* different, so it's not a serious limitation of OS X. Microsoft probably does the same, but my God, they must do a terrible job, because in terms of usability, their products are so far behind it's crazy. Don't like Messenger as an AIM client - and who would? - try disabling it in Windows XP. It takes serious effort to kill the damned thing, as a ton of other programs launch it as well. Want to run a script every time a connection is detected (like updating a dDNS connection, or setting up an open port on a wireless router)? Ha! Good luck. We all know these things are a joke to do inside the OS, but to normal people, these things just "aren't possible". Windows is worse than just "not user-friendly" - by being so pervasive it makes people think that it defines what a computer can do, and therefore, Windows' limitations become a computer's limitations.
Uh, yes it is. That's why so many people I try to introduce Linux to don't want to switch. Linux is too much of a hassle to use.
Read the post again. I said
The problem is that it's so pervasive that everyone's used to believing that user-friendly = Windows-like = "everything just works"
and
No, it's not the most Windows-like system around.
People who start off using Windows learn its quirks and idiosyncracies and think of them as "normal". They're not. Linux isn't "Windows-ex-user-friendly", but I'm glad it's not, because Windows isn't userfriendly to begin with. In fact, there are quite a few different paradigms that Linux has that the Windows paradigm doesn't have that are far more user-friendly. See WindowMaker, for instance, with the NeXTSTEP interface, or Emacs with almost everything bound to keybindings, or LyX.
A high learning curve does not make something non-user-friendly, especially when there are rewards for the high learning curve. There's absolutely no doubt that there are more powerful tools under Linux - Emacs was virtually designed from the ground up to allow people to edit files as fast and as easily as possible (hence the reason that cursor editing keys are all control-combinations of home row keys). A person who chooses not to go through the "hassle" of not climbing a learning curve which has obvious benefits is not avoiding the program because it's not user-friendly - they're avoiding the program because they're lazy.
Start off with people who have never used a computer (or at least, never used Windows), and are willing to learn to use one, and they'll learn Linux rather easily. That's how most of us did.
Only tech-nerds like us think that way. That's a made-up definition of user-friendly.
Am I a user? Yes. Is an operating system that doesn't let me do what I want non-friendly to me? Yes. Then it's not user-friendly, now is it?
Last time I checked the definition of user-friendly is "friendly to the user". If an operating system doesn't let the user do what he wants, it's not being friendly, now is it?:)
why she should change from something that "already works."
She already had a Windows mindset - that is, "Microsoft is smarter than you. You only want to do what Microsoft lets you do. You do not want to do anything else. Microsoft is good to you." She's not a user - she's a Windows user.
Windows doesn't "work". No operating system works. There's at least one thing broken about every operating system/distribution in existence. The question as to whether or not it's user friendly is whether or not you can deal with the broken parts well.
Whoops. I meant to add at the end "or OS X", because both of those really are the top end of user-friendliness around, though OS X is even better. Both of them allow you to do whatever you want, AND they try to give you a hand by doing as much of it for you as they can.
OS X is really a marvel, though - it's what Debian really should strive to become.
Second, if you use debian, you have no right to bitch about Linux being difficult to use. It's not exactly the most user-friendly system around.
No, it's not the most Windows-like system around. It is the most Linux-like system around, and it's absolute craploads easier to use than Windows. Everything is documented, and everything is modifiable.
As a simple example: Windows XP doesn't handle wireless connections terribly well - if I standby my laptop with one wireless connection that uses DHCP, and then wake it up in an area where it has a different wireless connection, it doesn't release/renew the DHCP lease. I have to do it myself. This is stupid - on a Linux system, if the distribution was screwed up, I could just script it in a moment's notice.
Windows's help system is also a joke - most of the programs don't properly document what things do (the number of times I've seen "There is no help available for this option...") and so you're left hoping that things work.
Windows is by far one of the least user-friendly operating systems around. The problem is that it's so pervasive that everyone's used to believing that user-friendly = Windows-like = "everything just works". That's not true, because no operating system just works, because no operating system knows everything you could possibly do with it.
Linux forces you to learn about a problem before solving it. That actually makes it very user-friendly, because it means that users can realize that they can do more than what they originally thought they could do - meaning the OS makes them more productive.
Windows isn't user-friendly. The simplest way to illustrate that is to ask this: how much does it allow you, the user, to do, and how much does it try to do it for you? An operating system that does everything for you and allows you to do nothing isn't user-friendly, because what if you don't want to do what it wants? An operating system that allows you to do everything but does nothing for you isn't user-friendly, because, well, it's a computer. It can do things automatically. The best operating system is one that tries to do everything for you, but allows you to do everything as well, and that's Debian.
One is proton decay, which means the building blocks of any sentient computer will eventually decay on their own.
You do realize that no one has seen one proton decay. Not one, right? Proton decay assumes supersymmetry is valid, and as many physicists have noted, supersymmetry is an excellent theory, which predicts a whole host of particles - half of which have been discovered.
Proton decay isn't real - not yet. And there is no a priori reason to assume that it is. Its current lower bound is 10^33 or so years. At this point you can't simply claim that proton decay happens - in fact, you'd rather claim that it doesn't.
I was under the impression that dark matter needs fine tuning to explain Tully-Fisher
Yes and no: The typical Tully-Fisher coefficients for Sa, Sb, and Sc type galaxies are 9.95, 10.2, and 11.0 or so. These are all within 10%, and for Sa and Sb types, within 5%, of 10. Simple assumptions get you a coefficient of 10, if you assume that the mass-to-light ratio is the same for all spirals, and that the surface brightness is the same for all spirals.
The first assumption (mass-to-light ratio) is a clearly idiotic assumption. It assumes that galaxies form with same proportions of light and dark matter, which we *know* is not true for other types of galaxies (dwarf ellipticals, in particular). Aside: This is also the "nail in MOND's coffin", more or less - MOND was hoping to replace the dark matter hypothesis by saying physics works differently at large distances. The problem is that galaxies which contain the same amount of light-emitting matter and have the same spatial extent should therefore have the same rotation curves. This isn't true. You then have to add a new parameter with MOND to fit it, which is OK, sure, but now you've started to lose the elegance originally intended, and now MOND becomes a more complicated theory than the dark matter hypothesis, which just says "well, that galaxy formed around less dark matter."
Anyway, back to the subject: the point is that those two assumptions clearly are not completely true, and therefore there's plenty of room for a 10% correction due to forming biases in spiral galaxy types. If the mass-to-light ratio is a very weak function of mass (which is believable - perhaps smaller galaxies formed when the dark matter density was slightly lower, due to their late formation times), you can easily get those corrections.
MOND allows you to get that 10% correction due to the parametric fit of the rotation curve, which is essentially identical to the way that it's done in the dark matter case - the corrections are due to the variation in the rotation curve, which MOND says is due to a modified Newtonian field, and dark matter says is due to a dark matter density. It's the same reasoning - one isn't more natural than the other.
(It should also be noted that the Tully-Fisher data has a crapload of spread to it, just like all astronomical data. Each galaxy varies a fair amount.)
It can also explain phenomena which the dark matter hypothesis can't, such as the Tully-Fisher relationship observed in the surface brightness of galaxies.
The Tully-Fisher relation has been explained by dark matter for some time. You can find a brief derivation in Carroll & Ostlie p. 1002, for instance. There's no need to invoke MOND at all - it just comes from the fact that the luminosity is proportional to the maximum velocity to the 4th power, which you can get by using the expression for total mass contained within the galaxy derived from rotational velocity curves.
Now it is clear that you have no idea on the concept of interferometry either.
No, of course I do. I never said it would be easy, but it's not impossible. I'm a bit confused about the necessity of a completely stable site (the 30 meter number came from the Thirty Meter Telescope, which has planned sites for places like Hawaii - not exactly what one would consider a "perfectly stable site") - I don't think that the lunar surface would be a tremendously bad placement for it. The control you're immediately throwing out as "ridiculous" is definitely within reach nowadays. Technology's amazing.
Your argument is basically: assuming no launch costs, no personnel costs, a cost-free fully-outfitted manufacturing facility (that needs to make mirror substrates out of the lunar soil!
Yes, that would be exactly what I'm assuming. What would need to be launched that wouldn't be used for something else? Not a tremendous amount, for one. The personnel are used for other things, the manufacturing facility is used for other things, etc.
If you're trying to say "if we go to the Moon to build a telescope, it'd be insanely expensive!" I agree with you. Very much so. But that's not what I've been saying. What I've been saying is "if we go to the moon, and establish a permanent presence, building a telescope is a good idea."
I'm not sure how far in the not-to-distant future you are dreaming
I'd say 50 years is a reasonable estimate. If Bush is serious about the return-to-the-Moon thing.
Your comment about steel mills is way off too. These mills live and die on the margins.
Um. Yeah. Duh. That's because there's no money in small-scale extremely expensive totally cost-inefficient steel production, which is what I said.
They are large and power hungry because of thermodynamics.
No, they're power hungry because of thermodynamics. They're large because of scale. They would lose money trying to make steel smaller-scale than they do now - it's not that it's impossible. Not unless you're trying to say that iron and/or steel has never been made in anything except gigantic foundries. Besides, that's again assuming there's no other material on the Moon that could make a strong support besides steel. I wouldn't even think of steel, as the materials aren't easily available.
I mean, c'mon. Do you think no one's thought about "how do we manufacture things on the Moon?" Of course they have, and there's an absolute flood of information about it. Here's just a start.
Anyone can simply say "oh, we can't do it, it's too complicated, requires too much work, etc." Not really. There's certainly no reason why we can't build a telescope out of materials on the Moon. It's just hard. And requires people to think, and be creative, and come up with new solutions to things, which is usually harder to convince people to do in a culture where people are constantly happy with the status quo.
Yes, but Dell doesn't sell computers that won't work with Dell brand processors, or will only work with Dell brand video cards, and Dell brand hard drives, and Dell brand monitors, and Dell brand power supplies.
Yah, they used to sell some of those (Dell brand power supplies, mainly). And that sucked. However market pressure (and the fact that standards are cheaper!) pushed them to using standard parts.
Dell follows a standard - the PC standard. They have a BIOS, so a RAID card plugged in will boot, standard USB ports, standard PCI slots, standard memory slots. Even Dell's laptops now use standard parts: SODIMMs and MiniPCI cards. Standardization, with systems being supplied by multiple vendors, is a good thing. Dell doesn't have to sell the individual parts - not at all. But they don't prevent you from obtaining compatible parts.
Of course they don't, though. People would just go to other vendors that you can buy the other parts for at Best Buy when you need them. But if Dell was a monopoly, they could. And then you'd be screwed. And left with a lot of incompatible parts strewn around Best Buy because Dell's abusing its monopoly.
That's a good analogy for Microsoft, because that's what RealPlayer, Mozilla, WinZip, and AIM are slowly becoming. Incompatible parts, because Microsoft's abusing its monopoly.
Well, that's the reason that the LGPL was originally called the "Library GPL", now wasn't it? OK, maybe not, but the LGPL is a much better fit for libraries as it doesn't have that fuzziness, unless you really want to prevent the code's use in proprietary software (which is their choice, of course).
I guess I should've been more specific... most of the libraries I've ever seen are LGPL. Briefly looking through Freshmeat, I can barely find any that are GPL rather than LGPL'd. (The FSF wants people to release libraries under the GPL to strengthen free software over proprietary, but I don't think many people see it that way. They just want to have the software available, and want due credit.) What library were you having concerns over?
To be honest, I doubt it would be that much of a chore to ask the library's developer to license it to you under the LGPL for a specific project. Monetary incentive could come in handy here, as well. It's important to remember that the GPL doesn't forbid dual licensing - it can't, of course: the author is free to license something however he or she wants. Granted, cases where there are many authors involved could be difficult.
The idea of using in situ resources is great. But I wonder if it will really be a short term or medium term solution (on the scale of twenty years, say.) We have certainly been talking about asteroid mining &c, but I wonder if the costs for sending up infrastructure are feasable?
The benefit is that we're talking about things that can directly feed money back into the economy again. Small-scale manufacturing? There are tons of places that would want that!
How much money would it cost to build a small probe designed to go to the Moon and try to manufacture alloys (say, bricks) out of lunar regolith? If that's *all* it was designed to do, probably not much - say, maybe, 100 million. So there's your testbed, and then you build a larger scale and send it. Maybe in pieces so you don't have to worry about it surviving completely intact. Then when humans land, the infrastructure's already there.
Think about it this way: the mass of the systems used to construct, test and characterise the PSF of a mirror are huge compared to the mass of the mirror itself. You would need (I think) technology on a near magical scale to get that down.
This is what adaptive optics have really helped in - the mirror itself constructs an ideal PSF. The JWST, Hubble's successor, is not going up as a completed mirror, but segmented, that will unfold when it gets there, with adaptive optics optimizing the shape. This is exactly what you would want to do on the Moon, just probably an order of magnitude or more larger.
The technology you have will improve continuously so that the best giant mirror you can make now will suck compared to the one you can make later.
Well, this is somewhat true, and somewhat not true. Aperture wins, all the time, so long as the other effects are marginal. Adaptive optics basically can mean that your mirror can be as good as your electronics, since it needs to adapt to thermal variations, and replacing electronics is easy (well... control electronics. I wouldn't suggest replacing thousands of transducers, though this probably would be possible). The age of big, giant mirrors is pretty much over. Only a few of the huge, large scale next-generation telescopes are single-mirror types: both JWST and the thirty-meter-telescope are segmented designs. This is because electronics has sped up at a rate much faster than optics.:)
But let's see NASA get together a working probe that can build a stop sign out of in situ materials so that we can characterize exactly how cost effective your suggestion is.
Part of the problem that I have with the current pace of NASA is that it's honestly insane. Small, proof-of-concept missions are really the best way to go. Like Pathfinder, or Deep Space 1, for instance. Actually, in my opinion, I wouldn't've kept Pathfinder running considering the cost of operating it. I would've put that money towards its successor. NASA used to say "faster, cheaper, better" - but I'd say just "more often" would be better. Momentum, momentum, momentum.
I mean, people out there have to have ideas about how to manufacture things off planet. Plenty of scientists post proposals about it all the time, so the science is there. It's just the engineering that needs a kickstart, and the best way to do that is incrementally.
You clearly have never seen, or are aware of, the infrastructure needed to manufacture and test very large telescope mirrors.
Yes, I am. Very much so. If I had been talking about building a single 30 meter mirror, you'd be right. That would definitely need one heck of a testing facility and a lot of support facilities as well. However, you don't need that. Adaptive optics have gotten really good in the past 10 years, and so you don't need one gigantic contiguous piece of glass. With adaptive optics, you can use many smaller, easy to manufacture pieces of glass very easily.
(until you go and see how big they are and how much power and other resources needed to run them);
Modern steel mills are large and power-hungry because they can be. There's no driver to make them smaller, and it would definitely cost more money. If you force someone to think about how to build a small-scale steel mill, they'll come up with it.
Whoa! That one about knocked me out of my chair!
Ah, so that explains why you didn't read the rest of what I said in the sentence you quoted!
Assuming that lunar human infrastructure exists, and lunar manufacturing capability could be constructed
would be the part that you left out. The expensive part about building a gigantic telescope in space are the launch costs, and any repairs that would need to be done, and the necessity of all of the support personnel. If the launch costs are eliminated by lunar manufacturing, the repair/retrofit costs are simply absorbed into the cost of maintaining a permanent lunar presence, and the support personnel are no longer constantly required, then it would be quite cheap compared to a space-based telescope.
Even if the GPL'd code is part of a binary static lib, there's still some fuzziness on the use of a GPL'd API. Enough fuzziness that my company's management didn't want to risk using it.
Huh? I'm confused. They didn't want to risk using a GPL'd library?
Yes, you can't statically link to a GPL'd library with a non-GPL'd piece of code. That's not stupid, because no one would know that you're using GPL'd code. In essence, you're saying "Thanks for the code - now screw off!" to the people that created it. They also don't know if you changed it, improved it, etc. Essentially you're stealing their work without helping them at all. I don't consider the prevention of this a bad thing.
But why would you care about statically linking a GPL'd library? What's wrong with dynamically linking it? If you're worried about people not having the dynamic library, ship it with the product - along with the source of the dynamic library, which you have, because it's GPL'd.
Does the Lunar Surface Still Offer Value As a Site for Astronomical Observatories?, by three members of JPL, Goddard and UT, and published in Space Policy (I guess NRO wasn't taking articles then) provides the full story.
Hmm, do the authors have a bias for space telescopes, I wonder?
All of their arguments are correct - for now. The problem is in one word scale. Physically, extremely large telescopes aren't really feasible to send into space - to LEO or to the Moon, regardless.
However, oddly enough, the moon - being basically made out of the same stuff Earth is - has all the raw materials to make mirrors.
So the idea isn't to send a telescope from the Earth to the Moon.
The idea is to build it on the Moon in the first place.
This would require a lot of R&D, and a lot of work - but that's just R&D, which can be much cheaper than the launch costs of a gigantic telescope (a many-meter telescope).
The authors also then go to say "why would we want to keep a state of the art instrument around for a long time?" implying, of course, that state of the art changes so rapidly that keeping something around for a long time is meaningless.
They somewhat fail to see one thing: a gigantic mirror will always be state of the art. If the instruments need to be changed periodically (say once every 10 years) then the marginal cost of replacing instruments on a telescope on the Moon (assuming that the capability of getting to the Moon exists) is going to be infinitely less than launching a new state of the art space telescope.
Basically, the authors are assuming that
Everything must be put on the Moon
There is little/no lunar human infrastructure
While this is certainly true now (and validates the reason that their one example - LUTE - wasn't chosen, as it would have been an unmanned observatory) - it is definitely not arbitrarily true for the future. Assuming that lunar human infrastructure exists, and lunar manufacturing capability could be constructed, the cost of building a telescope on the Moon is tremendously cheap.
At least we should boost Hubble to higher orbit, so when NASA gets additional funding, it can try again to bring it down. Putting it in a museum somewhere would really be a inspiration to many children to go into science.
C'mon, that's crazy. It's like suggesting we should've brought Mir down in pieces in a shuttle or something. Hubble's an old space telescope, and we've thrown many old space telescopes away.
You could build a replica for a fraction of the cost that it would take to bring it down. That'd be good enough for inspirational purposes.
The only reason that people are averse to doing it now is because somehow "Hubble" got a lot of public support, but it really doesn't deserve it. It's just an old telescope. Sure, it does good science - but so does any instrument if there are people operating it. The point is "is it worth spending money on something when that money could be better spent on a better replacement?" and the answer to that is "no".
NASA never should have let Hubble get into the public's eye this much. The pretty pictures can come from pretty much any other telescope (Spitzer put out a few nice ones over the weekend) - there's no reason to keep fawning over Hubble.
Because it's not complete. It describes how you *can use* the HTML DLL, not all the ways to use it. In other words, it would allow you to use it, but not to replace it.
(Note that giving a reference implementation would be remarkably simple, given the breadth of experience that Microsoft has with HTML rendering.)
Mozilla for a while had the beginnings of a replacement of mshtml.dll, but there were significant amounts of "undocumented features" with regards to the DLL, and Microsoft also tended to change them, making it basically impossible.
Now, maybe I am a little old (ok real old), but isn't Microsoft just completing the migration of the Windows UI into the OS?
First off, Microsoft has allowed themed UI since Windows 95. Which means they've been perfectly happy to give people control over the user interface. Seond off, they readily ignore their own UI anyway - see Windows Media Player.
Anyway, A web browser is not UI. It's an application that displays Web pages. It doesn't display controls, or anything else - just Web pages. It just parses HTML into a bitmap buffer, and Microsoft should have no a priori reason to have a preferred HTML renderer in the operating system.
They are not slowly gaining control of the desktop, they OWN the desktop.
No. They're slowly gaining control over it. It used to be they weren't the desktop. You had other Office apps, you had other email apps, etc. Slowly but surely, all of the applications that people used got usurped into Microsoft.
A "desktop environment" is all "productivity applications" - applications that make a computer a general workplace tool. Not something like a CAD program, or a video game - those are special purpose, and not everyone would use it. Microsoft never owned that - nor should they.
And, being as old as I am I remember when you had to do that, and they didn't work together, and I don't want it anymore either.
A broken implementation doesn't imply a broken idea. The idea is correct. The problem was lack of standards.
Let me transpose your idea into a completely similar idea...
"As a side note, I think most of the customers at Best Buy, OfficeMax, and other retail stores just want to buy a computer. They don't want to pick up a case, then a power supply, then a processor, then memory. And, being as old as I am, I remember when you had to do that."
Here's the thing. I can still buy a case, power supply, processor, and memory, and make them all work together, and I can go to Dell and just buy a computer, and trust them to do it for me. But everyone knows that tech support sucks, and so if you do it yourself, you've got a much better chance of knowing what went wrong, if you're careful. So there are reasons for just going to Dell, and there are reasons to not do it.
No one's saying "Microsoft, don't make a browser. And don't use an HTML rendering library." We're saying "Microsoft, start acting like the rest of the damned industry, and work together." Everyone else already makes standards-compliant parts (more or less).
That is why they were declared a monolopy!
Exactly. They are a monopoly. And that's what we're trying to fix.
Things like Internet Explorer and Windows Media Player have become part of the operating system? Why? It's good for developers. Need to view a HTML or XML based helpfile? Just use the built in Windows functions.
/etc/alternatives.
Plus, define a standard for the way that things are launched. If you want them to stay in library functions, publish the specs. Do you know how easy the Mozilla people could write a DLL for their HTML renderer? And have you seen Firefox lately? Dear God, it's so much faster than IE in rendering.
It's called an API. Microsoft is not publishing the API for the HTML DLL, and that's just crap. I can, of course, install Firefox on Windows, but Windows will still use the IE renderer anytime the DLL is called.
First we had IE. Next we have Media Player. Then Messenger. Then Zipped folders. (Notice no one complained about that?)
C'mon! Who doesn't see a pattern here? MS just needs to open the damned API, and everyone would be happy.
The problem is that Microsoft is extending the idea of "operating system" to equate to "desktop". Everyone who uses Linux knows this is a pile of crap. Microsoft does not control a desktop environment. They control an operating system, and their control over the operating system has allowed them to slowly start to gain control over the desktop environment. And again, that's crap.
because they simply won't make the corrupted version public.
Yah... wouldn't source code that's not public be... closed source?
So he's claiming that open source is dangerous because it could become closed source. And closed source is better, because it's more protected against... uh... wait.
I think you're failing to get the basic point. Our electric supply is 220V or 110V or whatever. High tension cables are 10kV or whatever. Batteries are 1.5V or 3V, not 5A. In almost all electrical circuits, we control the voltage first, and then vary the current by adjusting load. If I'm living under a high-voltage cable, I wouldn't write to the electricity company and say "please reduce your current", because that depends on what consumers are using. But I could say "please reduce your voltage"
No. The ozone production is related to the current generated across a material containing oxygen. You control that current by increasing the resistance. You do that by increasing the distance between the two things that have 10 kV across them. This changes the voltage gradient between the two objects. The current flowing through the wires only affects the magnetic field, not the static field and definitely NOT ozone production, regardless of what other people in this thread might think...
In other words, if you want to prevent ozone production between high tension wires, don't let them get close enough to emit a coronal discharge. Also don't let people get close enough to emit a coronal discharge.
In a very real, physical sense, voltage determines current and not the other way round. Not in this case - in this case, current (and time, I guess) determines current, because air, like all materials, is nonlinear near its dielectric breakdown point, and so the resistance is dependent upon the current flowing across it, which in turn determines the current flowing across it. In a normal ohmic device, you'd be right, but near dielectric breakdown, it's the current that matters. If the number of electrons moving through the medium is enough to create an ionization path, the resistance drops like a rock.
Air molecules bumping into it and picking up (or dropping) excess charge.
Again, though, it's a voltage gradient issue. How far do you need to go before you can consider "air" neutral? Thus, you can compute an effective "resistance" between those points, and then a current flowing from the balloon. That current is never going to be anywhere near breakdown, and so the resistance will be huge, and the current flow will be virtually nil, and the ozone production will be nothing.
Look, the basic point is that you can't just say "well, 1m away from a capacitor charged to 10 kV, ozone production is 10 ppb/hour." You need to know the spacing of the capacitor and its dielectric breakdown voltage. In other words, you need to know the current across the capacitor. Measuring current dynamically is easy - measuring resistance dynamically is impossible. If the plates had a current going through them of 2A, I can bet that it produced a lethal amount of ozone, as it was arcing the entire time! If it was 2 nA, then that's different, as it's exactly what you'd expect from a negative ion source generator.
Again, same question if they're dealing with rats. Was it arcing TO the rats? That can be a hard thing to tell (although typically the rats would be a bit jumpy:) ) You can't tell that with the voltage. You CAN tell it with the current.
Simple question here: if you have a high voltage source, can you tell if it's sparking by the voltage? No. You CAN tell from the current it's drawing. Since sparking produces drastically more ozone than not, I'd say the current's important.
Never smelt the ozone coming off a CRT monitor Look inside a monitor while it comes on. Sparks all over the place. 10 kV electron guns tend to do that. Coronal discharges. Small, yes, but lots of them.
or a laser printer when you turn it on? Coronal discharges again. From the static buildup on the drum.
But as someone who understands Ohm's law, you surely realise that - since the relevant 'resistance' is a property of the atmosphere - current is entirely determined by the voltage.
No, it's not. It's set by the resistance, which is determined by the current. Yah. The current is determined by the current. Sounds screwy, but it's right. Once the current through the air hits a threshold, the air undergoes dielectric breakdown, forms a conductive path, and the current goes to hell. Again, you could either say the current causes it, or the voltage gradient causes it, because the resistance per unit length is fixed. But it's the current (because the current is # of electrons/unit time, and # of electrons/unit time is proportional to # of oxygen ions generated) that matters.
(Proof: Right before a coronal discharge, you could have 10 kV, and, say, 1 nA of current. During the coronal discharge, you still have 10 kV, but probably a few mA of current. What changed? The current - 1 nA - reached a threshold, the air underwent dielectric breakdown, and the resistance changed dramatically. Both cases - 10 kV. One case: virtually no ozone production. Other case: huge amounts. It's all about the current.)
Now, as for air ionizers...
Air ionizers generate virtually no ozone. Like, parts per billion level. Coronal discharges generate thousands of times that level, because the current is many times higher.
Next example. Take that 10kV balloon floating in the air. Move it into the stratosphere. How long is it going to stay at that potential? A very, very long time. There's no easy path to ground, and so how exactly would it dissipate? It's got to be 10 kV between something and something else. Again. Voltage gradient, or current. Not voltage.
Hell, there could be 2 gigavolts between me and the Moon. Hell if I know. But there's also an abso-friggin-lutely huge resistance, and so no current flows, and no ozone is produced.
Slashdot Mirror
Windows users make a good point that they don't want to relearn how to do what they already can.
I dunno. This isn't a good point in my opinion - it just shows they aren't computer "users" - they're Windows users (so they don't know what a computer is really capable of, only what Windows is capable of), or that they're lazy.
Here's the point, I guess. How much time a day does an average person who uses a computer spend typing? Let's say four hours, just for fun. Now let's say that that person learned how to type poorly, probably self-taught. And that person types at, say, 30 wpm.
If you asked that person "why don't you learn to type properly?" they might say "I don't want to relearn what I already know how to do" - ignoring the fact that people who type properly can type over 100 wpm, and that would save them many times the investment on time in a few days.
The real reason is that the person is either a) doubtful that the effort will produce an improvement, which is silly considering you can see there's obvious room for improvement, or b) lazy.
To me, I'm just completely confused. If I buy something, or am going to work on something, I figure out exactly what it's capable of and how to best maximize the time that I spend using it. That used to mean reading the manual (gasp!) but manuals are so pathetically bad nowadays that I usually end up just poking around on the Web trying to find other people's experiences with it. If you aren't willing to spend time maximizing your ability to use a piece of equipment, you're lazy. It's just that simple.
Fedora kernel crash on all my SMP systems
I hate to say it, but someone coming from Fedora is still not a user - they're a Red Hat user, and Red Hat has done quite a bit to make the system "like" Windows. If you expect Debian to act "like" Red Hat, you'll be disappointed. If you expect Debian to act "like" Windows, you'll be disappointed.
If you don't expect anything of Debian, and examine it for what it is, you'll be very impressed.
I don't want to describe all technical problems I had
No operating system works. There are technical problems with every operating system in existence. I can't use plenty of hardware on a Windows XP box. There's no chance in hell of me using my old LANding Gear network adapter on Windows XP. It's a tremendous effort to get the Acer NeWeb WarpLink wireless adapter to work, too - and both of those took only a tiny amount of work in Linux. The question is not "how easy does this appear at first glance?" but "how easy can it become?"
Debian has a huge userbase, and a huge community. Did you ask for help? Did you search for help? File a bug?
It takes exactly 45 minutes and all work perfectly.
Amazing. A former Fedora user thinks Fedora works perfectly. Never would have guessed.
Or maybe it is if you have only debian systems and if you are ready to lose your stability if you don't follow the Debian Way to configure something, I don't know.
In other words, "this operating system sucks if you don't learn how to use it!" Good call. I've said this elsewhere, and I'll say it again. A high learning curve does not make something not user-friendly. Someone who isn't willing to climb a learning curve isn't avoiding software because it's not user-friendly - they're avoiding software because they're lazy.
A system that install itself without problem and *just work* is friendly.
Nothing like this will ever exist for all users and all configurations. What you want is a system that "just works" for you. If you're someone who works "exactly" like Red Hat wants you to work, then Red Hat will probably "just work".
Note: Don't reply to this if you just want to defend Red Hat/Fedora. Fedora's a good distribution, and it aims to be a lot like Debian. What I'm trying to point out is that if you put no effort into getting a new operating system/distribution to work - and only two days for a complete newbie is not effort - you can't claim "it's not user-friendly!" The truth is, you were lazy. Everyone has to put in effort to learn an operating system - even Windows (ever wonder why they sell those "learn Windows now!" CDs?). What makes a userfriendly distribution is one that is friendly to the user - that allows the user to do whatever he wants, and helps as much as possible.)
Wow, so from what you are saying, Debian is the Alpha and Omega, the beginning and the end, of all operating systems.
Click on a parent and read its replies before you reply - I accidentally forgot to add a bit afterwards (Submit and Preview are too close!). OS X is very much the "ideal", though the lack of completely open software is a slight drawback. It's better than Debian, though for people who are more technically inclined, I think the open source portion of Debian makes it more user-friendly.
(For instance, my SiPix Blink2 that I just bought works in gphoto2 perfectly fine. However, Debian's libgphoto2 package accidentally wasn't compiled with libjpeg support, so it didn't work. Whoops. However, a bit of poking around and I figured this out - asking on the gphoto2 Sourceforge site helped! - and a simple apt-get source libgphoto2-2, apt-get builddeps libgphoto2-2, dpkg-buildpackage, and everything's good. If a system library accidentally was forgotten on Mac OS X, you'd have to wait for Apple to come out with a fix. Which would happen, sure, but not in the less than 1 day that I fixed it under Linux. And by the way - it works better under Linux than Windows. The Windows driver postprocesses the image in an absolutely horrendous way. This isn't 100% true, of course, since I think gphoto2 works under OS X, but you get the general idea)
The mantra for any operating system, in terms of usability, should be "Attempt to do what is sensible, but in the end, always let the user decide."
Disagreeing with a comment, but not giving any reasons.
Here's a hint: that's the stupidest thing a person can say. I'm glad your brilliance outshines the rest of humanity so much that any comments you have would be lost on the rest of us.
For us silly mortals, we have to rely on intelligent discussion in order to actually learn.
(p.s.: with a UID of 200586, you either a) haven't been here very long, or b) haven't bothered to contribute to discussions at all. If a), then you lied, and if b) then you've no right to talk about what's dumb and what's not, as you've been silent the entire time.)
I didn't say that Debian should steal software from OS X, just implementation ideas (i.e., make a pretty GUI for editing all of the conffiles, AND allow people to edit the conffiles as well).
If you have software that's not open, and not free, fundamentally, it's not as user-friendly as software that is open source, because then, the user can change it, and the user is the only person who knows what his or her needs are.
It's the one limitation of OS X, but, honestly, Apple spends a crapload of time with usability focus groups, and most people's needs aren't *that* different, so it's not a serious limitation of OS X. Microsoft probably does the same, but my God, they must do a terrible job, because in terms of usability, their products are so far behind it's crazy. Don't like Messenger as an AIM client - and who would? - try disabling it in Windows XP. It takes serious effort to kill the damned thing, as a ton of other programs launch it as well. Want to run a script every time a connection is detected (like updating a dDNS connection, or setting up an open port on a wireless router)? Ha! Good luck. We all know these things are a joke to do inside the OS, but to normal people, these things just "aren't possible". Windows is worse than just "not user-friendly" - by being so pervasive it makes people think that it defines what a computer can do, and therefore, Windows' limitations become a computer's limitations.
Uh, yes it is. That's why so many people I try to introduce Linux to don't want to switch. Linux is too much of a hassle to use.
Read the post again. I said
The problem is that it's so pervasive that everyone's used to believing that user-friendly = Windows-like = "everything just works"
and
No, it's not the most Windows-like system around.
People who start off using Windows learn its quirks and idiosyncracies and think of them as "normal". They're not. Linux isn't "Windows-ex-user-friendly", but I'm glad it's not, because Windows isn't userfriendly to begin with. In fact, there are quite a few different paradigms that Linux has that the Windows paradigm doesn't have that are far more user-friendly. See WindowMaker, for instance, with the NeXTSTEP interface, or Emacs with almost everything bound to keybindings, or LyX.
A high learning curve does not make something non-user-friendly, especially when there are rewards for the high learning curve. There's absolutely no doubt that there are more powerful tools under Linux - Emacs was virtually designed from the ground up to allow people to edit files as fast and as easily as possible (hence the reason that cursor editing keys are all control-combinations of home row keys). A person who chooses not to go through the "hassle" of not climbing a learning curve which has obvious benefits is not avoiding the program because it's not user-friendly - they're avoiding the program because they're lazy
Start off with people who have never used a computer (or at least, never used Windows), and are willing to learn to use one, and they'll learn Linux rather easily. That's how most of us did.
Only tech-nerds like us think that way. That's a made-up definition of user-friendly.
Am I a user? Yes. Is an operating system that doesn't let me do what I want non-friendly to me? Yes. Then it's not user-friendly, now is it?
Last time I checked the definition of user-friendly is "friendly to the user". If an operating system doesn't let the user do what he wants, it's not being friendly, now is it?
why she should change from something that "already works."
She already had a Windows mindset - that is, "Microsoft is smarter than you. You only want to do what Microsoft lets you do. You do not want to do anything else. Microsoft is good to you." She's not a user - she's a Windows user.
Windows doesn't "work". No operating system works. There's at least one thing broken about every operating system/distribution in existence. The question as to whether or not it's user friendly is whether or not you can deal with the broken parts well.
Whoops. I meant to add at the end "or OS X", because both of those really are the top end of user-friendliness around, though OS X is even better. Both of them allow you to do whatever you want, AND they try to give you a hand by doing as much of it for you as they can.
OS X is really a marvel, though - it's what Debian really should strive to become.
Second, if you use debian, you have no right to bitch about Linux being difficult to use. It's not exactly the most user-friendly system around.
No, it's not the most Windows-like system around. It is the most Linux-like system around, and it's absolute craploads easier to use than Windows. Everything is documented, and everything is modifiable.
As a simple example: Windows XP doesn't handle wireless connections terribly well - if I standby my laptop with one wireless connection that uses DHCP, and then wake it up in an area where it has a different wireless connection, it doesn't release/renew the DHCP lease. I have to do it myself. This is stupid - on a Linux system, if the distribution was screwed up, I could just script it in a moment's notice.
Windows's help system is also a joke - most of the programs don't properly document what things do (the number of times I've seen "There is no help available for this option...") and so you're left hoping that things work.
Windows is by far one of the least user-friendly operating systems around. The problem is that it's so pervasive that everyone's used to believing that user-friendly = Windows-like = "everything just works". That's not true, because no operating system just works, because no operating system knows everything you could possibly do with it.
Linux forces you to learn about a problem before solving it. That actually makes it very user-friendly, because it means that users can realize that they can do more than what they originally thought they could do - meaning the OS makes them more productive.
Windows isn't user-friendly. The simplest way to illustrate that is to ask this: how much does it allow you, the user, to do, and how much does it try to do it for you? An operating system that does everything for you and allows you to do nothing isn't user-friendly, because what if you don't want to do what it wants? An operating system that allows you to do everything but does nothing for you isn't user-friendly, because, well, it's a computer. It can do things automatically. The best operating system is one that tries to do everything for you, but allows you to do everything as well, and that's Debian.
One is proton decay, which means the building blocks of any sentient computer will eventually decay on their own.
You do realize that no one has seen one proton decay. Not one, right? Proton decay assumes supersymmetry is valid, and as many physicists have noted, supersymmetry is an excellent theory, which predicts a whole host of particles - half of which have been discovered.
Proton decay isn't real - not yet. And there is no a priori reason to assume that it is. Its current lower bound is 10^33 or so years. At this point you can't simply claim that proton decay happens - in fact, you'd rather claim that it doesn't.
I was under the impression that dark matter needs fine tuning to explain Tully-Fisher
Yes and no: The typical Tully-Fisher coefficients for Sa, Sb, and Sc type galaxies are 9.95, 10.2, and 11.0 or so. These are all within 10%, and for Sa and Sb types, within 5%, of 10. Simple assumptions get you a coefficient of 10, if you assume that the mass-to-light ratio is the same for all spirals, and that the surface brightness is the same for all spirals.
The first assumption (mass-to-light ratio) is a clearly idiotic assumption. It assumes that galaxies form with same proportions of light and dark matter, which we *know* is not true for other types of galaxies (dwarf ellipticals, in particular). Aside: This is also the "nail in MOND's coffin", more or less - MOND was hoping to replace the dark matter hypothesis by saying physics works differently at large distances. The problem is that galaxies which contain the same amount of light-emitting matter and have the same spatial extent should therefore have the same rotation curves. This isn't true. You then have to add a new parameter with MOND to fit it, which is OK, sure, but now you've started to lose the elegance originally intended, and now MOND becomes a more complicated theory than the dark matter hypothesis, which just says "well, that galaxy formed around less dark matter."
Anyway, back to the subject: the point is that those two assumptions clearly are not completely true, and therefore there's plenty of room for a 10% correction due to forming biases in spiral galaxy types. If the mass-to-light ratio is a very weak function of mass (which is believable - perhaps smaller galaxies formed when the dark matter density was slightly lower, due to their late formation times), you can easily get those corrections.
MOND allows you to get that 10% correction due to the parametric fit of the rotation curve, which is essentially identical to the way that it's done in the dark matter case - the corrections are due to the variation in the rotation curve, which MOND says is due to a modified Newtonian field, and dark matter says is due to a dark matter density. It's the same reasoning - one isn't more natural than the other.
(It should also be noted that the Tully-Fisher data has a crapload of spread to it, just like all astronomical data. Each galaxy varies a fair amount.)
It can also explain phenomena which the dark matter hypothesis can't, such as the Tully-Fisher relationship observed in the surface brightness of galaxies.
The Tully-Fisher relation has been explained by dark matter for some time. You can find a brief derivation in Carroll & Ostlie p. 1002, for instance. There's no need to invoke MOND at all - it just comes from the fact that the luminosity is proportional to the maximum velocity to the 4th power, which you can get by using the expression for total mass contained within the galaxy derived from rotational velocity curves.
Now it is clear that you have no idea on the concept of interferometry either.
No, of course I do. I never said it would be easy, but it's not impossible. I'm a bit confused about the necessity of a completely stable site (the 30 meter number came from the Thirty Meter Telescope, which has planned sites for places like Hawaii - not exactly what one would consider a "perfectly stable site") - I don't think that the lunar surface would be a tremendously bad placement for it. The control you're immediately throwing out as "ridiculous" is definitely within reach nowadays. Technology's amazing.
Your argument is basically: assuming no launch costs, no personnel costs, a cost-free fully-outfitted manufacturing facility (that needs to make mirror substrates out of the lunar soil!
Yes, that would be exactly what I'm assuming. What would need to be launched that wouldn't be used for something else? Not a tremendous amount, for one. The personnel are used for other things, the manufacturing facility is used for other things, etc.
If you're trying to say "if we go to the Moon to build a telescope, it'd be insanely expensive!" I agree with you. Very much so. But that's not what I've been saying. What I've been saying is "if we go to the moon, and establish a permanent presence, building a telescope is a good idea."
I'm not sure how far in the not-to-distant future you are dreaming
I'd say 50 years is a reasonable estimate. If Bush is serious about the return-to-the-Moon thing.
Your comment about steel mills is way off too. These mills live and die on the margins.
Um. Yeah. Duh. That's because there's no money in small-scale extremely expensive totally cost-inefficient steel production, which is what I said.
They are large and power hungry because of thermodynamics.
No, they're power hungry because of thermodynamics. They're large because of scale. They would lose money trying to make steel smaller-scale than they do now - it's not that it's impossible. Not unless you're trying to say that iron and/or steel has never been made in anything except gigantic foundries. Besides, that's again assuming there's no other material on the Moon that could make a strong support besides steel. I wouldn't even think of steel, as the materials aren't easily available.
I mean, c'mon. Do you think no one's thought about "how do we manufacture things on the Moon?" Of course they have, and there's an absolute flood of information about it. Here's just a start.
Anyone can simply say "oh, we can't do it, it's too complicated, requires too much work, etc." Not really. There's certainly no reason why we can't build a telescope out of materials on the Moon. It's just hard. And requires people to think, and be creative, and come up with new solutions to things, which is usually harder to convince people to do in a culture where people are constantly happy with the status quo.
Yes, but Dell doesn't sell computers that won't work with Dell brand processors, or will only work with Dell brand video cards, and Dell brand hard drives, and Dell brand monitors, and Dell brand power supplies.
Yah, they used to sell some of those (Dell brand power supplies, mainly). And that sucked. However market pressure (and the fact that standards are cheaper!) pushed them to using standard parts.
Dell follows a standard - the PC standard. They have a BIOS, so a RAID card plugged in will boot, standard USB ports, standard PCI slots, standard memory slots. Even Dell's laptops now use standard parts: SODIMMs and MiniPCI cards. Standardization, with systems being supplied by multiple vendors, is a good thing. Dell doesn't have to sell the individual parts - not at all. But they don't prevent you from obtaining compatible parts.
Of course they don't, though. People would just go to other vendors that you can buy the other parts for at Best Buy when you need them. But if Dell was a monopoly, they could. And then you'd be screwed. And left with a lot of incompatible parts strewn around Best Buy because Dell's abusing its monopoly.
That's a good analogy for Microsoft, because that's what RealPlayer, Mozilla, WinZip, and AIM are slowly becoming. Incompatible parts, because Microsoft's abusing its monopoly.
Well, that's the reason that the LGPL was originally called the "Library GPL", now wasn't it? OK, maybe not, but the LGPL is a much better fit for libraries as it doesn't have that fuzziness, unless you really want to prevent the code's use in proprietary software (which is their choice, of course).
I guess I should've been more specific... most of the libraries I've ever seen are LGPL. Briefly looking through Freshmeat, I can barely find any that are GPL rather than LGPL'd. (The FSF wants people to release libraries under the GPL to strengthen free software over proprietary, but I don't think many people see it that way. They just want to have the software available, and want due credit.) What library were you having concerns over?
To be honest, I doubt it would be that much of a chore to ask the library's developer to license it to you under the LGPL for a specific project. Monetary incentive could come in handy here, as well. It's important to remember that the GPL doesn't forbid dual licensing - it can't, of course: the author is free to license something however he or she wants. Granted, cases where there are many authors involved could be difficult.
The idea of using in situ resources is great. But I wonder if it will really be a short term or medium term solution (on the scale of twenty years, say.) We have certainly been talking about asteroid mining &c, but I wonder if the costs for sending up infrastructure are feasable?
:)
The benefit is that we're talking about things that can directly feed money back into the economy again. Small-scale manufacturing? There are tons of places that would want that!
How much money would it cost to build a small probe designed to go to the Moon and try to manufacture alloys (say, bricks) out of lunar regolith? If that's *all* it was designed to do, probably not much - say, maybe, 100 million. So there's your testbed, and then you build a larger scale and send it. Maybe in pieces so you don't have to worry about it surviving completely intact. Then when humans land, the infrastructure's already there.
Think about it this way: the mass of the systems used to construct, test and characterise the PSF of a mirror are huge compared to the mass of the mirror itself. You would need (I think) technology on a near magical scale to get that down.
This is what adaptive optics have really helped in - the mirror itself constructs an ideal PSF. The JWST, Hubble's successor, is not going up as a completed mirror, but segmented, that will unfold when it gets there, with adaptive optics optimizing the shape. This is exactly what you would want to do on the Moon, just probably an order of magnitude or more larger.
The technology you have will improve continuously so that the best giant mirror you can make now will suck compared to the one you can make later.
Well, this is somewhat true, and somewhat not true. Aperture wins, all the time, so long as the other effects are marginal. Adaptive optics basically can mean that your mirror can be as good as your electronics, since it needs to adapt to thermal variations, and replacing electronics is easy (well... control electronics. I wouldn't suggest replacing thousands of transducers, though this probably would be possible). The age of big, giant mirrors is pretty much over. Only a few of the huge, large scale next-generation telescopes are single-mirror types: both JWST and the thirty-meter-telescope are segmented designs. This is because electronics has sped up at a rate much faster than optics.
But let's see NASA get together a working probe that can build a stop sign out of in situ materials so that we can characterize exactly how cost effective your suggestion is.
Part of the problem that I have with the current pace of NASA is that it's honestly insane. Small, proof-of-concept missions are really the best way to go. Like Pathfinder, or Deep Space 1, for instance. Actually, in my opinion, I wouldn't've kept Pathfinder running considering the cost of operating it. I would've put that money towards its successor. NASA used to say "faster, cheaper, better" - but I'd say just "more often" would be better. Momentum, momentum, momentum.
I mean, people out there have to have ideas about how to manufacture things off planet. Plenty of scientists post proposals about it all the time, so the science is there. It's just the engineering that needs a kickstart, and the best way to do that is incrementally.
You clearly have never seen, or are aware of, the infrastructure needed to manufacture and test very large telescope mirrors.
Yes, I am. Very much so. If I had been talking about building a single 30 meter mirror, you'd be right. That would definitely need one heck of a testing facility and a lot of support facilities as well. However, you don't need that. Adaptive optics have gotten really good in the past 10 years, and so you don't need one gigantic contiguous piece of glass. With adaptive optics, you can use many smaller, easy to manufacture pieces of glass very easily.
(until you go and see how big they are and how much power and other resources needed to run them);
Modern steel mills are large and power-hungry because they can be. There's no driver to make them smaller, and it would definitely cost more money. If you force someone to think about how to build a small-scale steel mill, they'll come up with it.
Whoa! That one about knocked me out of my chair!
Ah, so that explains why you didn't read the rest of what I said in the sentence you quoted!
Assuming that lunar human infrastructure exists, and lunar manufacturing capability could be constructed
would be the part that you left out. The expensive part about building a gigantic telescope in space are the launch costs, and any repairs that would need to be done, and the necessity of all of the support personnel. If the launch costs are eliminated by lunar manufacturing, the repair/retrofit costs are simply absorbed into the cost of maintaining a permanent lunar presence, and the support personnel are no longer constantly required, then it would be quite cheap compared to a space-based telescope.
Even if the GPL'd code is part of a binary static lib, there's still some fuzziness on the use of a GPL'd API. Enough fuzziness that my company's management didn't want to risk using it.
Huh? I'm confused. They didn't want to risk using a GPL'd library?
Yes, you can't statically link to a GPL'd library with a non-GPL'd piece of code. That's not stupid, because no one would know that you're using GPL'd code. In essence, you're saying "Thanks for the code - now screw off!" to the people that created it. They also don't know if you changed it, improved it, etc. Essentially you're stealing their work without helping them at all. I don't consider the prevention of this a bad thing.
But why would you care about statically linking a GPL'd library? What's wrong with dynamically linking it? If you're worried about people not having the dynamic library, ship it with the product - along with the source of the dynamic library, which you have, because it's GPL'd.
I never understood the "fuzziness" part...
Hmm, do the authors have a bias for space telescopes, I wonder?
All of their arguments are correct - for now. The problem is in one word scale. Physically, extremely large telescopes aren't really feasible to send into space - to LEO or to the Moon, regardless.
However, oddly enough, the moon - being basically made out of the same stuff Earth is - has all the raw materials to make mirrors.
So the idea isn't to send a telescope from the Earth to the Moon.
The idea is to build it on the Moon in the first place.
This would require a lot of R&D, and a lot of work - but that's just R&D, which can be much cheaper than the launch costs of a gigantic telescope (a many-meter telescope).
The authors also then go to say "why would we want to keep a state of the art instrument around for a long time?" implying, of course, that state of the art changes so rapidly that keeping something around for a long time is meaningless.
They somewhat fail to see one thing: a gigantic mirror will always be state of the art. If the instruments need to be changed periodically (say once every 10 years) then the marginal cost of replacing instruments on a telescope on the Moon (assuming that the capability of getting to the Moon exists) is going to be infinitely less than launching a new state of the art space telescope.
Basically, the authors are assuming that
While this is certainly true now (and validates the reason that their one example - LUTE - wasn't chosen, as it would have been an unmanned observatory) - it is definitely not arbitrarily true for the future. Assuming that lunar human infrastructure exists, and lunar manufacturing capability could be constructed, the cost of building a telescope on the Moon is tremendously cheap.
At least we should boost Hubble to higher orbit, so when NASA gets additional funding, it can try again to bring it down. Putting it in a museum somewhere would really be a inspiration to many children to go into science.
C'mon, that's crazy. It's like suggesting we should've brought Mir down in pieces in a shuttle or something. Hubble's an old space telescope, and we've thrown many old space telescopes away.
You could build a replica for a fraction of the cost that it would take to bring it down. That'd be good enough for inspirational purposes.
The only reason that people are averse to doing it now is because somehow "Hubble" got a lot of public support, but it really doesn't deserve it. It's just an old telescope. Sure, it does good science - but so does any instrument if there are people operating it. The point is "is it worth spending money on something when that money could be better spent on a better replacement?" and the answer to that is "no".
NASA never should have let Hubble get into the public's eye this much. The pretty pictures can come from pretty much any other telescope (Spitzer put out a few nice ones over the weekend) - there's no reason to keep fawning over Hubble.
Because it's not complete. It describes how you *can use* the HTML DLL, not all the ways to use it. In other words, it would allow you to use it, but not to replace it.
(Note that giving a reference implementation would be remarkably simple, given the breadth of experience that Microsoft has with HTML rendering.)
Mozilla for a while had the beginnings of a replacement of mshtml.dll, but there were significant amounts of "undocumented features" with regards to the DLL, and Microsoft also tended to change them, making it basically impossible.
Now, maybe I am a little old (ok real old), but isn't Microsoft just completing the migration of the Windows UI into the OS?
First off, Microsoft has allowed themed UI since Windows 95. Which means they've been perfectly happy to give people control over the user interface. Seond off, they readily ignore their own UI anyway - see Windows Media Player.
Anyway, A web browser is not UI. It's an application that displays Web pages. It doesn't display controls, or anything else - just Web pages. It just parses HTML into a bitmap buffer, and Microsoft should have no a priori reason to have a preferred HTML renderer in the operating system.
They are not slowly gaining control of the desktop, they OWN the desktop.
No. They're slowly gaining control over it. It used to be they weren't the desktop. You had other Office apps, you had other email apps, etc. Slowly but surely, all of the applications that people used got usurped into Microsoft.
A "desktop environment" is all "productivity applications" - applications that make a computer a general workplace tool. Not something like a CAD program, or a video game - those are special purpose, and not everyone would use it. Microsoft never owned that - nor should they.
And, being as old as I am I remember when you had to do that, and they didn't work together, and I don't want it anymore either.
A broken implementation doesn't imply a broken idea. The idea is correct. The problem was lack of standards.
Let me transpose your idea into a completely similar idea...
"As a side note, I think most of the customers at Best Buy, OfficeMax, and other retail stores just want to buy a computer. They don't want to pick up a case, then a power supply, then a processor, then memory. And, being as old as I am, I remember when you had to do that."
Here's the thing. I can still buy a case, power supply, processor, and memory, and make them all work together, and I can go to Dell and just buy a computer, and trust them to do it for me. But everyone knows that tech support sucks, and so if you do it yourself, you've got a much better chance of knowing what went wrong, if you're careful. So there are reasons for just going to Dell, and there are reasons to not do it.
No one's saying "Microsoft, don't make a browser. And don't use an HTML rendering library." We're saying "Microsoft, start acting like the rest of the damned industry, and work together." Everyone else already makes standards-compliant parts (more or less).
That is why they were declared a monolopy!
Exactly. They are a monopoly. And that's what we're trying to fix.
Things like Internet Explorer and Windows Media Player have become part of the operating system? Why? It's good for developers. Need to view a HTML or XML based helpfile? Just use the built in Windows functions.
/etc/alternatives.
Plus, define a standard for the way that things are launched. If you want them to stay in library functions, publish the specs. Do you know how easy the Mozilla people could write a DLL for their HTML renderer? And have you seen Firefox lately? Dear God, it's so much faster than IE in rendering.
It's called an API. Microsoft is not publishing the API for the HTML DLL, and that's just crap. I can, of course, install Firefox on Windows, but Windows will still use the IE renderer anytime the DLL is called.
First we had IE.
Next we have Media Player.
Then Messenger.
Then Zipped folders. (Notice no one complained about that?)
C'mon! Who doesn't see a pattern here? MS just needs to open the damned API, and everyone would be happy.
The problem is that Microsoft is extending the idea of "operating system" to equate to "desktop". Everyone who uses Linux knows this is a pile of crap. Microsoft does not control a desktop environment. They control an operating system, and their control over the operating system has allowed them to slowly start to gain control over the desktop environment. And again, that's crap.
because they simply won't make the corrupted version public.
Yah... wouldn't source code that's not public be... closed source?
So he's claiming that open source is dangerous because it could become closed source. And closed source is better, because it's more protected against... uh... wait.
Brilliant! What a moron.
I think you're failing to get the basic point. Our electric supply is 220V or 110V or whatever. High tension cables are 10kV or whatever. Batteries are 1.5V or 3V, not 5A. In almost all electrical circuits, we control the voltage first, and then vary the current by adjusting load. If I'm living under a high-voltage cable, I wouldn't write to the electricity company and say "please reduce your current", because that depends on what consumers are using. But I could say "please reduce your voltage"
No. The ozone production is related to the current generated across a material containing oxygen. You control that current by increasing the resistance. You do that by increasing the distance between the two things that have 10 kV across them. This changes the voltage gradient between the two objects. The current flowing through the wires only affects the magnetic field, not the static field and definitely NOT ozone production, regardless of what other people in this thread might think...
In other words, if you want to prevent ozone production between high tension wires, don't let them get close enough to emit a coronal discharge. Also don't let people get close enough to emit a coronal discharge.
In a very real, physical sense, voltage determines current and not the other way round.
Not in this case - in this case, current (and time, I guess) determines current, because air, like all materials, is nonlinear near its dielectric breakdown point, and so the resistance is dependent upon the current flowing across it, which in turn determines the current flowing across it. In a normal ohmic device, you'd be right, but near dielectric breakdown, it's the current that matters. If the number of electrons moving through the medium is enough to create an ionization path, the resistance drops like a rock.
Air molecules bumping into it and picking up (or dropping) excess charge.
Again, though, it's a voltage gradient issue. How far do you need to go before you can consider "air" neutral? Thus, you can compute an effective "resistance" between those points, and then a current flowing from the balloon. That current is never going to be anywhere near breakdown, and so the resistance will be huge, and the current flow will be virtually nil, and the ozone production will be nothing.
Look, the basic point is that you can't just say "well, 1m away from a capacitor charged to 10 kV, ozone production is 10 ppb/hour." You need to know the spacing of the capacitor and its dielectric breakdown voltage. In other words, you need to know the current across the capacitor. Measuring current dynamically is easy - measuring resistance dynamically is impossible. If the plates had a current going through them of 2A, I can bet that it produced a lethal amount of ozone, as it was arcing the entire time! If it was 2 nA, then that's different, as it's exactly what you'd expect from a negative ion source generator.
Again, same question if they're dealing with rats. Was it arcing TO the rats? That can be a hard thing to tell (although typically the rats would be a bit jumpy
Simple question here: if you have a high voltage source, can you tell if it's sparking by the voltage? No. You CAN tell from the current it's drawing. Since sparking produces drastically more ozone than not, I'd say the current's important.
Never smelt the ozone coming off a CRT monitor
Look inside a monitor while it comes on. Sparks all over the place. 10 kV electron guns tend to do that. Coronal discharges. Small, yes, but lots of them.
or a laser printer when you turn it on?
Coronal discharges again. From the static buildup on the drum.
But as someone who understands Ohm's law, you surely realise that - since the relevant 'resistance' is a property of the atmosphere - current is entirely determined by the voltage.
No, it's not. It's set by the resistance, which is determined by the current. Yah. The current is determined by the current. Sounds screwy, but it's right. Once the current through the air hits a threshold, the air undergoes dielectric breakdown, forms a conductive path, and the current goes to hell. Again, you could either say the current causes it, or the voltage gradient causes it, because the resistance per unit length is fixed. But it's the current (because the current is # of electrons/unit time, and # of electrons/unit time is proportional to # of oxygen ions generated) that matters.
(Proof: Right before a coronal discharge, you could have 10 kV, and, say, 1 nA of current. During the coronal discharge, you still have 10 kV, but probably a few mA of current. What changed? The current - 1 nA - reached a threshold, the air underwent dielectric breakdown, and the resistance changed dramatically. Both cases - 10 kV. One case: virtually no ozone production. Other case: huge amounts. It's all about the current.)
Now, as for air ionizers...
Air ionizers generate virtually no ozone. Like, parts per billion level. Coronal discharges generate thousands of times that level, because the current is many times higher.
Next example. Take that 10kV balloon floating in the air. Move it into the stratosphere. How long is it going to stay at that potential? A very, very long time. There's no easy path to ground, and so how exactly would it dissipate? It's got to be 10 kV between something and something else. Again. Voltage gradient, or current. Not voltage.
Hell, there could be 2 gigavolts between me and the Moon. Hell if I know. But there's also an abso-friggin-lutely huge resistance, and so no current flows, and no ozone is produced.