Relying on a hardware designer for process protection that should be provided by running on separate processors is an old idea and a bad one--unless you care more about convenience than real reliability.
Please keep in mind that MS Research is quite a different beast than the production departments of MS. MS Research does a lot of respected work. They also employ some of the most reputable researchers in software and OS development, including:
Tony Hoare (quicksort, CSP, Hoare logic, among other things)
Leslie Lamport (TLA+, the bakery algorithm, the byzantine generals problem, LaTeX, among other things)
I dislike MS production software and business practices as much as the next guy. But don't make the mistake of underestimating MS Research just because you dislike MS.
Unfortunately, I don't that Mountain Sun ships outside of Colorado. Does give you a good excuse to visit Boulder CO though:-) Their beer is certainly far better than the mass-produced-in-chemical-factories stuff.
You have to be able to consider each and every interaction that may occur.
Needless to say, having a formal model of your system will help substantially with the task of considering each and every interaction that may occur.
Not that I'm suggetsing you should rely exclusively on formal proof. Rather, that your arguments against formal proof are really arguments for some kind of formality.
Why can't I just double click on the executable and have the installation process start?
Sigh. Why do you Windows types always have to make everything so complicated? Double-clicking? Installation process? What's wrong with just dragging and dropping the.app you want to install into your Applications folder?
Aside: yes, yes, I know that there are some Mac apps that have "installers" that perform an installation process which writes files to various parts of the filesystem aside from the applications folder. For legacy Unix apps I suppose this makes sense. OTOH I can think of a number of apps that use their installer simply to write a bunch of data to ~/Library. Frankly, I see no reason why those apps couldn't simply be drag-n-drop.apps, which do the necessary ~/Library mods the first time they're run. The only other place where I think installer packages are a valid way to go is for core OS X stuff, which most likely needs to spray stuff around the filesystem. All of this is somewhat beside the point, which is that installation could even easier that the parent poster seems to think.
Funny. I regularly use Apple's X11 server - which adds another layer on top of the standard Mac OS graphics subsystem FFS - to run X11 apps on my Mac, and I don't experience any of the problems that you are alluding to. My only gripe is that the integration between X11 and Aqua isn't quite as nice as I would like, but that's a UI issue rather than a performance one (and it's pretty much a function of the different UI models, so I'm not sure how it could be fixed). As others have already pointed out, perhaps your problem isn't X11 at all...
I suspect that you'd get a better reaction if your writing was more coherent. I eventually managed to figure out what I think your message is ("minimum wage laws are bad" - please correct me if my interpretation is wrong). But it took a while, and required reading your responses to other comments.
But that doesn't mean that all attempts at any kind of proof (either through testing or through formal deduction and inference) are a waste of time. It's not a question of "proving something 100%". It's a question of gaining sufficient confidence that the system will work as intended. How much is "sufficient"? That depends on the application. In some cases, testing is more than sufficient to build the necessary level of confidence. The key thing to keep in kind is that formal logical proofs are able to cover a much wider range of inputs than testing can possibly cover, so for high-confidence applications they start to make more sense. That doesn't necessarily mean that you have to logically prove the entire app, just the "critical" parts (while only testing the non-critical parts).
As an aside, there's a good case to be made for logical proof actually being more cost-effective than testing - see, for example, this paper.
However, from there, I think the IDist would say, "life is too complicated to have arisen by Darwinian processes"
I think we need to be very careful here to separate evolution (which was well-known before Darwin, and well-supported by the fossil record) from Darwin's proposal for the mechanism by which evolution occurs, i.e. natural selection. Further, it's worth noting that Darwin's original proposal is not precisely the same as the curent accepted model for natural selection - more recent information than was available to Darwin has improved our understanding of how natural selection operates.
Furthermore, the kind of complexity found in living things is only found elsewhere in nature in things that are designed.
The reality is that designed artifacts and living creatures have quite different "kinds of complexity". For starters, living creatures don't decompose into modular parts with well-defined functions in the same way as human-designed artifacts. Instead, biological systems tend to have "components" that contribute to the performance of multiple, overlapping functions, often as part of varying groups of other components depending on which function we're talking about (human skin, for example, is, in conjunction with other parts of the body, responsible for thermal control, structural integrity, sensory perception, environmental segregation, UV protection, and so on). Amusingly enough, the best way that we know of to develop multi-functional designs like that in engineering is through the use of "genetic algorithm" based optimization techniques, which emulate natural selection to find a combination of design parameters that perform well against some user-defined fitness criteria.
I think we all agree that "life is very complicated"
Don't be so sure about that. There's work by people such as John Warfield that makes a pretty good case for "complexity" having as much to do with the perception of the observer as it does with any intrinsic properties of the observed.
I find it interesting that everyone can agree that my watch, or my laptop, is designed.
I find it interesting too. Why do we not assume that they evolved? What is it about these artifacts that makes us certain that they were designed? Is it perhaps the fact that we already know they were designed by people? Or does it have something to do with the difference in "kind of complexity" that I describe above? Is that what makes designed artifacts so readily distinguished from living creatures?
If so, what happens when we apply those techniques to things we know are designed an things we know that aren't?
If the universe was created by an intelligent designer, then what is there that wasn't designed? How do you propose to test your criteria for deciding if something was designed or not if there is nothing that you can use to check for false positives? That's one of the fundamental problems with ID - the hypothesis itself makes testing for "design" essentially impossible.
I don't quite see how "wearing a skirt" equates with "being gay". Just as one example, plenty of Scots men wore kilts back in the day. Some still do. I doubt they were all gay.
I don't think any would argue that we are all intelligently and uniquely designed.
Then you would be sadly mistaken. The fact that people do argue that we are not designed is why this debate is occuring at all. The idea that we are "intelligently designed" is certainly debatable, since there are many features of the human body that are rather idiotically "designed" (one such example being the use of the same passage for the ingestion of both air and food - which makes choking a serious hazard; I'm sure you can think of plenty of other examples if you try).
Students can make up their own minds or develop their own opinions about who they believe the "Creator" is.And where exactly did this "creator" come from? Is he/she/it not "irreducibly complex" as well? If not, did he/she/it "evolve"? If so, who designed the creator?
How did we come into being before we changed?
And how did the agent responsible for both creation and change come into being?
Intelligent design can and has been proved scientifically
Please provide links to this scientific proof.
Intelligence leaves behind a characteristic signature.
Really? And what exactly is this "characteristic signature"?
Ok, just to be clear here, the World Multi-Conference on Systemics, Cybernetics and Informatics is a well-known waste of academic space. It regularly spams people requesting submissions, has no obvious standards, and will accept pretty much any paper the authors are willing to pay to have published. It has no safeguards. It also has no respect. The MIT tool was developed specifically to prove that the conference in question was a sham.
I agree that building up from a drawing program may be a better approach. IMHO good slides are a visual aid to an oratory, not a collection of bullet points. At least, that's the ideal I aspire to - I don't often manage to get there:-( Anyway, returning to my point: since slides are a visual aid, they are better developed with a program suited to visual design, rather than one suited to textual design.
The problem is that many people try to design slides that can be both presented orally, and understood in isolation. I understand why, because I often try to read slide presentations myself, and get irritated when tehre's lots of missing detail. Realistically, that's where I see the value of your combined report/slide approach. OTOH Powerpoint and Impress already support a "notes" area for generating slide based handouts. But it hardly ever gets used. (Note to self - must use the notes area more often...)
Linux runs on Intel too. And OOo runs on Windows or Linux. I doubt Intel cares what sofwtare is running on what OS, so long as everything is "Intel Inside®".
The problem with LaTeX as a presentation generator is that good presentations in many fields (i.e. not math) are as much graphics as text, and LaTeX just isn't all that convenient to use for such presentations. Things like Powerpoint and Impress allow graphics to be dropped in, visually resized on the fly, cropped as necessary, visually arranged into a coherent layout, and combined with the built-in drawing tools. I suppose one could get a similar effect by using Draw, or some other drawing program, to combine and layout the graphics, and then exporting as a single graphic to include in LaTeX. But this is a bit of pain compared to just dragging and dropping. What might be nice is a LyX-like tool that allows a wysiwyg approach to slide layout, but emits LaTeX as the end result.
Well, maybe not New Zealand, but most federal governments.
New Zealand doesn't have a federal government. Nor do a lot of other countries. In New Zealand there's just a national government - no federation of smaller states. There are district councils and the like at the local level, but nothing like the state-level governing infrastructures that you see in, e.g. the US. In New Zealand the national police force is the only police force, your car is registered with the national government, and the law is the same nationwide, etc. etc.
We used to joke that the only redundant things in the entire systems were the heaters and the SSPAs.
And before anyone gets too carried away: yes, this was a *joke*. There were also redundant pyros, redundant thermostats, redundant switches and relays for key systems, redundant thruster valves, and the like. But the only major (i.e. large) components that were redundant were the SSPAs, and it seemed like there were about a million redundant heaters (yes, that's hyperbole as well).
In fact, to the best of my knowledge they are anything *but* overengineered.
Very true. The entire MER program was mass-constrained from the get-go. They barely fit on the launch vehicle. At some points during the design cycle the mass margin was negative, and the systems engineers had to hunt around for things to take off. There was no room to spare for over-engineering, because there just wasn't any spare mass for anything other than the bare minimum to achieve the mission. I speak from direct knowledge here, because I sat through the debates about whether or not to have two transponders (final decision: one - the SDST was considered reasonably reliable), and similar debates about the solid-state power amplifiers (the final word I heard was two SSPAs, due to their potential for failure, but that may have changed after I left the program). We used to joke that the only redundant things in the entire systems were the heaters and the SSPAs.
The/. lead-in completely misrepresents the article in question. Had the submitter actually read TFA, he would know that the Greens are actually very excited about the deal. The quotes from Gren spokespersons cover a lot of the standard ground for OSS advocacy. However, the article in question was written for the National Business review, and is primarily a "debunking" of OSS, and of the Greens' enthusiasm for open solutions.
Not that I'm saying you're wrong, but according to TFA it was "Gerald Ilukwe, the general manager of Microsoft Nigeria" who made most of these statements. I can't imagine that he is ignorant of the situation in Africa. I imagine that it's more likely he's simply trying to protect his market.
Slashdot Mirror
Relying on a hardware designer for process protection that should be provided by running on separate processors is an old idea and a bad one--unless you care more about convenience than real reliability.
I dislike MS production software and business practices as much as the next guy. But don't make the mistake of underestimating MS Research just because you dislike MS.
Unfortunately, I don't that Mountain Sun ships outside of Colorado. Does give you a good excuse to visit Boulder CO though :-) Their beer is certainly far better than the mass-produced-in-chemical-factories stuff.
Needless to say, having a formal model of your system will help substantially with the task of considering each and every interaction that may occur.
Not that I'm suggetsing you should rely exclusively on formal proof. Rather, that your arguments against formal proof are really arguments for some kind of formality.
Eh. This stuff has no alcohol content. No thanks! I'll take one of the Mountain Sun Brewery's Java Porters over this crap any day.
Sigh. Why do you Windows types always have to make everything so complicated? Double-clicking? Installation process? What's wrong with just dragging and dropping the .app you want to install into your Applications folder?
Aside: yes, yes, I know that there are some Mac apps that have "installers" that perform an installation process which writes files to various parts of the filesystem aside from the applications folder. For legacy Unix apps I suppose this makes sense. OTOH I can think of a number of apps that use their installer simply to write a bunch of data to ~/Library. Frankly, I see no reason why those apps couldn't simply be drag-n-drop .apps, which do the necessary ~/Library mods the first time they're run. The only other place where I think installer packages are a valid way to go is for core OS X stuff, which most likely needs to spray stuff around the filesystem. All of this is somewhat beside the point, which is that installation could even easier that the parent poster seems to think.
Funny. I regularly use Apple's X11 server - which adds another layer on top of the standard Mac OS graphics subsystem FFS - to run X11 apps on my Mac, and I don't experience any of the problems that you are alluding to. My only gripe is that the integration between X11 and Aqua isn't quite as nice as I would like, but that's a UI issue rather than a performance one (and it's pretty much a function of the different UI models, so I'm not sure how it could be fixed). As others have already pointed out, perhaps your problem isn't X11 at all...
I suspect that you'd get a better reaction if your writing was more coherent. I eventually managed to figure out what I think your message is ("minimum wage laws are bad" - please correct me if my interpretation is wrong). But it took a while, and required reading your responses to other comments.
No, it cannot.
But that doesn't mean that all attempts at any kind of proof (either through testing or through formal deduction and inference) are a waste of time. It's not a question of "proving something 100%". It's a question of gaining sufficient confidence that the system will work as intended. How much is "sufficient"? That depends on the application. In some cases, testing is more than sufficient to build the necessary level of confidence. The key thing to keep in kind is that formal logical proofs are able to cover a much wider range of inputs than testing can possibly cover, so for high-confidence applications they start to make more sense. That doesn't necessarily mean that you have to logically prove the entire app, just the "critical" parts (while only testing the non-critical parts).
As an aside, there's a good case to be made for logical proof actually being more cost-effective than testing - see, for example, this paper.
No. They were intelligently designed...
I think we need to be very careful here to separate evolution (which was well-known before Darwin, and well-supported by the fossil record) from Darwin's proposal for the mechanism by which evolution occurs, i.e. natural selection. Further, it's worth noting that Darwin's original proposal is not precisely the same as the curent accepted model for natural selection - more recent information than was available to Darwin has improved our understanding of how natural selection operates.
Furthermore, the kind of complexity found in living things is only found elsewhere in nature in things that are designed.
The reality is that designed artifacts and living creatures have quite different "kinds of complexity". For starters, living creatures don't decompose into modular parts with well-defined functions in the same way as human-designed artifacts. Instead, biological systems tend to have "components" that contribute to the performance of multiple, overlapping functions, often as part of varying groups of other components depending on which function we're talking about (human skin, for example, is, in conjunction with other parts of the body, responsible for thermal control, structural integrity, sensory perception, environmental segregation, UV protection, and so on). Amusingly enough, the best way that we know of to develop multi-functional designs like that in engineering is through the use of "genetic algorithm" based optimization techniques, which emulate natural selection to find a combination of design parameters that perform well against some user-defined fitness criteria.
I think we all agree that "life is very complicated"
Don't be so sure about that. There's work by people such as John Warfield that makes a pretty good case for "complexity" having as much to do with the perception of the observer as it does with any intrinsic properties of the observed.
I find it interesting that everyone can agree that my watch, or my laptop, is designed.
I find it interesting too. Why do we not assume that they evolved? What is it about these artifacts that makes us certain that they were designed? Is it perhaps the fact that we already know they were designed by people? Or does it have something to do with the difference in "kind of complexity" that I describe above? Is that what makes designed artifacts so readily distinguished from living creatures?
If so, what happens when we apply those techniques to things we know are designed an things we know that aren't?
If the universe was created by an intelligent designer, then what is there that wasn't designed? How do you propose to test your criteria for deciding if something was designed or not if there is nothing that you can use to check for false positives? That's one of the fundamental problems with ID - the hypothesis itself makes testing for "design" essentially impossible.
I don't quite see how "wearing a skirt" equates with "being gay". Just as one example, plenty of Scots men wore kilts back in the day. Some still do. I doubt they were all gay.
Neither can a lot of "humans". For that matter, there were plenty of people around ebfore the concept of PDEs even existed. Were they not human?
Lack of belief in a thing is not the same as believing that there is not a thing.
Then you would be sadly mistaken. The fact that people do argue that we are not designed is why this debate is occuring at all. The idea that we are "intelligently designed" is certainly debatable, since there are many features of the human body that are rather idiotically "designed" (one such example being the use of the same passage for the ingestion of both air and food - which makes choking a serious hazard; I'm sure you can think of plenty of other examples if you try).
Students can make up their own minds or develop their own opinions about who they believe the "Creator" is.And where exactly did this "creator" come from? Is he/she/it not "irreducibly complex" as well? If not, did he/she/it "evolve"? If so, who designed the creator?
How did we come into being before we changed?
And how did the agent responsible for both creation and change come into being?
Intelligent design can and has been proved scientifically
Please provide links to this scientific proof.
Intelligence leaves behind a characteristic signature.
Really? And what exactly is this "characteristic signature"?
Ok, just to be clear here, the World Multi-Conference on Systemics, Cybernetics and Informatics is a well-known waste of academic space. It regularly spams people requesting submissions, has no obvious standards, and will accept pretty much any paper the authors are willing to pay to have published. It has no safeguards. It also has no respect. The MIT tool was developed specifically to prove that the conference in question was a sham.
The problem is that many people try to design slides that can be both presented orally, and understood in isolation. I understand why, because I often try to read slide presentations myself, and get irritated when tehre's lots of missing detail. Realistically, that's where I see the value of your combined report/slide approach. OTOH Powerpoint and Impress already support a "notes" area for generating slide based handouts. But it hardly ever gets used. (Note to self - must use the notes area more often...)
Linux runs on Intel too. And OOo runs on Windows or Linux. I doubt Intel cares what sofwtare is running on what OS, so long as everything is "Intel Inside®".
The problem with LaTeX as a presentation generator is that good presentations in many fields (i.e. not math) are as much graphics as text, and LaTeX just isn't all that convenient to use for such presentations. Things like Powerpoint and Impress allow graphics to be dropped in, visually resized on the fly, cropped as necessary, visually arranged into a coherent layout, and combined with the built-in drawing tools. I suppose one could get a similar effect by using Draw, or some other drawing program, to combine and layout the graphics, and then exporting as a single graphic to include in LaTeX. But this is a bit of pain compared to just dragging and dropping. What might be nice is a LyX-like tool that allows a wysiwyg approach to slide layout, but emits LaTeX as the end result.
New Zealand doesn't have a federal government. Nor do a lot of other countries. In New Zealand there's just a national government - no federation of smaller states. There are district councils and the like at the local level, but nothing like the state-level governing infrastructures that you see in, e.g. the US. In New Zealand the national police force is the only police force, your car is registered with the national government, and the law is the same nationwide, etc. etc.
And before anyone gets too carried away: yes, this was a *joke*. There were also redundant pyros, redundant thermostats, redundant switches and relays for key systems, redundant thruster valves, and the like. But the only major (i.e. large) components that were redundant were the SSPAs, and it seemed like there were about a million redundant heaters (yes, that's hyperbole as well).
Very true. The entire MER program was mass-constrained from the get-go. They barely fit on the launch vehicle. At some points during the design cycle the mass margin was negative, and the systems engineers had to hunt around for things to take off. There was no room to spare for over-engineering, because there just wasn't any spare mass for anything other than the bare minimum to achieve the mission. I speak from direct knowledge here, because I sat through the debates about whether or not to have two transponders (final decision: one - the SDST was considered reasonably reliable), and similar debates about the solid-state power amplifiers (the final word I heard was two SSPAs, due to their potential for failure, but that may have changed after I left the program). We used to joke that the only redundant things in the entire systems were the heaters and the SSPAs.
There is no complaint. The submitter is an idiot who has misrepresented what TFA is about.
The /. lead-in completely misrepresents the article in question. Had the submitter actually read TFA, he would know that the Greens are actually very excited about the deal. The quotes from Gren spokespersons cover a lot of the standard ground for OSS advocacy. However, the article in question was written for the National Business review, and is primarily a "debunking" of OSS, and of the Greens' enthusiasm for open solutions.
Not that I'm saying you're wrong, but according to TFA it was "Gerald Ilukwe, the general manager of Microsoft Nigeria" who made most of these statements. I can't imagine that he is ignorant of the situation in Africa. I imagine that it's more likely he's simply trying to protect his market.