There are so many factors to performance it's not funny. You have the classical memory BUS speed, the periferal bus speed ( if you're overclocking, or have a strange multiplier ), you have your MB chipset buffering scheme ( how many delinquent requests can it keep going ). You have the performance of the chipset itself. Of course you have the huge variable of different periferals strenths and weaknesses which pull you every which way but sunday. Then of course you have hardware drivers which tend to be optimized for specific processors ( the norm being intel, but in some circumstances, like AMD's 3DNow for Voodoo's Glide ).
Then inside / around the CPU you have cache, which is a HUGE variable. You have the raw MHZ speed, you have the pipeline depth, and the latency, both of which are negatively affected by larger caches ( due to address resolution logic ). Then you have the port deth ( how many parallel accesses can the cache access ). And finally the size and bandwidth of the cache. AMD / Cyrix have gone with bigger but lower performance caches, while Intel has gone with more complex but smaller caches. Hypothetically, a larger, simpler cache will be cheaper to design, but will take up more surface area, and thus provide lower yield. To make matters worse, some programs require high speed access to a very small data-set ( and thus benifit Intel ), while other applications just use a lot of data, and anything that minimizes main memory access boosts speed. I believe Quake qualifies for the former, while Office apps ( and scripting languages in general ) benifit the latter.
AMD and Cyrix also, for a while there, worked at enhancing the internal instruction flow algorithms. Making huge branch prediction buffers, and in the case of Cyrix, producing all sorts of algorithmic optimizations that Intel strangely didn't implement.
I believe the main reason AMD and Cyrix didn't work as hard at their FPU was because it's a _really_ ugly design project. It's more fun to work on general purpose flow design and playing and tweaking a simplistic cache design, than to get dirty with all the possible combinations of floating point logic ( especially one as ugly as the 8087 family. I believe Intel owns several patents on some highly optimized implementations, so the others would have to devote some big bucks to tweak theirs without violating any laws. Not to mention, making it faster often times means taking up more silicon. Thus you have a larger die ( thus lowering yields ) and the logic is expensive to design / debug / implement to boot.
The next issue was latency. Intel, with the 80686 line ( I hate their non-informative naming conventions ), went super-pipelined, which worked great for sequential operations, but performed horribly in random branching contexts. AMD and Cyrix both opted for a narrow pipe-depth, with an emphasis on branch prediction. Thus even failed predictions had minimal penalty.
The fastest possible processor will be non-pipelined and have n-wide execution components. The reason being that each pipeline stage introduces a store and forward delay. Some stages may perform minimal operations, thus wasting 75% of a clock tick. This really hurts data-dependancy delays, since a pipelined FPU might take 15-150% longer to complete a Divide which the very next instruction requires. If all other data-dependacy paths are blocked, all the pipelining in the world won't do you any good. In the integer world, this is very common. I'm about to perform a cache missed memory fetch, but first I must calculate the address. If every other instruction is based on the contents of that memory cell, then pipelining can only hurt this particular case.
The biggest opponent to complex and optimized operations was that they would slow down the rest of the processor ( by requiring slower clock ticks ). But the device manufacturers are learning how to make different parts of the CPU run at different frequencies. ( They've long since learned how to run the BUS at a fraction of the Core ). Intel's next 80686 processor varient will have a clock doubled integer core, for example.
Still, the main reason we don't see a return to complex optimizations is that having 32 ADD components is extremly more expensive than having 2 16 deep add components. Even though you'll get a significant performance boost ( assuming you can manage that huge bus, and a potentially large number of register ports ), you probably won't make up for the added expense in shere complexity and yeild loss ( due to extra size ).
Thus, Intel went for a partially pipelined FPU which had heavy latency penalties, but improved overall operations ( especially for non data-dependant operations ).
AMD Finally headed this off by making multiple independant and fully pipelined FPU's in their Athalon. ( they spent the extra bucks to remove many of the stalling conditions caused by sharing of resources by seperate components ).
Personally I like SUN's java-multi-threaded CPU concept ( even if it never succeeds ). Basically, you have 4 parallel fully functional, non-related, non-pipelined, fully optimized functional units. There are no resource contention issues, no scheduling problems, a simplified logic design. And it's cheaper because you take away pipelining. The best part is that each of these extrememly simple components are just cookie cuts. You spend all your time tweaking the hell out of one tiny unit, then make 32 copies. Almost as easy as cache design.
I believe the Crusoe could learn from this. They already have their simplified design, they could take it a step further. Say, keep a single CPU implementation for power-critical devices. Then replicate that core 8, 16 or 32 times for a desk-top varient. Since you can control your wrapper code, you can determine what is the optimal CPU-width. I'm sure there are many cases that would allow you to submit 32 parallel instructions ( at least for the compiler ).
But you're wrong again. Perl typically executes faster on text processing benchmarks because of the builtin regexp. But you cannot conclude from this, that Perl is faster than Java per se. Perl is fast at what it does (and I speak as someone who has written over 100,000 lines of Perl code and have been writing it since 1991).
I agree that mathematical algorithms are slower in perl. I too have written nested loops of matrix manipulations, and they were inexcusibly slow. Even when you specifiy integer only, and use for( 0..100) instead of for( $i = 0 ; $i ( notice my aversion towards the bastard word "proof" ).
Perl is glue, and should be treated as such. You can't write big perl programs without having an intimiate understanding of C ( since you're bound to call a great many c-library functions, and you really should consider writing stuff in C in tandem ).
In this respect, perl has the potential to be faster than Java for certain classes of projects. In general, an all native project will run faster in java IF, the delay in JIT ( or VM loading/ initialization for that matter ) compilation is marginalized away by either a sufficiently large computationally intensive core, or through persistent operation ( serverlet, etc ). Otherwise, gcc optimizations are bound to make better work of the perl-c libraries, than JIT will on the java-byte-code.
In my experience, the load time of java applications is very frustrating. Perl has very little to initialize, and is single processed at that, so will run faster than java-native-threads for simple applications. Plus I rarely find a perl-daemon eating up 100% of my CPU time, which I have regularly found in java ( at least netscape's version of it ). As a disclaimer, these are _my_ experiences with Java, and I do not claim this to be the general case. I'm sure bad code design had much to do with it.
On OOP, I agree that Perl's backwardly compatible method is akin to the P-III's x86 mess. Still with Perl5.005, you are capable of performing most OOP operations. namely: the pseudo-hash's ability to stricly enforce hash contents, and to provide private class variables. local var strict Data-typing based on class names. ( ie "my Base::FooClass $x = new Base::FooClass" )
Unfortunately, my main reason for using perl is the vast feature set ( CPAN, integration with Apache, support for QT, Gnome, GL, etc ). If python were that supported, I would probably switch over in a heart beat. It is good to reinvent yourself every now and again..
A lesser used purpose of perl for me is for quick and powerful command-lines or shell scripting. I completely believe Perl is best suited for this, and I do not see python as a replacement here. The idiocyncracies of perl really shine in a shell environment, and I believe this to be the major problem with most programmers.. Shell attributes do not regularly lend themselves to general purpose programmers, who then label perl as cryptic.
To fully appretiate perl, you must reguarly perform such system administrative functions.
As with other comments I've read. Lack of IP would be akin to Communism. The discoverer or, more importantly, the entrepeneur should have a right to command that which they create ( at least for a period of time ). I would rather that we lead by example than by force. Let something be free because the creater wills it. Otherwise we have anarchy ( If you won't free that bone with meat on it so that the community can share, then we'll do it for you. Never mind that you almost died trying to acquire it )
And also as was written in this thread, the GPL is a form of IP ( albeit, so that others can take it form you and then sue YOU for using it ). But this is perfectly possible when communism goes awry.
-Michael p.s. Sorry if I'm sounding political.. it's not my intent.
I'll have to disagree with you here. I've done some FPGA programming myself, and it currently can't even begin to compete against commercial high end processors. Granted with the super-RISC-like nature of the Crusoe is better suited for this, and I really am pulling for the FPGA revolution, but there are several issues:
1. Production Volume. In order to be cost effective ( less than a grand a chip, more than 10 manufactured a day ), you're going to need a serious chunk of fab hardware. ( which seperates itself from the near-zero mass reproduction cost of Software )
2. Performance. This is only going to fly if it can have an acceptible performance. FPGA is too unoptimized ( not familiar enough with ASICs though I know them to be faster, though not reprogrammable ).
3. Size. For FPGA to have the same amount of logic, it requires an incredible number of unused wires and gates. The cost should be minimal because we're dealing with mirrored arrays ( like RAM ), but size and power consumption will be significantly greater for an equivalent chip.
4. Power consuption. One of the Crusoe's key design targets was power consuption ( since they knew they couldn't make the fastest chip around, they found the best balanace of technologies ). FPGA and I assume ASICs would not allow them such a level of efficiency.
In short, yes you could increase the market ( and thus lower the cost ) of straight FPGA, then provide open source logic to do something like Crusoe ( or even a purely custom-Linux Supported Core ), but it's not going to be competative.. And since you can purchase a $40 CPU with far greater power.. Why would you?
I don't want to stamp out innovation or the free market.. If you believe you can succeed, then by all means try. Heck, I'd love to be proven wrong.
Transmeta has spent a great deal on R&D, and therefore needs to recouperate both those costs, and that of outsourcing fabrication [to IBM?]. The code itself is intimately related to inner details of the processor. With the Crusoe being as simplistic as it is, published source code would greatly accelerate reverse engineering. Therefore the big boys could easily steal the idea with minimal development cost.
This is the idea of intellectual property. If someone else can completely know how you did something, then what is to stop them from doing it. Or as with the nuclear arms race between China and the US, their knowledge of how we did certain things helps them to not waste time making the same mistakes that we made.
As others pointed out, Hardware open source is a completely different beast than that of software. I can't just build the crusoe in my spare time as a recreation. I have to devote serious resources. And so I feel a bigger loss if some other industry steals my ideas, and makes their product and nobody buys mine.
Still, I think it would be nice to eventually have all Crusoe code public domain. They could benifit from the same sorts of optimizations that key software technologies have had ( e.g. with gcc or the core of Linux, where everyone has a vested interest in focusing ). But unfortunately, it becomes an economic issue. I doubt they benifit enough from Open source to outweigh the competative negatives.
I read the article, and I don't see how you can draw those conclusions. I interpreted this to mean that the brain was trying its best to adapt to critical conditions. When sleep deprived, various parts of the brain begin to shut down ( as a self defense mechanism, or simply because they cieze to function ). But in order to maximize the human capacity to think and survive, other more dormant portions of the brain are activated to take over certain operations. The specific operation sited in this article was linguistics.
I don't really see why linguistics would be important for survival. They said math skills went down the toilet ( because it is such a complex and consuming activity that there weren't any reserve portions of the brain to divy activity out to ).
They specificly raise the health risks of sleep deprevation, and how much poorer those test subjects performed. What it does show, however, is the incredible adaptibility of the human mind. Much like Terminator 2, re-routing neural connections on the fly to redundant reserve systems.
What this also suggests to me is that we are capable of using other portions of our brain when we try hard enough.. Perhaps we can meditate and awaken those dormant sections. Who knows, maybe during finals, we can get a suddent burst of intelligence even.:)
-Michael
The spell checker was the single worst contributor to the modern written word.
A few points. First, never say never. Even now.:) I believe the major problem with QM is that it is based on probabalistic waves. An earlier poster spoke about the inability to completely describe a single particle, but that these probabalistic equations describe the ratio of temporal / spatial existences ( sorry, oddly worded; am too tired ). My belief is that one day we will be intelligent enough to mathematically describe the practical reality of QM. Just as Newton had to invent a new math ( eg Calculus ) to describe his physical world. There may yet be more mathematical realms not yet understood. I've been turned on to String theory, simply because they tease us with equations that almost look right. Then go on to describe how each decade they discover a new math which gives us greater insight. Whether or not it is ultimately true is hardly relavant so long as they can continually better their predictions of our physical world. As for the statement about being boggled or not understanding. I've read that on many occasions. I believe it is a good litmus test for the present day. But I do not believe it will always hold true ( that whole never say never thing again:)
I have to disagree with you a bit here. It's very easy to learn yes and no. Even a dog or cat can learn it. It's all about associations ( pavlovian(sp?) training ). Biologically we are wired ( which would be your "prewiring" ) to react positively and negatively to various stimuli. This goes back to single sell action / reactions. Assuming evolution and natural selection, then we are the products of organisms that have successfully learned how to react positively to things that promote our life, and visa versa. The higher level organisms can associate things to those basic hard wired "good" and "bad" stimuli. I believe a dog could have 3 or so levels of association, and humans were something like 8. ( The dog hears a bell, and anticipates the light that use to mean food was to be prepared and so the dog salivates even though no food is procured ). Likewise the repeated use of the word no to a child alongside a spanking associates the word no with pain to the child. Eventually the child forgets the pain and associates no with "bad". Similarly we "learn" the word good. IMHO we have very little hard wired in. Humans in particular are probably de-evolved as far as instincts and hardwiring go. Instead, we've become dependant on our superior associativity skills. With that we're more adaptable then say a bee that naturally reacts to the ultraviolet colors in a flower.
I give little credibility to our prewiring having anything to do with our parental mating since the above adaquately explains things for me. I've not really seen any credible cases of children seperated at birth that took on the traits of their parents that could not be properly explained by genetic predisposition. Perhaps your article provides more compelling evidence. As to the relationship to quantum evolution. I'm torn on the idea. The idea of a multi-verse is personally unsettling. I never liked the sci-fi plays on the concept. But any such hypothesis that provides an essence or "will" to a seemingly chaotic or random course of events aids the understanding of things such as AI. A program, as you say may seem to only be capable of spitting back answers that we once fed it, but this is becuse they are algorithmically bound. Even our random number generators are deterministic. But what if we used a quantum random-number generator. If there were any truth to the "will" of sub-atomic particles, then strange non-random patters should arise in algorithms that make use of such randomizers. I totally believe that this sort of theory is testable. Perhaps it is the next step towards true AI. Life bound by sets of rules that man has created. Perhaps a little scary.
Re:Why not just use the Crusoe as a G4?
on
Darwin on Crusoe?
·
· Score: 1
Originally RISC => "Reduced" Instruction Set Computing. I'm not sure where you've learned that achronym.
Additionally, there are almost no true RISC CPU's any more. All modern CPU's should be considered Post RISC. Eg, not all instructions are 32 bits ( as par ALPHA and 64 bit constant loading, etc ). Additionally, they hardly execute in a single clock anymore. This was only possible when no FPU was utilized ( or at least was used as a seperate FCPU with CPU synchronization ). Additionally, the number of instructions is once again sky rocketing in these post-modern CISC chips.
I can just see chips like ALPHA, PowerPC and Sparc as having the exact same sort of legacy compatibility issues as modern CISC. Heck, Sparc's register window really can get in the way in multi-threading / tasking.
What I really like about the cruso idea is that you can make a CPU that has vector processing if you like, then translate MMX / SSE / AltiVec / 3DNow as need be to utilize your own proprietary method. This way, when a non-compatible method comes along later, you don't have to spend the silicon on compatibility.
If you read the MS FOF, then you'd know that MS DID purposefully make changes that broke DR Dos with their windows code back in the day.
I don't mind AOL being big, popular, easy to use, having proprietary content, etc. But what I do mind, is their purposefully sabotaging the competition. This goes far beyond 5.0's alleged breaking of other ISP components. It involves blatant incompatibilities with mainstream internet standards, which involves, among other things basic email.
Unfortunately, I don't think this case stands a chance. But I do hope that it hurts AOL's pristine image, and more importantly that it causes AOL to change their ways. They are fast replacing MS as the corporate evil of the computer world, IMHO. In fact, I think they will be much more dangerous than MS ever was. AOL has the ability to disrupt the operations of entire companies with it's network connections. There is little or no privacy with any material that passes their their gates. If they are successful in stomping out non subsidized ISP's ( corporate, university, phone, etc ) over the next 10 years then they will have unprecidented control over information. Something MS never really had.
First, It amazes me the maturity level one finds on public discussion boards. I am very grateful for the moderation system, so that, in general, I do not have to be subjected to this. But on topics such as this, I feel compelled to actually read everything. How people's opinion of Tom Christianson relates to this escapes me. He has contributed considerable intelligent information that can only help the Open Source community. I'd like to present the idea of evolution, if I may. A man evolves by growing more nimble fingers, a larger brain with greater capacity for memory, problem solving, etc. He evolves by being able to withstand harsh cold ( or hot ) environments. Often times, man finds himself with strange and seemingly useless biology: An appendix, toes, back hair, various skin tones and shapes, etc. Life, on the other hand, evolves through natural selection of species. When one species can no longer adapt ( analogously finding a local minima that they can not surmount ), it is replaced by others. The same holds true for our silicone world. A technology is created. The environment changes, and the technology must adapt or die. Death involves reinventing the wheel with our new life form; starting from scratch on the technical design, AND convincing a target audience to support the new technology. Thus, just like in nature, the predisposition of a technology to adapt is greater than it is to die off and be replaced by something new. As a case example, take Intel's x86 ISA. A venerable and intelligently adapted system. Sure, the original draft of x86 didn't take into account 32 bit memory, pipelining, n-way parallel instruction execution, etc. Yet that same language has been adapted to systems today that are among the fastest and cheapest in the world. This of course, was not to the merit of the appauling instruction set, but to the adaptiveness / inventiveness of their designers who knew how to maintain their market. Before shreding the idea of the x86 ISA, try, yourself, to invent a hardware language today that can cleanly adapt to technologies 10 years from now. ( neural message passing systems possibly? ) It's really hard to do. The only real way is to have enough of an influence to periodically reinvent yourself in an incompatible way ( as Apple or SUN was able to do. Though their lack of a perfectly competative market reduces their value to the computing society as a whole. ) Perl, likewise could not have have forseen all the uses that it has today ( and therefore produced a language suitable for everything ). Written first as a scripting language, then later as a full general purpose language. Never being fully without bugs, constantly reinventing itself, yet keeping backward compatibility as best possible. Still later, we add certain modern programming techniques ( again without breaking compatibility ) such as OOP, and threading. Threading isn't fully stable, and many important packages ( such as DBI ) warn against using it, while OOP is something of a perversion of classical approach. It is difficult and inconvinient to consistently define class objects. When dealing with OO, I much prefer java ( and to some small degree, Python ) since these were created after the OO revolution ( or skirmish, if you prefer ). Still, perl seemed better apt to handle these two revolutions than C ( which isn't pleasent for writing threading code, and had to be mangled in order to handle OOP ). Without breaking backward [symantic] compatibility, perl has managed to encorporate many exciting modern features. It's nature has even allowed it to adapt to web server-side embedded scripting ( in the form of ePerl, embPerl, mason, etc ). My focus is not that perl has conqured the world, but that it has survived the changing world. It has its scars and strange looking unused appendages, but for it's intended purpose, it provided zero down time through the evolving years ( referring to porting time ). People that buy a new linux distribution could still use their old perl scripts as is. But additionally taking advantage of the newer features in the new release. Much like one can do with the x86 architecture. If we take a look at java for a moment, we'll see yet another example of the uglyness of evolution. Originally concieved of as a fresh new concept taking advantage of the state of the art in computer science ( If you'll forgive the proposition ) we're already in our 3'rd revision, which has made heavy use of the term deprecation. First we do it one way, then decide that there really is a better way of doing it. But now we've discovered a more elegant approach to event handling and change so we change things yet again. Each time, it was essential to maintain backward compatibility ( this is still a fledgling system, and supporters would not accept such unreliable lack of support with legacy applications as they might with the conversion from SunOS to Solaris, or mc6800 -> powerPC ). Java is in the same boat as perl and x86 in that as long as the language evolves, it will be plauged with legacy features. Thankfully I can think of little that is hurt by the existing evolutionary steps in Java, but it is still much younger than perl or x86. Give it a decade and we'll see how well it fairs against new and shining languages. And some will support Java even though it is considered archaic and too difficult to keep up with the new boys. Or stead-fast C coders will mock Java, saying C was the true language all along ( as some humorously still do about assembly today ). The key is really whether a life form ( or a technology ) has outlived it's usefullness. Java obviously hasn't. It's a good programming model ( with far fewer points for error than C++, with what I'd call, similar support for the technology ). ( Notice I said model, since, obviously the execution time is subject to debate ). C is going to be around for as long as UNIX. Assembly as long as the discrete instruction model exists ( I'm all for neural messaging systems personally. Interpret your code directly without compilation!!;-) ). The x86 is still alive and kicking, simply because the technology is such that the underlying negative effects for a non-optimal instruction set are offset by pipelining and caching ( as in the case of AMD's micro-op cache ). To boot, this performance increase in changing the ISA is miniscule compared to the inconvinience of developers supporting a new platform. The added cost of compatibility translation hardware is offset by the massive economies of scale. As far as I'm concerned, we were taught a long time ago to not program in assembly when you can program in C, ( and consequently, not program in C if you can use something more aptly suited to the task ). Thus the processor is a black box, and I could care less what instruction set it uses. Give me performance / cost ( and if I'm in a laptop, additionally devide that by power consuption ). So what of perl? Originally designed as a reporting language ( before the days of the Dynamic web ) it now finds itself primarily as the glue between c libraries. To my knowledge, perl was not designed as the creator or the repository of information, but for the "extraction", and "reporting" of that information. Thus the DBI hook to mysql and the CGI front end allows it to format the data found in so many of our web sites. But this specific functionality has been built anew by php. Doesn't the [cached] CGI method of c-like logic control structures seem outdated in the modern world of table generation and data reporting. There have been no new op codes in perl to conviniently generate an HTML table in perl ( though embPerl and the like do so nicely but at a much higher level ( and thus less efficiently ) ). Thus has perl been replaced by php or the like in the realm of simple DHTML DB reporting? And since HTML is a far more common report medium than vterminals, has the underlying theme of perl been superceeded? I'll leave this as an open item. Whatever perl's role in the web, it still has roles in System programming. I find perl much more powerful and pleasant to program in than sed, awk, or even bash. If you need to write a filter or data converter from one database to another, I doubt that you're going to use php, and possibly not even C ( unless you really don't know any others ). If you want to provide a summary of slack space used on your hard drive ( sorting by directories, then by files, etc ), I'd find a hard time doing this in any other language than perl. ( Not impossible mind you.. ) Also I find perl very useful as a configuration script or simple socket daemon. Additionally, I can take advantage of many libraries that have perl hooks into them. OpenGL, XML parsers, TK, QT, Gnome, OLE, NT registry, socketing, etc. I can take advantage of these without having to learn all the intracacies of each language. I can play with new technologies found on CPAN, and decide early on if they're worth persuiing. The late binding and non-strict type checking really aid in such a transparent interface. Peronsonally, I find these times exciting. Who will win out in the language wars? Of course there will be more than one winner, because there is more than one target audience. As each new computing need arises, existing languages will adapt, and new tailored languages will appear. It is this trade off that the devoloper has to make. Personally, I like the advantage of perl's wealth of c-library compatibility, so I write many things in perl ( and even write emulators of HTML templating, or even the FastCGI / java-serverlet model to suite my own needs, just so that I can make use of certain libraries that are not readily availble in php or Java ) It also means I can have a common library package for both configuration / database manipulation and CGI scripting. I like the open nature of perl ( hard to compile away into hidden libraries ), I like the "more than one way to do it" methodology. I like that I _can_ write strongly typed modules ( through perl 5.005's "my Foo $var = new Foo;" and the "-w" and "use strict" directives ), but can also write "perl -pe '$x++ if/^A/; END {print $x }' " and other such one line programs that I can test the concept of in less than 10 seconds. ( though the obove seems clearer with 'egrep "^A" | wc ' ). When solving such specific pre-defined dillemas, perl excels. I should hope that I have to say very little about the most important contribution of perl; it's regular expression library. Highly optimized C ( depending on which version of perl you're using ), has managed to take an NFA construct and extend both it's efficiency and feature set. I have never seen such an extensive reg-ex package in all my years of programming. ( I would definitely be interested to learn of more powerful ones if any of you is able to provide me with such info ). php has seen fit to include perl reg-ex's, but I'm not sure how up to date, or how completely compatible it is. To be able to mathematically speak to your computer in such a way as to have it almost understand what you are thinking ( in terms of syntactic identification and manipulation ) never cieses to bring a smile to my face. Having the ability to intelligently merge a group of HTML files and into a single one with 3 lines of reg-ex code expresses such a mathematic bueaty I can not properly describe it. It isn't about doing it in fewer lines of code, it's about commanding power over symantics. Of course it seems cryptic to the uninitiated. So does calculus. Is that any excuse for shunning the language? Neuton invented Calculus because it was too tedius ( and in fact impossible ) to correctly describe bodies in motion without it. Yes I can write an optimized and easy to understand filter or parser in simplistic C constructs, but reg-ex allows you to attack the problem from a completely different and extensible approach. I've regularly written entire parsers in perl ( memory management alone eases the pain of writing a parser ) as a simple OO module that gets loaded into a larger program. Meta language is therefore much more obtainable in a smaller development group. Even today, perl's reg-ex is being extended with parenthesis matching, and the refinement of a procedural approach. So long as perl contains the most powerful and unmatched reg-ex library, I will not lose use for it. But meanwhile, nay-sayers will scoff at perl code littered with such "unreadible" expressions, happily ignorant of the calculus of the programming world.
Lastly, on the topic of a first time language ( which seems closest to the original topic ). Perl is great in that it allows the most straightforward possible program. perl -e 'print "Hello World"'( as with BASIC ). You can even enter the debugger and type this in manually ( again, a la BASIC ). You can create an initialized variable the first time you use it. $x = 2 + 3; print $x;
Essentially, I'm claiming that perl is every bit as good as BASIC as a first time language. Moreover, it's better than the old GWBasic, Apple Basic or TI Basic that I first learned on, in that we're not dealing with antiquted line numbers, or goto's. Intead we have very nice flow control structures ( last, next, redo ). Aside from that, you could completely take a BASIC lesson plan and "port" it over to perl, with the advantage that the student now contains some real world experience ( as opposed to BASIC ). I doubt system programming is a good second step. Perhaps CGI's, though you really do need to have _some_ programming experience before you're allowed to crash your school's local web server.
Just for reference. I learned the
Basic -> Pascal -> Scheme \ C -> C++ -> System programming ( insert thousands of languages here including Java and Perl )
I kind of like this approach, but am basically questioning if Perl could replace Basic as I've said above. Real work should _not_ be performed until a C like language is well understood. But there is no reason you couldn't return to perl at the system level.
Well, this is long, and time is short, so I shall end you here. -Michael
Great Expectations and the great space barrier
on
On to Mars
·
· Score: 1
Ahh, my fellow comrades. The dreamers, the doers, the opinionated. We who love SCI Fi, we who dream of a free and unified world. We that look to the stars... Of course, not all of us have our head fully screwed in properly. Do it, damned the cost ( since money is of little value to us right? ) Even in a moneyless "Star Trek" society, there is a cost in everything: The human man hours of skilled and talented workers ( opportunity cost ), the physical resources, loss of life or general working conditions ( quality of life ). Communist society, for example, would have us all living more humbly for the "greater good" ( which of course is in the eye of the statesmen ).
In reaction to colonization of the moon, mars ( or even space stations ) as a supposed solution to the failing of NASA, I say, Action with no incentive is foolhardy indeed.
One reader pointed out that we can not hail NASA's ( and hence man's ) conquest of the moon, since it's incentive was, in fact, a life threatening cold war. Something greater, and more profound than man's perseverance and ingenuity, is _life's_ fight for survival. We did not conquer the moon because we could, but because we had to; when backed into the corner, this cat lashed out with all it's might ( and resources ). This devotion of resources is something that finds no such motivation today. And I would challenge anyone that craves the back-firing black magic that brings back such motivation.
What we are left with are millions of dreamers pointing fingers saying, "but you did it before". And a government agency doing what I assume is their best to efficiently convert their scare resources into visible science.
As for a moon base. I'd like to point out that we have YET to successfully produce a bio-dome HERE ON EARTH! We do not yet fully understand the delicate balance of nature, less how to command it. Colonization, could therefore not be fully sustained by these currently lifeless worlds. The cost would be unimaginable. And of course we are talking about residences, which of course would have to be regulated. I'm sure the ACLU and other rights organizations would have a field day with many of the near-death experiences. If you think the set backs in the simple non-permanent space station are appalling...
As was also pointed out, we have little human critical experience with the extra-terrestrial. How do you mend broken bones and other ailments? Do we send groups of people within the next decade to their doom? Do we risk the life of our most talent and promising astronaughts?
I believe it to be naive to say that we're wasting money on these conservative, yet expensive projects, so we should go the full gambit and colonize! There are several orders of magnitude of complexity, risk and COST!
I would call this the brute force solution ( much like the lunar landing ). I have always idealized the intelligent, dynamic solution; one that isn't fully apparent at the outset. Do what is best at the moment ( hopefully as adaptable as possible ), and recheck yourself regularly to see if changes can be made to either get you further ahead, or ( as is often necessary ), to reinvent yourself.
I believe the "commercialization" of space is a good thing. The private sector ( and our potential entrepreneurial pioneers ) are welcome to build upon off-the-shelf rocket science and find some personal goal ( which just happens to further mankind ). Zero gravity Space products, and possibly lunar / mars mining are interesting incentives. The main thing that I like about them is that their motivation would be "economics", ( which doesn't mean money, btw ). You do what is most efficient at the moment. When things look bad, you kill it with little after-thought ( don't throw good money at bad ). When someone seems to be succeeding, other venture capitalists will jump on the band-wagon, thus expanding our knowledge and experience base. When dead-ends are reached, people will look elsewhere.
Since there doesn't seem to be enough vested interest, this will take time. The entry barrier is still too high for this model. But the incentives and the means are becoming ever more apparent. This is my personal hope for the future of space-technology.
My big issue is that government agencies do not follow the model of [economic] efficiency, but instead follow task oriented mandates. "Go to the moon", "stop crime", "Make hubble work". Unfortunately, their support ends with the completion of a given task. There is no competition, no reward system, no punishment ( other than general monetary prioritization and national pride ). Either a private sector NASA needs to come into existence ( possibly as with the post office ), or we'll have to continue to live with this task-mastering until it becomes economically feasible to surmount this "great space barrier".
In the short run, I support a congressional review of NASA. It may bring to light the negative effects of cutting costs by reducing redundancy. It may shed light on administrative weaknesses. It may very well bring about a more efficient and productive NASA. Sadly, it will most likely involve cuts in funding. But as such, I do not personally find value in mars exploration ( gasp, blasphemy! ). If there is life, it will still be there in a hundred years. Even if not, it would only be one of the billions of interesting facts that we've lost. The Matrix: You humans seem to define your reality by your misery. Me: That misery becomes our newfound motivation. Without it, we become complacent and wasteful.
Maybe I'm missing something here, but if we're so behind on outer space technology and "breaking low earth orbit", then how do we send out all those mars, venis, and solar-system probes. Yes they're small, but the point is that the technology is there. We simply have learned more intelligent methods than the brute force F=ma.
Just through I'd give my opinion on the matter after reading as much as I could get my hands on.
Pros: Allows me to communicate with my grandparents, or any known non-techno savy people. Allows me to contact relatives I don't know the email of, or to people I haven't contacted in years and can not be sure of their email address( and don't feel like making a phone call ). What I find interesting is that the USPS is probably in a better situation to do this than commerical companies, since they already have access to the address of every man-woman and child, so to speak.
Cons: If not done properly, it can lead to spam, unwanted charges, or wasted use of paper. It is essential that they do it right the first time. Even worse would be the exploits that we ( or their committee ) don't think of. You now have a new form of a publicly viewable Social Security Number. It's only a matter of time before companies REQUIRE your private / personal email address for services.. Simple string parsing would validate the request. Now, just like the Pentium III Serial ID, or your social, malicious companies could exploit known shared info. Possibly credit history would be attached, medical records, etc. In fact, all SSN info could be mirrored by your USPS ID. Heck, some sites would probably try and link SSN to USPS ID. The solution to this, I believe is to not allow a per-person email address, even though this doesn't fully eliminate the problem ( now we just track house-holds, instead of individuals.. But it's a house-hold that buys a product, so it's still valuable info ). Additionally, since USPS is federally subsidized, any significant innefficiencies are partially passed along to the people. It is important that this not become a huge multi-million/billion dollar flop which requires federal bailout. Thankfully I doubt the system would allow a perpetually innefficient system to exist.
Proposals: Optionally allow mail forwarding to existing address. Their site would simply provide a consistent address. This minimizes their cost, since they wouldn't have to store the mail long term if we didn't regularly log in. Also doesn't require us to have yet another email account to check daily / weekly. As with USPS mail, the sender is billed. This alleviates much junk mail / spam. The downside to this is in auto-reply email, where a person registers with a web site, then puts their USPS-email address which for some reason is mapped for printout. Now my poor free web site is being charged for many "potential customers". Additionally, it provides for a seriously expensive DOS attack. You now have the ability to rake up millions of dollars worth of USPS bills if a target email-responder site is repeatedly hit. The solution, in my opinion is to set up a billable account with USPS, and then potentially billable email would have to be authenticated and authorized ( just as in any e-commerce transaction ). The default auto-response web-site would obviously not provide a mechanism to send billable email. Unfortuntaely, this would either require a client side program ( possibly in java ), or to make use of CGI's that require either cut-paste, or browser-file-uploading. None of these are ideal, since they don't make full use of your existing email programs. Another method would be to simply send the email, then if it requires payment, a notification is returned, requiring you to log onto their web site and authorize the transaction.. Unfortunately, this makes it easier to spam, since everyone can be mailed, and the payment-based transactions would simply be ignored.
In order to alleviate spam, the central site could possibly monitor mail volume, and automatically charge accounts that exceed a certain volume, with the notion that spam/ junk mail is the intention. High volume is expensive for the central web site in any case ( due to excessive local storage, etc ). Another thing I like about this, is that it minimizes chain mail, since you'd be addressing dozens, or hundreds of people regularly. I don't consider this stiffling of free communication, since you'd still have your other email addresses to use for such time-wasting things. I'm not a big fan of email-based mailing lists anyway. That's what bulliten boards are for. If it's supposed to be daily, then they can regularly check the bulliten board, with little or no excuse.
I definately think this issue deserves attention, since there is a lot at stake; Our privacy, financial obligations( for both sender, receiver and USPS ), and our dear forests. I do not, however believe that ostrige-like-fear should hamper progress.
I believe there is a certain security issue with this. I remember in days of old, certain stores acted as a local post office, but robbing such a place could also be considered a federal offence. I am under the impression that post offices today are isolated, partially to prevent it from being a direct target of crime( as would be the case in a grocery store ).
I'm sure it's possible to reside in a mall, but I get the feeling that there's a good reason to not. -Michael
- CNN was obviously erroneous about the Patmos using AMD K6's instead of AMD Athalons. -- 200MHZ bus and the words cheap obviously seemed like a K6 to the semi-technically knowledgable author. In fact, ( as has been pointed out here several times ), the only 200MHZ bus available for a PC is the Athalon. -- They mentioned Patmos reaching 1GHZ w/in the year. The K6 does not have the potential to make this speed, in fact, few processors in the world other than the Athalon are currently capable of reaching this goal within a year. Especially without requiring a motherboard; which obviously would be bad for Patmos, since I assume they provide _some_ form of custimization of their motherboard. Or at the very least have carefully selected a board, and would not think kindly of choosing a new board so soon after their initial product release.
- The unqualified use of the word super-computer.
I've noticed several posts about people thinking that they could design a "super computer" even cheaper than Patmos. But really, all they could achieve is a "theoretically fast" machine. A supercomputer is the sum of all it's parts, and therefore the weakest link can break the chain.
As a disclaimer, I am not formally trained in super-computer concepts, but much of this is based on my experience and common sence ( which may differ from horse to horse ).
A super-computer must have top tier performance ( obviously ), must have data-integrity ( you don't spend half a mil just to have a core dump, or system freeze ), reliability ( 24 / 7 uptime while performing it's work ). It should also be scalable ( grows with your company or the task as is fiscally justifiable ).
--Simple points: When selling a super computer, you must choose high quality parts ( or at least make things n-way redundant ).
In my experience, IDE drives don't cut the mustard, due to their high volume, minimal quality price-focused nature. ( skimp on a heat-sink or shock obsorber to save 5 bucks per drive, etc ). When you buy IDE, you think disk space and low price. When you buy SCSI you think of performance and quality ( and usually expense ). Thus they are designed based on that marketing paradigm.
An IDE drive also uses an IDE controller, and is thus inherently sequential. A SCSI device can queue multiple independant requests, so as to perform disk geometry to determine an optimal seek path. Additionally, due to the paradigm above, more cache and higher rotatial speeds are applied to SCSI devices ( not that they couldn't be applied to IDE.. but why should they? )
As for a network, some referred to HUBs and ethernet. Ether does _not_ scale well. Sure you can get a faster / more intelligent HUB, but you never achieve maximal theoretical bandwidth. I'm not completely sure of the network technology used here, but it seems to be peer to peer and bi-direction ( to facility rapid acq's ).
--Memory. This is really the key to a good super-computer design. SGI made use of wide 256 bit multi-ported memory busses with interleaved memory ( 16 consecutive bytes was segmented across 4 memory controllers, thus linear reads were faster, AND independent concurrent accesses had a statistical speed-up ). Of course a 256 memory BUS is expensive, especially in a multi-CPU configuration. SUN's starfire, for example, has up to 64 fully interconnected CPU's ( don't recall the BUS width ). This required a humongous backplane with added cost.
AFAIK, the Athalon uses regular SDRAM ( and a cost effective solution would have made use of off-the-shelf parts ). SDRAM is nicer than older PC-based memory in that it's pipelining allowed multiple concurrent memory-row access within a chip. Several memory addresses within the same row could be in the pipeline, and up to 2 rows could have elements in the pipeline. This is a more sophisticated approach than interleaved memory, BUT, it introduces a longer / slower pipeline. RAMBUS furthers this concept by narrowing the BUS width and furthering segmentation. It allows greater concurrency, but latency ( and thus linear logical dependency ) is increased.
RAMBUS's theory of high latency, high concurrency benifits non-linear programming, such as Italium's ( Intel's Merced ) deep speculative memory prefetching, or ALPHA and SUN's multi-threaded processors ( where cache misses cause an internal thread-level rapid context switch, thus hiding the latency ). Existing x86 architectures, however can not fully take advantage of such concurrency, and the net effect is slower execution time for linearly dependent algorithms ( non-local/consistent branching, and non-parallelizable math calculations ). In this case, making use of high speed / low latency interleaved EDO ( as is / was done in several graphics boards ) seems a better alternative ( but hasn't come to pass in mainstream motherboards ).
--mutl-CPU. This is an interesting topic. Mutiple CPU's can connect to the same memory ( with large internal caches ), or can have a numa architecure with shared segmented memory ( isolated, with interconnecting buses ). Or they can be autonimous units connected via a network. There are pros and cons to each mechanism. The last requires the most redundant hardware ( which is actually good in terms of hot-plugability ), and has the slowest inter-CPU communication. It thus works well in message passing systems, as opposed to symmetric decomposition of large data arrays ( eg parallel vector processors ). Personally, I like the NUMA approach the best, but it requires proprietary hardware, and hot-plugability would have to take the form of a VME bus etc.
It would seem that the approach here is multi-CPU ( 2 or 4 ) to perform a single task. Concurrent threads are distributed across machines in a message passing system ( hopefully minimal data-sharing ). The AI controller probably handles messaging, arbitration, in addition to the advertised load balancing. The multiple CPU's on a board are most likely for redundancy. My guess is that 2 or 3 CPU's are used for user-thread redundancy and a 2 or 1 CPU's are dedicated for OS operations ( using spin loops in the user threads ). Thus minimal context switches are required, minimal memory bandwidth is used ( since only one virtual CPU is ever accessing memory at a time ( though 2 CPU's are simultaneously requesting that information ). They may actually allow the Linux scheduler to rotate proc's, but as I've learned, this isn't Linux's strong point. A single tasked CPU is a happy CPU.
I know SUN has optimizations for context switching ( keeping most pertinent info w/in the CPU, along with a unified virtual memory model, as opposed to the offset-based x86 virtual memory model ). Unfortunately this is offset by register window swapping, but such context-switching centric processors would allow for more efficient concurrent operations such as multi-threaded apps ( such as java. Before you laugh, one application of this low-cost supercomputer is web servers.. And serverlets are an emerging technology, people will ask the question, how can I make this existing code run faster in a short period of time ).
-Concept. ASICs / FPGA's. SGI had the concept of making a simple, cheap, reliable, and fast logic CPU, then couple it with extremely proprietary logic / processing chips that offset the code logic. Combinational logic is faster than any sequential logic, though much more prone to bugs, and higher production costs. High performance reprogrammable FPGA's could help the industry, since the hardware could maintain a high volume, low cost ( as with current CPU's ). Thus you could make PCI / AGP expansion boards that handle load-balancing, message passing, java-extensions, OS-operations etc.
I'm sure their AI logic is done similarly, but it's a completely seperate box, my thought would be that the "boxen" would have these expansion boards, and the customer could request optimizers for say, the web, or weather calculation, chess designs, what-have-you. The goal being that these expansion boards become as common as modern graphics accelerators, modems, sound cards etc, without having to go through all the trouble of designing the hardware of those boards.
-Michael
Re:Interesting but not well enough thought out
on
The Regulon
·
· Score: 2
One further point. My initial statement was that exponentiality is only bad in a scarsity.. A good example is math.. I can write a seemingly infinitely large number... But in reality I have a scarcity of paper or time to write it.. But I can be creative enough to represent astronomical numbers efficiently. Arguably, those that sought to find pi with inifinite precision in days of old were challenged with this feat. But they too did not consume 100% paper, or skip sleep or their wives..
I hate when humans are compared to a virus because all a virus does is eat and reproduce. We have art and inspiration.. So long as we don't become mindless consumer zombies, we'll have an edge on exponentiality.
I think the original author's point was that was the danger. But myself and my children will not be plagued with this; not because I'll keep the TV and internet off, but because I'll excite them with the many physical wonders of the world as my father did with me.
-Michael
Interesting but not well enough thought out
on
The Regulon
·
· Score: 2
exponentiality only plays a factor when there is or can be a scarcity. There is only so much land, sunlight, soil, free O2, etc. So mold growth, viral replication, etc when grown unchecked will quickly consume all resources and start dying out and rapidly.
Information consumes several resources. First and fore-most it consumes time. If "media" grew to consume 100% of time, then we'd have a recession and then a depression, and thus a death of society; course humans are more pro-active then a virus. We'd make laws, etc when enough people understood the danger.
Another resource consumed is disk space. You can't store every piece of information ever gathered by anyone. The ideal would be to have an MP3 recorder on and with you 24/7 so as to reference anything you've ever heard. Likewise with a video camera.. Beyond that, you could sense and record everything in the world 24/7 - thereby approaching omnipotence on a local scale. But the storage of this information would be impossible. Typically we record onto ferrite, and there's obviously a scarcity of metal in the world. Other's are using phosphors or plastics, but they too would be a scarcity.
So what happens in the "hard drive revolution"? Well, we use metal for quite a bit, and so to take metal away from one form of consumption produces competition and weighing of values. Look at DVD-ROM drives and cell phones.. They share a common component (don't know what, but I've read it often enough on slashdot). Because cell phones are 'exponentially growing', DVD-ROM's are prohibitively expensive. Manufacturers of course try to get around it by making them faster and thus justifying the added cost.
The danger again is the value storage above all things. So we lose the ability to use the materials for other purposes, and again humanity becomes pro-active and regulates. In reality, we have a diminishing marginal enjoyment. We're not going to record and playback information 24/7; we're going to be highly interested in other forms of recreation, or in rare cases; work.:)
As other's have pointed out, there is also the prospect of human memory.. Our brain is a finite storage system. To my understanding we "memorize" relations between experiences. We comprehend a sound, and relate it to some higher level concept (such as a vocabulary, or fear, etc). Most of the experiences we have are lost, and only the gist of what went on is retained - some more than others. Beyond that we'll have a fleeting interest in all the info out there. Some would love to have ready and repetitous access to espn channels 50 through 3,000. Other's would be interetsed in the discovery series 50 through 100, etc. So different personalities are going to filter that info uniquely.
Next is the issue of information predators. Well, first and foremost is the minimal resource requirements of information. Second is the reusibility of most of these resources (except time and energy consumption). You can only eat a plankton once, but you can record different MP3's to the same spot on a hard drive over time. The real predators are interest and time. You may be facinating today and thus over-write older material, but tomorrow you're old news; bye bye.. Garbage In, Garbage Out.
Next comes Darwin with Natural Selection. Information that gets discarded quickly is selected for deletion; they'll have their moment.. Mutated from similar topics of the past. Copied web sites, re-reporting news on CNN, etc. Each time they're a tiny bit different, and they just might have found the right audience; maybe Bill Mawr had a comical spin to the same info that made a certain class suddenly interested in it. But ultimately, eveything has it's time to die. And information can die at a much more massive rate than life; just like an over-population of viri in a host, once the host dies. Entire hard drives can be wiped; entire broad-casts can be over, without ever recording them.
The saturation point is where we max out all possible bandwidth. Our entire visual perifera is filled with mini windows of streaming content. Any additional information beyond this immediately dies because of lack of additional audience. Even if they are recorded, there is little chance that every moment will be revisited until a new baby is born or some other content dies. But now we're back to postulate 1 above that we're spending 100% time in information consumption. So my argument is now concluded.
Ethics is a highly subjective and volital thing, but is none the less an essential part of science / engineering.
My personal value system suggests that manipulation ( in all its forms ) of single celled life falls in yellow catagory which means acceptable, but with precautions. Research such as this should be public knowledge ( genetic war-fare is a potential down-side )
Personally, I don't think that there is anything special about life. We are simply bueatiful machines crafted in an elegant fashion. This is completely independant of whether we were created by an intelligent being or not. Just as we seek to explore the physical world around us, why not seek to better understand the most basic workings of our being. If you are relgious and are having issues with "defining-life at a genetic level", then think of it as simply better appretiating "God's work".
As a point of reflection, Judao-Christian Dogma states that "Thou shalt have no other Gods before me". If God can be defined ( at least in part ) as the creator, then surely mans ability to manipulate atoms and now create life places us dangerously close to defiance of that tenant.
I'm one of those egnostic types, somewhere between atheism and religious so I like to straddle the fence ( It's more interesting than only taking one side ).
I remember the old days of BBS's. We'd have monthly get-togethers that included pizza. We were all a big group of friends; the younger crowd talking about games, the older crowd talking about hardware all in the same room. Sure enough, the feel of community is removed when you're 10 thousand miles away from everyone.
When I went to a University, I actually found a nice replacement for BBS's in the campus central UNIX machines. I'd log in and track who else was on. I could be productive, while chatting with 3 or 4 other people. There were bulliten boards, download sites. Before the web was popular, we'd play with our.plan and.profile's. We had a sence of our own little space that we could decorate and other friends come come and visit.
Around that time muds/mush's/mucs were popular. There were tales of people failing out of college playing muds 18 / 7. Those muds allowed chatting in a pseudo IRC fashion, but just as on the UNIX machines, we had our own little piece of the virtual world that we could decorate and share with others.
The combination of these provided ample substitute for me in the mid 90's. BBS's were still popular ( and I activately sought them out ), but as FTP sites fullfilled my download cravings, I had removed the last of the BBS advantages.
Role playing, chatting ( with local people ), files, information, home decoration...
Now AOL's Instant Messenger, ICQ, and Yahoo pager are my primary chat resources ( localized apps are much faster and virsitile than an IP packet per character ). My Web page is all the person information I need to keep. The web itself handles all my files. And I've been too busy to seek out those MUD's of old.
Actually, another good thing about the BBS's was that you couldn't be a junky. You only had 2 hours a day, and all your games were turn based, which limited so many actions per day. There is no reason why tradewars can't be loaded onto a telnet compatible server and still use the turn based mentality. I would hope that it's already been done.
I have 4/5 computers at home that perform various functions. My main computer has dual video and is my gaming / DVD / internet machine ( win98 sadly ). For this I have a dedicated high quality monitor / keyboard. But I also have a second video card in it connected to a second, lesser quality switched monitor.
Because several of my other machines run in Text mode ( servers / burners ), the video quality was not an issue for me. However I still do 1024 X and have a second NT / 98 box ( for networked games ). So I have all high quality vid cables ( at my local computer fair, they cost me $10 as opposed to the $4 for cheaper VGA ). The switch box does introduce noticable distortion, but only for very small text. It's been fine for Linux development ( so long as I choose my font / color scheme appropriately ).
My switch box cost me $25 at the fair ( 4-way KVM ). I run into the problem of having to switch over in order to boot the machines, and on some boxes I need to "gpm -k ; gpm -t ps2" to reactivate my mouse after switching to other boxes. ( anybody know if there's a better way? )
I seriously doubt this cheap solution is any good for very high res monitors. It only works if you can dedicate one monitor / machine for your main work, and just have others for occasional use ( where loss of quality is acceptible )
Still, I like toys, and am willing to sacrafice quality over quantity of toys.:)
If you have the money though, $100 active boxes are your best bet. Personally I don't find value in the KB key capturing for switching ( unless you really do it often ). ( In my mind, the greatest value of switch boxes are to reduce the cost of monitors.. Buying a thousand dollar box ( as I think I've seen suggested here ) really seems to push it ).
Alternatively, getting eXceed ( or the free MixServer ) for windows, or VNC for Windows / Linux will allow you to consolidate displays.
Slashdot Mirror
I'd be curious to learn more.
There are so many factors to performance it's not funny. You have the classical memory BUS speed, the periferal bus speed ( if you're overclocking, or have a strange multiplier ), you have your MB chipset buffering scheme ( how many delinquent requests can it keep going ). You have the performance of the chipset itself. Of course you have the huge variable of different periferals strenths and weaknesses which pull you every which way but sunday. Then of course you have hardware drivers which tend to be optimized for specific processors ( the norm being intel, but in some circumstances, like AMD's 3DNow for Voodoo's Glide ).
Then inside / around the CPU you have cache, which is a HUGE variable. You have the raw MHZ speed, you have the pipeline depth, and the latency, both of which are negatively affected by larger caches ( due to address resolution logic ). Then you have the port deth ( how many parallel accesses can the cache access ). And finally the size and bandwidth of the cache. AMD / Cyrix have gone with bigger but lower performance caches, while Intel has gone with more complex but smaller caches. Hypothetically, a larger, simpler cache will be cheaper to design, but will take up more surface area, and thus provide lower yield. To make matters worse, some programs require high speed access to a very small data-set ( and thus benifit Intel ), while other applications just use a lot of data, and anything that minimizes main memory access boosts speed. I believe Quake qualifies for the former, while Office apps ( and scripting languages in general ) benifit the latter.
AMD and Cyrix also, for a while there, worked at enhancing the internal instruction flow algorithms. Making huge branch prediction buffers, and in the case of Cyrix, producing all sorts of algorithmic optimizations that Intel strangely didn't implement.
I believe the main reason AMD and Cyrix didn't work as hard at their FPU was because it's a _really_ ugly design project. It's more fun to work on general purpose flow design and playing and tweaking a simplistic cache design, than to get dirty with all the possible combinations of floating point logic ( especially one as ugly as the 8087 family. I believe Intel owns several patents on some highly optimized implementations, so the others would have to devote some big bucks to tweak theirs without violating any laws. Not to mention, making it faster often times means taking up more silicon. Thus you have a larger die ( thus lowering yields ) and the logic is expensive to design / debug / implement to boot.
The next issue was latency. Intel, with the 80686 line ( I hate their non-informative naming conventions ), went super-pipelined, which worked great for sequential operations, but performed horribly in random branching contexts. AMD and Cyrix both opted for a narrow pipe-depth, with an emphasis on branch prediction. Thus even failed predictions had minimal penalty.
The fastest possible processor will be non-pipelined and have n-wide execution components. The reason being that each pipeline stage introduces a store and forward delay. Some stages may perform minimal operations, thus wasting 75% of a clock tick. This really hurts data-dependancy delays, since a pipelined FPU might take 15-150% longer to complete a Divide which the very next instruction requires. If all other data-dependacy paths are blocked, all the pipelining in the world won't do you any good. In the integer world, this is very common. I'm about to perform a cache missed memory fetch, but first I must calculate the address. If every other instruction is based on the contents of that memory cell, then pipelining can only hurt this particular case.
The biggest opponent to complex and optimized operations was that they would slow down the rest of the processor ( by requiring slower clock ticks ). But the device manufacturers are learning how to make different parts of the CPU run at different frequencies. ( They've long since learned how to run the BUS at a fraction of the Core ). Intel's next 80686 processor varient will have a clock doubled integer core, for example.
Still, the main reason we don't see a return to complex optimizations is that having 32 ADD components is extremly more expensive than having 2 16 deep add components. Even though you'll get a significant performance boost ( assuming you can manage that huge bus, and a potentially large number of register ports ), you probably won't make up for the added expense in shere complexity and yeild loss ( due to extra size ).
Thus, Intel went for a partially pipelined FPU which had heavy latency penalties, but improved overall operations ( especially for non data-dependant operations ).
AMD Finally headed this off by making multiple independant and fully pipelined FPU's in their Athalon. ( they spent the extra bucks to remove many of the stalling conditions caused by sharing of resources by seperate components ).
Personally I like SUN's java-multi-threaded CPU concept ( even if it never succeeds ). Basically, you have 4 parallel fully functional, non-related, non-pipelined, fully optimized functional units. There are no resource contention issues, no scheduling problems, a simplified logic design. And it's cheaper because you take away pipelining. The best part is that each of these extrememly simple components are just cookie cuts. You spend all your time tweaking the hell out of one tiny unit, then make 32 copies. Almost as easy as cache design.
I believe the Crusoe could learn from this. They already have their simplified design, they could take it a step further. Say, keep a single CPU implementation for power-critical devices. Then replicate that core 8, 16 or 32 times for a desk-top varient. Since you can control your wrapper code, you can determine what is the optimal CPU-width. I'm sure there are many cases that would allow you to submit 32 parallel instructions ( at least for the compiler ).
I agree that mathematical algorithms are slower in perl. I too have written nested loops of matrix manipulations, and they were inexcusibly slow. Even when you specifiy integer only, and use for( 0..100) instead of for( $i = 0 ; $i ( notice my aversion towards the bastard word "proof" ).
Perl is glue, and should be treated as such. You can't write big perl programs without having an intimiate understanding of C ( since you're bound to call a great many c-library functions, and you really should consider writing stuff in C in tandem ).
In this respect, perl has the potential to be faster than Java for certain classes of projects. In general, an all native project will run faster in java IF, the delay in JIT ( or VM loading/ initialization for that matter ) compilation is marginalized away by either a sufficiently large computationally intensive core, or through persistent operation ( serverlet, etc ). Otherwise, gcc optimizations are bound to make better work of the perl-c libraries, than JIT will on the java-byte-code.
In my experience, the load time of java applications is very frustrating. Perl has very little to initialize, and is single processed at that, so will run faster than java-native-threads for simple applications. Plus I rarely find a perl-daemon eating up 100% of my CPU time, which I have regularly found in java ( at least netscape's version of it ). As a disclaimer, these are _my_ experiences with Java, and I do not claim this to be the general case. I'm sure bad code design had much to do with it.
* $bar = $x->[ 2 ]->{ e }->[ 1 ]; # yields 'g'
Even easier:
$bar = $x->[ 2 ]{ e }[ 1 ];
*Finally, consider this python statement: x=['a','b',{'c':'d','e':['f','g']},['h',['i','j']] ].
Try
$r_x = [ 'a', 'b', { c => 'd', e => [ 'f', 'g' ] }, [ 'h', [ 'i', 'j' ] ] ]
What's the problem????
On OOP, I agree that Perl's backwardly compatible method is akin to the P-III's x86 mess. Still with Perl5.005, you are capable of performing most OOP operations.
namely:
the pseudo-hash's ability to stricly enforce hash contents, and to provide private class variables.
local var strict Data-typing based on class names. ( ie "my Base::FooClass $x = new Base::FooClass" )
Unfortunately, my main reason for using perl is the vast feature set ( CPAN, integration with Apache, support for QT, Gnome, GL, etc ). If python were that supported, I would probably switch over in a heart beat. It is good to reinvent yourself every now and again..
A lesser used purpose of perl for me is for quick and powerful command-lines or shell scripting. I completely believe Perl is best suited for this, and I do not see python as a replacement here. The idiocyncracies of perl really shine in a shell environment, and I believe this to be the major problem with most programmers.. Shell attributes do not regularly lend themselves to general purpose programmers, who then label perl as cryptic.
To fully appretiate perl, you must reguarly perform such system administrative functions.
* These comments powered by Printf
:)
Well, since this is perl, most likely it's powered by write.
As with other comments I've read. Lack of IP would be akin to Communism. The discoverer or, more importantly, the entrepeneur should have a right to command that which they create ( at least for a period of time ). I would rather that we lead by example than by force. Let something be free because the creater wills it. Otherwise we have anarchy ( If you won't free that bone with meat on it so that the community can share, then we'll do it for you. Never mind that you almost died trying to acquire it )
And also as was written in this thread, the GPL is a form of IP ( albeit, so that others can take it form you and then sue YOU for using it ). But this is perfectly possible when communism goes awry.
-Michael
p.s. Sorry if I'm sounding political.. it's not my intent.
I'll have to disagree with you here. I've done some FPGA programming myself, and it currently can't even begin to compete against commercial high end processors. Granted with the super-RISC-like nature of the Crusoe is better suited for this, and I really am pulling for the FPGA revolution, but there are several issues:
1. Production Volume. In order to be cost effective ( less than a grand a chip, more than 10 manufactured a day ), you're going to need a serious chunk of fab hardware. ( which seperates itself from the near-zero mass reproduction cost of Software )
2. Performance. This is only going to fly if it can have an acceptible performance. FPGA is too unoptimized ( not familiar enough with ASICs though I know them to be faster, though not reprogrammable ).
3. Size. For FPGA to have the same amount of logic, it requires an incredible number of unused wires and gates. The cost should be minimal because we're dealing with mirrored arrays ( like RAM ), but size and power consumption will be significantly greater for an equivalent chip.
4. Power consuption. One of the Crusoe's key design targets was power consuption ( since they knew they couldn't make the fastest chip around, they found the best balanace of technologies ). FPGA and I assume ASICs would not allow them such a level of efficiency.
In short, yes you could increase the market ( and thus lower the cost ) of straight FPGA, then provide open source logic to do something like Crusoe ( or even a purely custom-Linux Supported Core ), but it's not going to be competative.. And since you can purchase a $40 CPU with far greater power.. Why would you?
I don't want to stamp out innovation or the free market.. If you believe you can succeed, then by all means try. Heck, I'd love to be proven wrong.
Transmeta has spent a great deal on R&D, and therefore needs to recouperate both those costs, and that of outsourcing fabrication [to IBM?]. The code itself is intimately related to inner details of the processor. With the Crusoe being as simplistic as it is, published source code would greatly accelerate reverse engineering. Therefore the big boys could easily steal the idea with minimal development cost.
This is the idea of intellectual property. If someone else can completely know how you did something, then what is to stop them from doing it. Or as with the nuclear arms race between China and the US, their knowledge of how we did certain things helps them to not waste time making the same mistakes that we made.
As others pointed out, Hardware open source is a completely different beast than that of software. I can't just build the crusoe in my spare time as a recreation. I have to devote serious resources. And so I feel a bigger loss if some other industry steals my ideas, and makes their product and nobody buys mine.
Still, I think it would be nice to eventually have all Crusoe code public domain. They could benifit from the same sorts of optimizations that key software technologies have had ( e.g. with gcc or the core of Linux, where everyone has a vested interest in focusing ). But unfortunately, it becomes an economic issue. I doubt they benifit enough from Open source to outweigh the competative negatives.
I read the article, and I don't see how you can draw those conclusions. I interpreted this to mean that the brain was trying its best to adapt to critical conditions. When sleep deprived, various parts of the brain begin to shut down ( as a self defense mechanism, or simply because they cieze to function ). But in order to maximize the human capacity to think and survive, other more dormant portions of the brain are activated to take over certain operations. The specific operation sited in this article was linguistics.
:)
I don't really see why linguistics would be important for survival. They said math skills went down the toilet ( because it is such a complex and consuming activity that there weren't any reserve portions of the brain to divy activity out to ).
They specificly raise the health risks of sleep deprevation, and how much poorer those test subjects performed. What it does show, however, is the incredible adaptibility of the human mind. Much like Terminator 2, re-routing neural connections on the fly to redundant reserve systems.
What this also suggests to me is that we are capable of using other portions of our brain when we try hard enough.. Perhaps we can meditate and awaken those dormant sections. Who knows, maybe during finals, we can get a suddent burst of intelligence even.
-Michael
The spell checker was the single worst contributor to the modern written word.
A few points. First, never say never. Even now. :) I believe the major problem with QM is that it is based on probabalistic waves. An earlier poster spoke about the inability to completely describe a single particle, but that these probabalistic equations describe the ratio of temporal / spatial existences ( sorry, oddly worded; am too tired ). :)
My belief is that one day we will be intelligent enough to mathematically describe the practical reality of QM. Just as Newton had to invent a new math ( eg Calculus ) to describe his physical world. There may yet be more mathematical realms not yet understood.
I've been turned on to String theory, simply because they tease us with equations that almost look right. Then go on to describe how each decade they discover a new math which gives us greater insight. Whether or not it is ultimately true is hardly relavant so long as they can continually better their predictions of our physical world.
As for the statement about being boggled or not understanding. I've read that on many occasions. I believe it is a good litmus test for the present day. But I do not believe it will always hold true ( that whole never say never thing again
I have to disagree with you a bit here. It's very easy to learn yes and no. Even a dog or cat can learn it. It's all about associations ( pavlovian(sp?) training ). Biologically we are wired ( which would be your "prewiring" ) to react positively and negatively to various stimuli. This goes back to single sell action / reactions. Assuming evolution and natural selection, then we are the products of organisms that have successfully learned how to react positively to things that promote our life, and visa versa. The higher level organisms can associate things to those basic hard wired "good" and "bad" stimuli. I believe a dog could have 3 or so levels of association, and humans were something like 8. ( The dog hears a bell, and anticipates the light that use to mean food was to be prepared and so the dog salivates even though no food is procured ). Likewise the repeated use of the word no to a child alongside a spanking associates the word no with pain to the child. Eventually the child forgets the pain and associates no with "bad". Similarly we "learn" the word good.
IMHO we have very little hard wired in. Humans in particular are probably de-evolved as far as instincts and hardwiring go. Instead, we've become dependant on our superior associativity skills. With that we're more adaptable then say a bee that naturally reacts to the ultraviolet colors in a flower.
I give little credibility to our prewiring having anything to do with our parental mating since the above adaquately explains things for me. I've not really seen any credible cases of children seperated at birth that took on the traits of their parents that could not be properly explained by genetic predisposition. Perhaps your article provides more compelling evidence.
As to the relationship to quantum evolution. I'm torn on the idea. The idea of a multi-verse is personally unsettling. I never liked the sci-fi plays on the concept. But any such hypothesis that provides an essence or "will" to a seemingly chaotic or random course of events aids the understanding of things such as AI. A program, as you say may seem to only be capable of spitting back answers that we once fed it, but this is becuse they are algorithmically bound. Even our random number generators are deterministic. But what if we used a quantum random-number generator. If there were any truth to the "will" of sub-atomic particles, then strange non-random patters should arise in algorithms that make use of such randomizers. I totally believe that this sort of theory is testable. Perhaps it is the next step towards true AI. Life bound by sets of rules that man has created. Perhaps a little scary.
Originally RISC => "Reduced" Instruction Set Computing. I'm not sure where you've learned that achronym.
Additionally, there are almost no true RISC CPU's any more. All modern CPU's should be considered Post RISC. Eg, not all instructions are 32 bits ( as par ALPHA and 64 bit constant loading, etc ). Additionally, they hardly execute in a single clock anymore. This was only possible when no FPU was utilized ( or at least was used as a seperate FCPU with CPU synchronization ). Additionally, the number of instructions is once again sky rocketing in these post-modern CISC chips.
I can just see chips like ALPHA, PowerPC and Sparc as having the exact same sort of legacy compatibility issues as modern CISC. Heck, Sparc's register window really can get in the way in multi-threading / tasking.
What I really like about the cruso idea is that you can make a CPU that has vector processing if you like, then translate MMX / SSE / AltiVec / 3DNow as need be to utilize your own proprietary method. This way, when a non-compatible method comes along later, you don't have to spend the silicon on compatibility.
-Michael
If you read the MS FOF, then you'd know that MS DID purposefully make changes that broke DR Dos with their windows code back in the day.
I don't mind AOL being big, popular, easy to use, having proprietary content, etc. But what I do mind, is their purposefully sabotaging the competition. This goes far beyond 5.0's alleged breaking of other ISP components. It involves blatant incompatibilities with mainstream internet standards, which involves, among other things basic email.
Unfortunately, I don't think this case stands a chance. But I do hope that it hurts AOL's pristine image, and more importantly that it causes AOL to change their ways. They are fast replacing MS as the corporate evil of the computer world, IMHO. In fact, I think they will be much more dangerous than MS ever was. AOL has the ability to disrupt the operations of entire companies with it's network connections. There is little or no privacy with any material that passes their their gates. If they are successful in stomping out non subsidized ISP's ( corporate, university, phone, etc ) over the next 10 years then they will have unprecidented control over information. Something MS never really had.
-Michael
First, It amazes me the maturity level one finds on public discussion boards. I am very grateful for the moderation system, so that, in general, I ;-) ). The x86 is still alive and kicking, /^A/; END {print $x }' " and other such one line programs that I can test the concept of in less than 10 seconds.
do not have to be subjected to this. But on topics such as this, I feel compelled to actually read everything. How people's opinion of Tom
Christianson relates to this escapes me. He has contributed considerable intelligent information that can only help the Open Source community.
I'd like to present the idea of evolution, if I may. A man evolves by growing more nimble fingers, a larger brain with greater capacity for
memory, problem solving, etc. He evolves by being able to withstand harsh cold ( or hot ) environments. Often times, man finds himself with
strange and seemingly useless biology: An appendix, toes, back hair, various skin tones and shapes, etc. Life, on the other hand, evolves
through natural selection of species. When one species can no longer adapt ( analogously finding a local minima that they can not surmount ), it
is replaced by others.
The same holds true for our silicone world. A technology is created. The environment changes, and the technology must adapt or die. Death
involves reinventing the wheel with our new life form; starting from scratch on the technical design, AND convincing a target audience to
support the new technology. Thus, just like in nature, the predisposition of a technology to adapt is greater than it is to die off and be replaced by
something new.
As a case example, take Intel's x86 ISA. A venerable and intelligently adapted system. Sure, the original draft of x86 didn't take into account
32 bit memory, pipelining, n-way parallel instruction execution, etc. Yet that same language has been adapted to systems today that are among
the fastest and cheapest in the world. This of course, was not to the merit of the appauling instruction set, but to the adaptiveness / inventiveness
of their designers who knew how to maintain their market. Before shreding the idea of the x86 ISA, try, yourself, to invent a hardware language
today that can cleanly adapt to technologies 10 years from now. ( neural message passing systems possibly? ) It's really hard to do. The only
real way is to have enough of an influence to periodically reinvent yourself in an incompatible way ( as Apple or SUN was able to do. Though
their lack of a perfectly competative market reduces their value to the computing society as a whole. )
Perl, likewise could not have have forseen all the uses that it has today ( and therefore produced a language suitable for everything ). Written
first as a scripting language, then later as a full general purpose language. Never being fully without bugs, constantly reinventing itself, yet
keeping backward compatibility as best possible. Still later, we add certain modern programming techniques ( again without breaking
compatibility ) such as OOP, and threading. Threading isn't fully stable, and many important packages ( such as DBI ) warn against using it,
while OOP is something of a perversion of classical approach. It is difficult and inconvinient to consistently define class objects. When dealing
with OO, I much prefer java ( and to some small degree, Python ) since these were created after the OO revolution ( or skirmish, if you prefer ).
Still, perl seemed better apt to handle these two revolutions than C ( which isn't pleasent for writing threading code, and had to be mangled in
order to handle OOP ). Without breaking backward [symantic] compatibility, perl has managed to encorporate many exciting modern features.
It's nature has even allowed it to adapt to web server-side embedded scripting ( in the form of ePerl, embPerl, mason, etc ).
My focus is not that perl has conqured the world, but that it has survived the changing world. It has its scars and strange looking unused
appendages, but for it's intended purpose, it provided zero down time through the evolving years ( referring to porting time ). People that buy a
new linux distribution could still use their old perl scripts as is. But additionally taking advantage of the newer features in the new release. Much
like one can do with the x86 architecture.
If we take a look at java for a moment, we'll see yet another example of the uglyness of evolution. Originally concieved of as a fresh new
concept taking advantage of the state of the art in computer science ( If you'll forgive the proposition ) we're already in our 3'rd revision, which
has made heavy use of the term deprecation. First we do it one way, then decide that there really is a better way of doing it. But now we've
discovered a more elegant approach to event handling and change so we change things yet again. Each time, it was essential to maintain
backward compatibility ( this is still a fledgling system, and supporters would not accept such unreliable lack of support with legacy applications
as they might with the conversion from SunOS to Solaris, or mc6800 -> powerPC ). Java is in the same boat as perl and x86 in that as long as
the language evolves, it will be plauged with legacy features. Thankfully I can think of little that is hurt by the existing evolutionary steps in Java,
but it is still much younger than perl or x86. Give it a decade and we'll see how well it fairs against new and shining languages. And some will
support Java even though it is considered archaic and too difficult to keep up with the new boys. Or stead-fast C coders will mock Java, saying
C was the true language all along ( as some humorously still do about assembly today ).
The key is really whether a life form ( or a technology ) has outlived it's usefullness. Java obviously hasn't. It's a good programming model (
with far fewer points for error than C++, with what I'd call, similar support for the technology ). ( Notice I said model, since, obviously the
execution time is subject to debate ). C is going to be around for as long as UNIX. Assembly as long as the discrete instruction model exists (
I'm all for neural messaging systems personally. Interpret your code directly without compilation!!
simply because the technology is such that the underlying negative effects for a non-optimal instruction set are offset by pipelining and caching (
as in the case of AMD's micro-op cache ). To boot, this performance increase in changing the ISA is miniscule compared to the inconvinience
of developers supporting a new platform. The added cost of compatibility translation hardware is offset by the massive economies of scale. As
far as I'm concerned, we were taught a long time ago to not program in assembly when you can program in C, ( and consequently, not program
in C if you can use something more aptly suited to the task ). Thus the processor is a black box, and I could care less what instruction set it
uses. Give me performance / cost ( and if I'm in a laptop, additionally devide that by power consuption ).
So what of perl? Originally designed as a reporting language ( before the days of the Dynamic web ) it now finds itself primarily as the glue
between c libraries. To my knowledge, perl was not designed as the creator or the repository of information, but for the "extraction", and
"reporting" of that information. Thus the DBI hook to mysql and the CGI front end allows it to format the data found in so many of our web
sites. But this specific functionality has been built anew by php. Doesn't the [cached] CGI method of c-like logic control structures seem
outdated in the modern world of table generation and data reporting. There have been no new op codes in perl to conviniently generate an
HTML table in perl ( though embPerl and the like do so nicely but at a much higher level ( and thus less efficiently ) ). Thus has perl been
replaced by php or the like in the realm of simple DHTML DB reporting? And since HTML is a far more common report medium than
vterminals, has the underlying theme of perl been superceeded? I'll leave this as an open item.
Whatever perl's role in the web, it still has roles in System programming. I find perl much more powerful and pleasant to program in than sed,
awk, or even bash. If you need to write a filter or data converter from one database to another, I doubt that you're going to use php, and possibly
not even C ( unless you really don't know any others ). If you want to provide a summary of slack space used on your hard drive ( sorting by
directories, then by files, etc ), I'd find a hard time doing this in any other language than perl. ( Not impossible mind you.. ) Also I find perl very
useful as a configuration script or simple socket daemon. Additionally, I can take advantage of many libraries that have perl hooks into them.
OpenGL, XML parsers, TK, QT, Gnome, OLE, NT registry, socketing, etc. I can take advantage of these without having to learn all the
intracacies of each language. I can play with new technologies found on CPAN, and decide early on if they're worth persuiing. The late binding
and non-strict type checking really aid in such a transparent interface.
Peronsonally, I find these times exciting. Who will win out in the language wars? Of course there will be more than one winner, because there
is more than one target audience. As each new computing need arises, existing languages will adapt, and new tailored languages will appear. It
is this trade off that the devoloper has to make. Personally, I like the advantage of perl's wealth of c-library compatibility, so I write many
things in perl ( and even write emulators of HTML templating, or even the FastCGI / java-serverlet model to suite my own needs, just so that I
can make use of certain libraries that are not readily availble in php or Java ) It also means I can have a common library package for both
configuration / database manipulation and CGI scripting.
I like the open nature of perl ( hard to compile away into hidden libraries ), I like the "more than one way to do it" methodology. I like that I
_can_ write strongly typed modules ( through perl 5.005's "my Foo $var = new Foo;" and the "-w" and "use strict" directives ), but can also
write "perl -pe '$x++ if
( though the obove seems clearer with 'egrep "^A" | wc ' ). When solving such specific pre-defined dillemas, perl excels.
I should hope that I have to say very little about the most important contribution of perl; it's regular expression library. Highly optimized C (
depending on which version of perl you're using ), has managed to take an NFA construct and extend both it's efficiency and feature set. I have
never seen such an extensive reg-ex package in all my years of programming. ( I would definitely be interested to learn of more powerful ones
if any of you is able to provide me with such info ). php has seen fit to include perl reg-ex's, but I'm not sure how up to date, or how completely
compatible it is. To be able to mathematically speak to your computer in such a way as to have it almost understand what you are thinking ( in
terms of syntactic identification and manipulation ) never cieses to bring a smile to my face. Having the ability to intelligently merge a group of
HTML files and into a single one with 3 lines of reg-ex code expresses such a mathematic bueaty I can not properly describe it. It isn't about
doing it in fewer lines of code, it's about commanding power over symantics. Of course it seems cryptic to the uninitiated. So does calculus. Is
that any excuse for shunning the language? Neuton invented Calculus because it was too tedius ( and in fact impossible ) to correctly describe
bodies in motion without it. Yes I can write an optimized and easy to understand filter or parser in simplistic C constructs, but reg-ex allows you
to attack the problem from a completely different and extensible approach. I've regularly written entire parsers in perl ( memory management
alone eases the pain of writing a parser ) as a simple OO module that gets loaded into a larger program. Meta language is therefore much more
obtainable in a smaller development group. Even today, perl's reg-ex is being extended with parenthesis matching, and the refinement of a
procedural approach. So long as perl contains the most powerful and unmatched reg-ex library, I will not lose use for it. But meanwhile,
nay-sayers will scoff at perl code littered with such "unreadible" expressions, happily ignorant of the calculus of the programming world.
Lastly, on the topic of a first time language ( which seems closest to the original topic ). Perl is great in that it allows the most straightforward
possible program. perl -e 'print "Hello World"'( as with BASIC ). You can even enter the debugger and type this in manually ( again, a la
BASIC ). You can create an initialized variable the first time you use it.
$x = 2 + 3; print $x;
Essentially, I'm claiming that perl is every bit as good as BASIC as a first time language. Moreover, it's better than the old GWBasic, Apple
Basic or TI Basic that I first learned on, in that we're not dealing with antiquted line numbers, or goto's. Intead we have very nice flow control
structures ( last, next, redo ). Aside from that, you could completely take a BASIC lesson plan and "port" it over to perl, with the advantage that
the student now contains some real world experience ( as opposed to BASIC ). I doubt system programming is a good second step. Perhaps
CGI's, though you really do need to have _some_ programming experience before you're allowed to crash your school's local web server.
Just for reference. I learned the
Basic -> Pascal -> Scheme \
C -> C++ -> System programming ( insert thousands of languages here including Java and Perl )
I kind of like this approach, but am basically questioning if Perl could replace Basic as I've said above. Real work should _not_ be performed
until a C like language is well understood. But there is no reason you couldn't return to perl at the system level.
Well, this is long, and time is short, so I shall end you here.
-Michael
Ahh, my fellow comrades. The dreamers, the doers, the opinionated. We who love SCI Fi, we who dream of a free and unified world. We that look to the stars...
Of course, not all of us have our head fully screwed in properly. Do it, damned the cost ( since money is of little value to us right? ) Even in a moneyless "Star Trek" society, there is a cost in everything: The human man hours of skilled and talented workers ( opportunity cost ), the physical resources, loss of life or general working conditions ( quality of life ). Communist society, for example, would have us all living more humbly for the "greater good" ( which of course is in the eye of the statesmen ).
In reaction to colonization of the moon, mars ( or even space stations ) as a supposed solution to the failing of NASA, I say, Action with no incentive is foolhardy indeed.
One reader pointed out that we can not hail NASA's ( and hence man's ) conquest of the moon, since it's incentive was, in fact, a life threatening cold war. Something greater, and more profound than man's perseverance and ingenuity, is _life's_ fight for survival. We did not conquer the moon because we could, but because we had to; when backed into the corner, this cat lashed out with all it's might ( and resources ). This devotion of resources is something that finds no such motivation today. And I would challenge anyone that craves the back-firing black magic that brings back such motivation.
What we are left with are millions of dreamers pointing fingers saying, "but you did it before". And a government agency doing what I assume is their best to efficiently convert their scare resources into visible science.
As for a moon base. I'd like to point out that we have YET to successfully produce a bio-dome HERE ON EARTH! We do not yet fully understand the delicate balance of nature, less how to command it. Colonization, could therefore not be fully sustained by these currently lifeless worlds. The cost would be unimaginable. And of course we are talking about residences, which of course would have to be regulated. I'm sure the ACLU and other rights organizations would have a field day with many of the near-death experiences. If you think the set backs in the simple non-permanent space station are appalling...
As was also pointed out, we have little human critical experience with the extra-terrestrial. How do you mend broken bones and other ailments? Do we send groups of people within the next decade to their doom? Do we risk the life of our most talent and promising astronaughts?
I believe it to be naive to say that we're wasting money on these conservative, yet expensive projects, so we should go the full gambit and colonize! There are several orders of magnitude of complexity, risk and COST!
I would call this the brute force solution ( much like the lunar landing ). I have always idealized the intelligent, dynamic solution; one that isn't fully apparent at the outset. Do what is best at the moment ( hopefully as adaptable as possible ), and recheck yourself regularly to see if changes can be made to either get you further ahead, or ( as is often necessary ), to reinvent yourself.
I believe the "commercialization" of space is a good thing. The private sector ( and our potential entrepreneurial pioneers ) are welcome to build upon off-the-shelf rocket science and find some personal goal ( which just happens to further mankind ). Zero gravity Space products, and possibly lunar / mars mining are interesting incentives. The main thing that I like about them is that their motivation would be "economics", ( which doesn't mean money, btw ). You do what is most efficient at the moment. When things look bad, you kill it with little after-thought ( don't throw good money at bad ). When someone seems to be succeeding, other venture capitalists will jump on the band-wagon, thus expanding our knowledge and experience base. When dead-ends are reached, people will look elsewhere.
Since there doesn't seem to be enough vested interest, this will take time. The entry barrier is still too high for this model. But the incentives and the means are becoming ever more apparent. This is my personal hope for the future of space-technology.
My big issue is that government agencies do not follow the model of [economic] efficiency, but instead follow task oriented mandates. "Go to the moon", "stop crime", "Make hubble work". Unfortunately, their support ends with the completion of a given task. There is no competition, no reward system, no punishment ( other than general monetary prioritization and national pride ). Either a private sector NASA needs to come into existence ( possibly as with the post office ), or we'll have to continue to live with this task-mastering until it becomes economically feasible to surmount this "great space barrier".
In the short run, I support a congressional review of NASA. It may bring to light the negative effects of cutting costs by reducing redundancy. It may shed light on administrative weaknesses. It may very well bring about a more efficient and productive NASA. Sadly, it will most likely involve cuts in funding. But as such, I do not personally find value in mars exploration ( gasp, blasphemy! ). If there is life, it will still be there in a hundred years. Even if not, it would only be one of the billions of interesting facts that we've lost.
The Matrix: You humans seem to define your reality by your misery.
Me: That misery becomes our newfound motivation. Without it, we become complacent and wasteful.
-Michael
Maybe I'm missing something here, but if we're so behind on outer space technology and "breaking low earth orbit", then how do we send out all those mars, venis, and solar-system probes. Yes they're small, but the point is that the technology is there. We simply have learned more intelligent methods than the brute force F=ma.
Just through I'd give my opinion on the matter after reading as much as I could get my hands on.
Pros:
Allows me to communicate with my grandparents, or any known non-techno savy people.
Allows me to contact relatives I don't know the email of, or to people I haven't contacted in years and can not be sure of their email address( and don't feel like making a phone call ).
What I find interesting is that the USPS is probably in a better situation to do this than commerical companies, since they already have access to the address of every man-woman and child, so to speak.
Cons:
If not done properly, it can lead to spam, unwanted charges, or wasted use of paper. It is essential that they do it right the first time.
Even worse would be the exploits that we ( or their committee ) don't think of.
You now have a new form of a publicly viewable Social Security Number. It's only a matter of time before companies REQUIRE your private / personal email address for services.. Simple string parsing would validate the request. Now, just like the Pentium III Serial ID, or your social, malicious companies could exploit known shared info. Possibly credit history would be attached, medical records, etc. In fact, all SSN info could be mirrored by your USPS ID. Heck, some sites would probably try and link SSN to USPS ID. The solution to this, I believe is to not allow a per-person email address, even though this doesn't fully eliminate the problem ( now we just track house-holds, instead of individuals.. But it's a house-hold that buys a product, so it's still valuable info ).
Additionally, since USPS is federally subsidized, any significant innefficiencies are partially passed along to the people. It is important that this not become a huge multi-million/billion dollar flop which requires federal bailout. Thankfully I doubt the system would allow a perpetually innefficient system to exist.
Proposals:
Optionally allow mail forwarding to existing address. Their site would simply provide a consistent address. This minimizes their cost, since they wouldn't have to store the mail long term if we didn't regularly log in. Also doesn't require us to have yet another email account to check daily / weekly.
As with USPS mail, the sender is billed. This alleviates much junk mail / spam. The downside to this is in auto-reply email, where a person registers with a web site, then puts their USPS-email address which for some reason is mapped for printout. Now my poor free web site is being charged for many "potential customers". Additionally, it provides for a seriously expensive DOS attack. You now have the ability to rake up millions of dollars worth of USPS bills if a target email-responder site is repeatedly hit.
The solution, in my opinion is to set up a billable account with USPS, and then potentially billable email would have to be authenticated and authorized ( just as in any e-commerce transaction ). The default auto-response web-site would obviously not provide a mechanism to send billable email. Unfortuntaely, this would either require a client side program ( possibly in java ), or to make use of CGI's that require either cut-paste, or browser-file-uploading. None of these are ideal, since they don't make full use of your existing email programs.
Another method would be to simply send the email, then if it requires payment, a notification is returned, requiring you to log onto their web site and authorize the transaction.. Unfortunately, this makes it easier to spam, since everyone can be mailed, and the payment-based transactions would simply be ignored.
In order to alleviate spam, the central site could possibly monitor mail volume, and automatically charge accounts that exceed a certain volume, with the notion that spam/ junk mail is the intention. High volume is expensive for the central web site in any case ( due to excessive local storage, etc ). Another thing I like about this, is that it minimizes chain mail, since you'd be addressing dozens, or hundreds of people regularly. I don't consider this stiffling of free communication, since you'd still have your other email addresses to use for such time-wasting things. I'm not a big fan of email-based mailing lists anyway. That's what bulliten boards are for. If it's supposed to be daily, then they can regularly check the bulliten board, with little or no excuse.
I definately think this issue deserves attention, since there is a lot at stake; Our privacy, financial obligations( for both sender, receiver and USPS ), and our dear forests. I do not, however believe that ostrige-like-fear should hamper progress.
-Michael
I believe there is a certain security issue with this. I remember in days of old, certain stores acted as a local post office, but robbing such a place could also be considered a federal offence.
I am under the impression that post offices today are isolated, partially to prevent it from being a direct target of crime( as would be the case in a grocery store ).
I'm sure it's possible to reside in a mall, but I get the feeling that there's a good reason to not.
-Michael
A Few points:
- CNN was obviously erroneous about the Patmos using AMD K6's instead of AMD Athalons.
-- 200MHZ bus and the words cheap obviously seemed like a K6 to the semi-technically knowledgable author. In fact, ( as has been pointed out here several times ), the only 200MHZ bus available for a PC is the Athalon.
-- They mentioned Patmos reaching 1GHZ w/in the year. The K6 does not have the potential to make this speed, in fact, few processors in the world other than the Athalon are currently capable of reaching this goal within a year. Especially without requiring a motherboard; which obviously would be bad for Patmos, since I assume they provide _some_ form of custimization of their motherboard. Or at the very least have carefully selected a board, and would not think kindly of choosing a new board so soon after their initial product release.
- The unqualified use of the word super-computer.
I've noticed several posts about people thinking that they could design a "super computer" even cheaper than Patmos. But really, all they could achieve is a "theoretically fast" machine. A supercomputer is the sum of all it's parts, and therefore the weakest link can break the chain.
As a disclaimer, I am not formally trained in super-computer concepts, but much of this is based on my experience and common sence ( which may differ from horse to horse ).
A super-computer must have top tier performance ( obviously ), must have data-integrity ( you don't spend half a mil just to have a core dump, or system freeze ), reliability ( 24 / 7 uptime while performing it's work ). It should also be scalable ( grows with your company or the task as is fiscally justifiable ).
--Simple points: When selling a super computer, you must choose high quality parts ( or at least make things n-way redundant ).
In my experience, IDE drives don't cut the mustard, due to their high volume, minimal quality price-focused nature. ( skimp on a heat-sink or shock obsorber to save 5 bucks per drive, etc ). When you buy IDE, you think disk space and low price. When you buy SCSI you think of performance and quality ( and usually expense ). Thus they are designed based on that marketing paradigm.
An IDE drive also uses an IDE controller, and is thus inherently sequential. A SCSI device can queue multiple independant requests, so as to perform disk geometry to determine an optimal seek path. Additionally, due to the paradigm above, more cache and higher rotatial speeds are applied to SCSI devices ( not that they couldn't be applied to IDE.. but why should they? )
As for a network, some referred to HUBs and ethernet. Ether does _not_ scale well. Sure you can get a faster / more intelligent HUB, but you never achieve maximal theoretical bandwidth. I'm not completely sure of the network technology used here, but it seems to be peer to peer and bi-direction ( to facility rapid acq's ).
--Memory. This is really the key to a good super-computer design. SGI made use of wide 256 bit multi-ported memory busses with interleaved memory ( 16 consecutive bytes was segmented across 4 memory controllers, thus linear reads were faster, AND independent concurrent accesses had a statistical speed-up ). Of course a 256 memory BUS is expensive, especially in a multi-CPU configuration. SUN's starfire, for example, has up to 64 fully interconnected CPU's ( don't recall the BUS width ). This required a humongous backplane with added cost.
AFAIK, the Athalon uses regular SDRAM ( and a cost effective solution would have made use of off-the-shelf parts ). SDRAM is nicer than older PC-based memory in that it's pipelining allowed multiple concurrent memory-row access within a chip. Several memory addresses within the same row could be in the pipeline, and up to 2 rows could have elements in the pipeline. This is a more sophisticated approach than interleaved memory, BUT, it introduces a longer / slower pipeline. RAMBUS furthers this concept by narrowing the BUS width and furthering segmentation. It allows greater concurrency, but latency ( and thus linear logical dependency ) is increased.
RAMBUS's theory of high latency, high concurrency benifits non-linear programming, such as Italium's ( Intel's Merced ) deep speculative memory prefetching, or ALPHA and SUN's multi-threaded processors ( where cache misses cause an internal thread-level rapid context switch, thus hiding the latency ). Existing x86 architectures, however can not fully take advantage of such concurrency, and the net effect is slower execution time for linearly dependent algorithms ( non-local/consistent branching, and non-parallelizable math calculations ). In this case, making use of high speed / low latency interleaved EDO ( as is / was done in several graphics boards ) seems a better alternative ( but hasn't come to pass in mainstream motherboards ).
--mutl-CPU. This is an interesting topic. Mutiple CPU's can connect to the same memory ( with large internal caches ), or can have a numa architecure with shared segmented memory ( isolated, with interconnecting buses ). Or they can be autonimous units connected via a network. There are pros and cons to each mechanism. The last requires the most redundant hardware ( which is actually good in terms of hot-plugability ), and has the slowest inter-CPU communication. It thus works well in message passing systems, as opposed to symmetric decomposition of large data arrays ( eg parallel vector processors ). Personally, I like the NUMA approach the best, but it requires proprietary hardware, and hot-plugability would have to take the form of a VME bus etc.
It would seem that the approach here is multi-CPU ( 2 or 4 ) to perform a single task. Concurrent threads are distributed across machines in a message passing system ( hopefully minimal data-sharing ). The AI controller probably handles messaging, arbitration, in addition to the advertised load balancing. The multiple CPU's on a board are most likely for redundancy. My guess is that 2 or 3 CPU's are used for user-thread redundancy and a 2 or 1 CPU's are dedicated for OS operations ( using spin loops in the user threads ). Thus minimal context switches are required, minimal memory bandwidth is used ( since only one virtual CPU is ever accessing memory at a time ( though 2 CPU's are simultaneously requesting that information ). They may actually allow the Linux scheduler to rotate proc's, but as I've learned, this isn't Linux's strong point. A single tasked CPU is a happy CPU.
I know SUN has optimizations for context switching ( keeping most pertinent info w/in the CPU, along with a unified virtual memory model, as opposed to the offset-based x86 virtual memory model ). Unfortunately this is offset by register window swapping, but such context-switching centric processors would allow for more efficient concurrent operations such as multi-threaded apps ( such as java. Before you laugh, one application of this low-cost supercomputer is web servers.. And serverlets are an emerging technology, people will ask the question, how can I make this existing code run faster in a short period of time ).
-Concept. ASICs / FPGA's. SGI had the concept of making a simple, cheap, reliable, and fast logic CPU, then couple it with extremely proprietary logic / processing chips that offset the code logic. Combinational logic is faster than any sequential logic, though much more prone to bugs, and higher production costs. High performance reprogrammable FPGA's could help the industry, since the hardware could maintain a high volume, low cost ( as with current CPU's ). Thus you could make PCI / AGP expansion boards that handle load-balancing, message passing, java-extensions, OS-operations etc.
I'm sure their AI logic is done similarly, but it's a completely seperate box, my thought would be that the "boxen" would have these expansion boards, and the customer could request optimizers for say, the web, or weather calculation, chess designs, what-have-you. The goal being that these expansion boards become as common as modern graphics accelerators, modems, sound cards etc, without having to go through all the trouble of designing the hardware of those boards.
-Michael
One further point. My initial statement was that exponentiality is only bad in a scarsity.. A good example is math.. I can write a seemingly infinitely large number... But in reality I have a scarcity of paper or time to write it.. But I can be creative enough to represent astronomical numbers efficiently. Arguably, those that sought to find pi with inifinite precision in days of old were challenged with this feat. But they too did not consume 100% paper, or skip sleep or their wives..
I hate when humans are compared to a virus because all a virus does is eat and reproduce. We have art and inspiration.. So long as we don't become mindless consumer zombies, we'll have an edge on exponentiality.
I think the original author's point was that was the danger. But myself and my children will not be plagued with this; not because I'll keep the TV and internet off, but because I'll excite them with the many physical wonders of the world as my father did with me.
-Michael
exponentiality only plays a factor when there is or can be a scarcity. There is only so much land, sunlight, soil, free O2, etc. So mold growth, viral replication, etc when grown unchecked will quickly consume all resources and start dying out and rapidly.
:)
Information consumes several resources. First and fore-most it consumes time. If "media" grew to consume 100% of time, then we'd have a recession and then a depression, and thus a death of society; course humans are more pro-active then a virus. We'd make laws, etc when enough people understood the danger.
Another resource consumed is disk space. You can't store every piece of information ever gathered by anyone. The ideal would be to have an MP3 recorder on and with you 24/7 so as to reference anything you've ever heard. Likewise with a video camera.. Beyond that, you could sense and record everything in the world 24/7 - thereby approaching omnipotence on a local scale. But the storage of this information would be impossible. Typically we record onto ferrite, and there's obviously a scarcity of metal in the world. Other's are using phosphors or plastics, but they too would be a scarcity.
So what happens in the "hard drive revolution"? Well, we use metal for quite a bit, and so to take metal away from one form of consumption produces competition and weighing of values. Look at DVD-ROM drives and cell phones.. They share a common component (don't know what, but I've read it often enough on slashdot). Because cell phones are 'exponentially growing', DVD-ROM's are prohibitively expensive. Manufacturers of course try to get around it by making them faster and thus justifying the added cost.
The danger again is the value storage above all things. So we lose the ability to use the materials for other purposes, and again humanity becomes pro-active and regulates. In reality, we have a diminishing marginal enjoyment. We're not going to record and playback information 24/7; we're going to be highly interested in other forms of recreation, or in rare cases; work.
As other's have pointed out, there is also the prospect of human memory.. Our brain is a finite storage system. To my understanding we "memorize" relations between experiences. We comprehend a sound, and relate it to some higher level concept (such as a vocabulary, or fear, etc). Most of the experiences we have are lost, and only the gist of what went on is retained - some more than others. Beyond that we'll have a fleeting interest in all the info out there. Some would love to have ready and repetitous access to espn channels 50 through 3,000. Other's would be interetsed in the discovery series 50 through 100, etc. So different personalities are going to filter that info uniquely.
Next is the issue of information predators. Well, first and foremost is the minimal resource requirements of information. Second is the reusibility of most of these resources (except time and energy consumption). You can only eat a plankton once, but you can record different MP3's to the same spot on a hard drive over time. The real predators are interest and time. You may be facinating today and thus over-write older material, but tomorrow you're old news; bye bye.. Garbage In, Garbage Out.
Next comes Darwin with Natural Selection. Information that gets discarded quickly is selected for deletion; they'll have their moment.. Mutated from similar topics of the past. Copied web sites, re-reporting news on CNN, etc. Each time they're a tiny bit different, and they just might have found the right audience; maybe Bill Mawr had a comical spin to the same info that made a certain class suddenly interested in it. But ultimately, eveything has it's time to die. And information can die at a much more massive rate than life; just like an over-population of viri in a host, once the host dies. Entire hard drives can be wiped; entire broad-casts can be over, without ever recording them.
The saturation point is where we max out all possible bandwidth. Our entire visual perifera is filled with mini windows of streaming content. Any additional information beyond this immediately dies because of lack of additional audience. Even if they are recorded, there is little chance that every moment will be revisited until a new baby is born or some other content dies. But now we're back to postulate 1 above that we're spending 100% time in information consumption. So my argument is now concluded.
-Michael
Ethics is a highly subjective and volital thing, but is none the less an essential part of science / engineering.
My personal value system suggests that manipulation ( in all its forms ) of single celled life falls in yellow catagory which means acceptable, but with precautions. Research such as this should be public knowledge ( genetic war-fare is a potential down-side )
Personally, I don't think that there is anything special about life. We are simply bueatiful machines crafted in an elegant fashion. This is completely independant of whether we were created by an intelligent being or not. Just as we seek to explore the physical world around us, why not seek to better understand the most basic workings of our being. If you are relgious and are having issues with "defining-life at a genetic level", then think of it as simply better appretiating "God's work".
As a point of reflection, Judao-Christian Dogma states that "Thou shalt have no other Gods before me". If God can be defined ( at least in part ) as the creator, then surely mans ability to manipulate atoms and now create life places us dangerously close to defiance of that tenant.
I'm one of those egnostic types, somewhere between atheism and religious so I like to straddle the fence ( It's more interesting than only taking one side ).
I remember the old days of BBS's. We'd have monthly get-togethers that included pizza. We were all a big group of friends; the younger crowd talking about games, the older crowd talking about hardware all in the same room. Sure enough, the feel of community is removed when you're 10 thousand miles away from everyone.
.plan and .profile's. We had a sence of our own little space that we could decorate and other friends come come and visit.
When I went to a University, I actually found a nice replacement for BBS's in the campus central UNIX machines. I'd log in and track who else was on. I could be productive, while chatting with 3 or 4 other people. There were bulliten boards, download sites. Before the web was popular, we'd play with our
Around that time muds/mush's/mucs were popular. There were tales of people failing out of college playing muds 18 / 7. Those muds allowed chatting in a pseudo IRC fashion, but just as on the UNIX machines, we had our own little piece of the virtual world that we could decorate and share with others.
The combination of these provided ample substitute for me in the mid 90's. BBS's were still popular ( and I activately sought them out ), but as FTP sites fullfilled my download cravings, I had removed the last of the BBS advantages.
Role playing, chatting ( with local people ), files, information, home decoration...
Now AOL's Instant Messenger, ICQ, and Yahoo pager are my primary chat resources ( localized apps are much faster and virsitile than an IP packet per character ). My Web page is all the person information I need to keep. The web itself handles all my files. And I've been too busy to seek out those MUD's of old.
Actually, another good thing about the BBS's was that you couldn't be a junky. You only had 2 hours a day, and all your games were turn based, which limited so many actions per day. There is no reason why tradewars can't be loaded onto a telnet compatible server and still use the turn based mentality. I would hope that it's already been done.
Just my 2 bits of nostalgia.
I have 4/5 computers at home that perform various functions. My main computer has dual video and is my gaming / DVD / internet machine ( win98 sadly ). For this I have a dedicated high quality monitor / keyboard. But I also have a second video card in it connected to a second, lesser quality switched monitor.
:)
Because several of my other machines run in Text mode ( servers / burners ), the video quality was not an issue for me. However I still do 1024 X and have a second NT / 98 box ( for networked games ). So I have all high quality vid cables ( at my local computer fair, they cost me $10 as opposed to the $4 for cheaper VGA ). The switch box does introduce noticable distortion, but only for very small text. It's been fine for Linux development ( so long as I choose my font / color scheme appropriately ).
My switch box cost me $25 at the fair ( 4-way KVM ). I run into the problem of having to switch over in order to boot the machines, and on some boxes I need to "gpm -k ; gpm -t ps2" to reactivate my mouse after switching to other boxes. ( anybody know if there's a better way? )
I seriously doubt this cheap solution is any good for very high res monitors. It only works if you can dedicate one monitor / machine for your main work, and just have others for occasional use ( where loss of quality is acceptible )
Still, I like toys, and am willing to sacrafice quality over quantity of toys.
If you have the money though, $100 active boxes are your best bet. Personally I don't find value in the KB key capturing for switching ( unless you really do it often ). ( In my mind, the greatest value of switch boxes are to reduce the cost of monitors.. Buying a thousand dollar box ( as I think I've seen suggested here ) really seems to push it ).
Alternatively, getting eXceed ( or the free MixServer ) for windows, or VNC for Windows / Linux will allow you to consolidate displays.
-Michael