Non-linear editing takes as much time if not more as old fashioned film painting. Even with digital video you've got to composite five layers of video, whether it be film or bytes. To do production quality digital editing you need a very expensive facility. Remember 35mm film is REALLY high resolution, so everything needs to be editing composed and rendered at that resolution. That takes forever even with really powerful computers. Digital equipment is initially expensive, very very expensive. Once it saturates sure you have your money's worth and can reuse a good deal of it but regular film can be alot cheaper because you can go back to cut and paste editing if need be not so with digital. Digital projectors are expensive as can be because theaters need to rip out all their old equipment. If you buy a newer film camaera you can still use your $75k light housing for it. Do some research rather than getting evangelical due to other people's posts.
One of the staed goals of the P4 though is to really whomp ass in terms of clock speed which is something I think it can be agreed thay've done. They needed to put something out that could do two things 1) make the home PC buying drool and drivel over the word gigahertz and 2) do things people wanted to do quickly. If you have a processor that really flies at video rendering and sound processing Dell and Gateway stick FireWire ports in it and classify it a multimedia editing PC to get eyes turned away from iMacs. Non-techies see the numerical difference between 1.2 and 1.5GHz and their eyesbrows go up because they understand that the higher number MUST be better. You're very correct though in saying render farms of these things will sink anyone. raytracing and picture editing use algorithms over and over again and are best kept in the cache until served with a cold soup. Coppermines and Willamettes have tiny caches that make it difficult to keep more complex (read larger) algorithms in the L2. When you buy an SGI Onyx system the processors have pretty heavy cache sizes which facilitate repeticious functions like this. If businesses want/need "high-performace" they ought to be buying Xeons if they want to stick with IA32 or maybe the UltraSPARC3. Quality hardware + quality software == justifiable cost.
Dude, what the fuck good would a "Linux" FPGA do for anyone? Ohhh, I get DSP performance out of my kernel? Big fucking deal. Its the 3D rendering or Quake3 I want DSP performance out of. And more to the guts of your post. Most modern RISC processors out now have a large amount of specialization either in their instructions or processing units. MIPS binaries don't run on SPARCs now do they? RISC is a good chip architecture but it is no reason to thrash CISC merely because one CISC implimenter is adding beacoup instructions to their chips. Extending your instructions is a good way for people to get high level optimization out of a processor. If certain functions that are popular among an entire class of products are turned into chip level instructions, anyone writing code in that catagory gets to replace a large chunk of code with a single instruction. This is good for programmers as they have finite time to complete a project and writing/debugging a chuck of code takes alot longer than using some hardware optimization.
I think two very valuable architectural features would be 1) a well defined CORBA (insert more component interface protocols here) interface. It would be a really nice thing to be able to include features in the server program itself directly into office apps connected to it. If you've got a well structured interface I can write a small collaboration app that uses the DB parser of your server (and the server's processing power) to find messages relevant to my project. Secondly I think a good interface to web servers is also pretty important. I mean sure you can have CGI or some such communicate through abstraction (and extra processes) to the calendar server. Let Apache talk directly to the server and make queries from an apache module.
I don't know if this goes out of the bounds of a calendar server but I think a very good addition would be either collaboration functions (a la Domino) or just the ability to store information about your projects in the calendar as relational objects. Being able to stick a new project on my calendar and then store meetings, trips, ect in my calendar and have them related to the project I'm working on. Collaboration features are also a big plus in the office, especially if it's all OS so you can add the ability to link adendum notes or comments to a document that reside on an intranet server. This feature set can also be referenced to your calendar so you can keep track of everything in a related way. Above all else it ought to be able to make better coffee than the secretary.
It isn't Intel marketing fooling anybody. Its the fact that for some applications you really need more processors and a larger cache. Sun sells SPARC boxes with enormous cache sizes for a reason. If you're running a high powered web server with an efficiently compiled binary you can fit most if not all of the most frequently used instructions in the cache making CPU-to-memory retreival unimpoortant and making the full chip speed cache a big advantage. Big fast cache == benefit in performace but hinderance in cost.
Fighting over the newest and most bestest Pentium based system is getting a bit old. The P4 is DIFFERENT from the P3 it isn't merely a core modification like the P3 was to the P2 and the P2 was to the PPro. The P4 is something Intel wants to promote as a real solution until they are able to roll out IA64 in the mainstream. If it costs too much or doesn't do what you want, don't buy the fucking thing. I'm not going to buy one. Let the OEMs buy a shitload of them and package them in consumer PCs. In terms of actual core quality, I'd vie for a Xeon over a regular P3 or P4. I get the option of lots of memory, better pipelining, and better branch prediction. Along with that I get the option of an enormous L2 cache which is very important when you're using the same instructions over and over, say when I'm doing 3D rendering or have a web server running; it's also going to increase performance on large numbers of non-repeating instructions (i.e. a large number of individual apps running all doing their own thing, say on a large multi-user system). The P4 is in the same situation the K6-3 was, AMD haters hoped it would kill AMD while supports defended it as more of a testbed than anything else. The P4 is a production test of some new approaches to things. It won't sell well in retail sales but I bet OEMs will eat it up because their customers will want it. The Duron many of you bought is a combination of technologies used in the K6-3 and Athlon, if the K6-3 had never been released the Duron might cost you a lot more due to the fact that some techniques were untried which means unrefined. In production cases, anything you try to do without first perfecting will cost you alot of money.
Intel's offers neigh hold a candle to Sun's offerings. An E10000 can have 64 processors and 64 gigs of RAM (which is SDRAM by the way). P3s max out at two processors and ~6.8 gigs of RAM (IIRC) but due to practical hardware limitations usually max out at 4 gigs. Xeons on the other hand can page as much memory as SPARC chips but the GX chipset limits the number of processors to 8 meaning you have to build a cluster in order to get 64 chips. The E10000 scales to 64 processors with no clustering which is an inherent speed advantage. There are workarounds for everything of course, do theoretically you CAN hack together some P3 or P4 boxes to handle huge databases and do a bunch of fancy shit but it will still cost you alot of money.
Dude, don't even start. HTTP knows what sort of files it is sending out, where those files are going to, and who is downloading them. If 20% of your downloads are audio/mpeg rather than text/html they can run a parser over the log and find out. They can also tell when image/jpeg and image/gifs are downloaded in conjunction with text/html pages. This is not something an ISP won't be offended by. But if you've got a bunch of files archived on the server and have one HTML page that links to them all, it is in the ISP's interest to more closely examine your web logs.
Consider me this, Adobe gets pissed that Photoshop is one of the most pirated pieces of software on warez sites, they file an injunction with some major ISPs and hosts that says they ought to scan and/or delete files named ps6*.arj or.zip. Faced with an inunction the ISPs and hosts are going to have to do this or else face a legal battle that many web-based companies can neigh afford. If you're a CEO of a company that is in this sort of situation you would be hard pressed to satisfy a small handful of users with legitimate MP3s. You can either lose a handful of customers or get your company embroiled in a legal battle where you might end up losing your entire company. Hmmm, handful of users buying commodity bandwidth or going out of business...hmm indeed. You can get new users to replace the handful you lost, you cannot (or rarely) however get a management position if you ran a company to the ground.
If you are unhappy with the service provided by your/an/someone's ISP pay a little money for a co-location. Jesus, you're bitching about an ISP called Half-Priced Hosting. Thats just poetic justice not the heavy hand of totalinarianism. When you read your TOS with an ISP put yourself in the shoes of said ISP and then you'll realize that the stuff you have qualms about as a customer are things you yourself would put in a TOS to cover your technogeek ass. You fucking GPL commies think you own absolutely everything but luckily you don't. Because an ISP deletes MP3s doesn't mean you need to get out your pitchforks because they think it's illegal. ISPs own the hard drives and computers those files are hosted on you're merely renting that space from them. Let say Metallica decides they are going to sue anyone carrying a Metallica MP3, the ISP is now liable for your content. Its like me stashing my pot in your house and you getting caught with it. If you ran an ISP (which has an abyssmally slender profit margin) you usually can't afford to fight some legal battle or pay some fuckers damn royalty fees. Communists seem to think that because they work only in GroupThink that a corporation is a single thing which is evil. A corporation comprises of all levels of workers from secretaries to VPs, if an ISP has to pay 10% of its yearly earnings in legal fees and then loses and has to pay another 40% thats some people getting pink slips or investors having their portfolios crash. You might be one of those investors who just bought some stock with E-Trade by the way.
So you ask "Well what can I do oh sultan of sarcasm who has little patience for our GPL communist GroupThink RMS clone ideals?" Go to a hosting company and get a co-location box. Don't rent or lease this box though, build it yourself and have them plug it into their network. In this situation you're no longer renting the host's hardware, merely bandwidth. Hosts usually don't give a fuck what is inside the packets flooding out from their network they just want to make surer you're paying for the bytes you're transfering. Or even better, pay for your own fucking dedicated connection and run a host from your house. If you do either of these things you get to write your own damn TOS and then you can stop whining and try to get Apache running on Hurd.
The difference between your usage of your cable modem and the usage they take a dislike to is that you're not getting a DNS entry for your computer and running some huge server. They don't want you registering sanemind.org and running a webserver hundreds of people are going to access. You're paying 50$ for your cable modem but they are paying many thousands of dollars to keep your city hooked up with enough bandwidth to satisfy all of their customers of which you are but one of many. Many early bouts of cable modem service resulted in people paying 50$ for T1 speeds and running websites off of their home machines. If you download 1bigfile.tar.gz that might take an hour and be of little concequence but if everyone and their mother is downloading 1bigfile.tar.gz from you that is a tax on their service. Besides all of that justification, you don't own the fucking cable line or the network connections going anywhere. You can't make demands on their service like you're some Machiavellian ruler. You're paying for a service as dictacted by a contract you accepted. You are legally liable for the upholding of said contract. You can't get your panties in an uproar over a contract you're supposed to have read.
What do you think HTTP already does? You store (or link) the files you want people to access in your www directory. Your server receives a HTTP request looking for a file, the HTTP server either finds it and sends it back or doesn't find it and tells the person so. You can transfer anything you want over HTTP and it allows for contextual information to be transfered (size and MIME type) of the object requested. DNS doesn't touch your system's internal components it just helps computers find each other whcih is all it ought to do.
The handle approach wants to extend and replace URLs. Right now you type a URL into your browser and it goes to a DNS server and places a query. The DNS server matches up slashdot.org with an IP address which your browser sends an HTTP request to. The slashdot server then finds the file and sends it to you. With the handle approach the file and address are stored in a directory so you type in slashdot.org.index and your browser authomatically goes to the index file on the slashdot server. That directory entry to slashdot.org.index is dynamic though, if the file moves to a different computer or server the name still points to it.
Thats what the dudes who came up with EJB wanted to do. Objects pass around the net and the person accessing said object only becomes aware if there is some response lag. Lets say instead of just a reference to a static web page you accessed a web page object. You could then pass that object onto someone else. But you could also substitute the object's contents with your own content that someone will access. Even with a powerful encryption and validation system in place you could still spoof the contents of an object. You could also make the contents of an object interactive so you could make a BO-type spoofed object that totally fucks someone over if it is ever accessed. The host-based internet sidesteps some of the security problems of an all peer internet by centralizing files and objects. Besides, I don't want to share my comparitively paltry dialup bandwidth with some other dude because I have a web page object on my system he wants to see.
It seems to me this sort of discussion has been handled many times with the same results. Object descriptions and addresses ought to remain separate. DOI looks like a big directory structure for the net; your objects be they computers, printers, or individual files are given handles which are are then in turn given directory registrations. Am I following so far? It seems like this is just restructuring overhead without making it particularly more efficient or effective. TCP/IP packets can be run through a stack which pretty quickly gives the receiver information about the packet but leaves the content alone. This is very simple and amorphous which is why it caught on (you can even use different routing/addressing schemes as long as it follows the header-has-little-to-do-with-the-packet concept). Directory structures on the other hand need alot of overhead due to the fact something somewhere has to know exactly where something is. Lets say all of the DNS servers around today had to hold references for every file available on the internet. That is amazing overhead just to access a text file on a server someplace. Overhead that is distributed over the WHOLE network (the entire internet) as you've only got so many directory servers you can possibly access. TCP/IP combined with transfers that overhead to the computers that are actually talking rather than the entire network. Its easy to upgrade the speed of your hardware to handle an increased demand or whatnot which is generating the extra overhead but is truely hard to squeeze more umph out of a network that is forced to access a limited number of nodes to do absolutely everything.
Re:Look back into the history of the Mac...
on
No Love For Darwin?
·
· Score: 2
Making development teams compete makes Jobs a jackass? Hmm it seems to me the people working for Apple are not only there voluntarily but they are getting paid regardless. Around the same time you had Microsoft's one development team trying like mad to get a GUI (read Xerox clone) out of a compiler and into a product box. The Macintosh made a huge splash when it was released and really did its part to advance personal computing, in ways moreso than Windows.
AmigaDE is a completely different beast. With AmigaDE Amiga does all the hard work for you, they build a bunch of libraries and a runtime interpreter for different architectures under a single API. You then write apps for DE which get compiled to a P-code sort of state, they will not run without the Amiga runtime. You can load up the runtime from any OS and then run your Amiga apps natively on Linux or Win32 or MIPS-64 or PPC or any other microprocessor.
It always amuses me to see the non-newbie Linux zealots get their panties in an uproar if there is ever any news about Microsoft. Yes Whistler has an option to only let certified applications be run by users or user groups. Big fucking deal. If you're a serious Unix admin you spend hours configuring your system so foreign binaries don't get executed on your box and fuck everything up. Certificates are the next logical step in system security. Say I have a bunch of Whistler workstations and I write a cool VB script to automate some tasks and keep the systems in top shape even when I'm not in the office. Wait a second, it's a bad idea to allow VB scripts to run on your system because of malicious code. Solution, sign MY script with a signature (anything I can import, not just VeriSign) and allow it to run unfettered because it is signed by me. When you write a shell script do you not make sure it runs with an SUID but can't be executed by unprivilaged users? Oh yeah huh. Microsoft with Whistler is going to start integration of the.NET concept, security is going to be an enormous concern in this area. NextBestThing.exe fires up and uses some ActiveX components somewhere on the internet if there is no handshaking and authorization going on in this transaction you've just opened a huge security hole in your system and network. The EJB 1.0 spec has a problem of this sort that crops up. It supports authorization as you can have a user have access to a certain set of objects but beyond that it is up to the developer how to impliment more security. If I was planning to start offering applications that resided or executed remotely I would want some proof they were the real shebang. I'd actually like to see a good certificate implimentation on Linux, especially in a professional environment where you need real reliability and authorization, not just strong crypto on your pr0n and mp3s.
finally ventured away from the PentiumPro core! Yay Intel, you go. The concept of SSE2 is interesting, I'm still waiting for software using SSE other than a small handful of Photoshop filters and a 3D model of the Soalr System. I guess to be a wet blanket I'll complain about lack of innovative architecture. The P4 and Athlon haven't inspired me at all. Wow more powerful scalar operations yay Intel, you go. Why hasn't anyone building x86 chips stuck a vector unit in one yet? Is there some architectural design limit I'm missing? Whatever happened to the insta-recompile code morphing the P4 was supposed to have (I haven't read all the articles posted yet, only Sharkey's)? Since they lowered the IPC and lock you into a go-nowhere processor line with extraordinarily priced memory and hardware who exactly is the tarket sucke..market for the P4?
the other guy said, is wireless broadband (or even broadband) the way to go? An ISDN or leased line can be had for a decent price and are usually pretty available. You can maximize bandwidth by putting a NAT server between you and the internet with cache'ing set so you only need to download the bare minimum of a page's components. Another option is to use something like DirecPC which has a fair downstream bandwidth but is limited to your modem upload speed (and severe latency). The drawbacks to all of these is a high initial cost (networking equipment and wiring along with ISDN/Leased line charges) but I think it is still more affordable than wireless broadband. Wireless just isn't "here" in quality and quantity that would make it a real option in my opinion.
I've found the best way to plan for future upgrades is to overcompensate, at least a little bit. Fiber and Wireless stuff is hip now but not really worth the overall cost you're incurring. Wireless is pretty unreliable and is exceptionally limited in bandwidth and overall functionality. Something like AirPort can only have a relatively small number of users and has a fraction of the bandwidth of wired options. Fiber is robust but is still an expensive proposition that doesn't have enough benefits (in your case) to merit use. So that leaves copper wiring. With Cat-5 wiring you have the option of a 10, 100 or 1000Mb network which allows you to save some money now with 100Mb equipment and then upgrade to 1000Mb when you really need it (and when it is a bit cheaper).
After you've got the wire run you need to manage it properly to get the most bang for your buck. I've seen pretty large networks with expensive equipment bottleneck because the original designers didn't plan things out well enough. Put a switch on every floor with computers on it, having a switched network makes peer-to-peer communication faster and more efficient. Connect the switches and any servers all into a set of central routers. Good routers are going to balance the load of the network and connect you to the outside world. Do not merely put powered hubs on every floor and connect them to a single server that is supposed to do routing in software and manage your network. Switches are going to speed up peer-to-peer stuff which is increasingly replacing client-server transaction (although servers are still needed) and plugging those switches into routers is going to let the routers balace the load of the entire network and help out client-server communication.
Aside from copper networking wire and the like, run some audio and video wire and leave plenty of open space for more wiring. I really wish my school has had the forsight a while ago to add the A/V cabling. A few cables inserted into labeled outlets and you've got a projector hooked up to a workstation in another room. Run some regular RJ-11 wire for a decent PBX, while it may not be glamorous the faculty will enjoy the ability to use a phone from a lab or lecture hall rather from only in their offices.
Well I hope that is at least helpful, if someone wants to point out errors on my part or suggestions feel free. Just don't flame.
Probes ARE reprogrammable, the problem with the Mars probe that ate atmosphere was a miscalculation of an orbital trajectory. Having a server you could update to wouldn't solve shit if your math is wrong. An internet server on a decive with a very small range communication channel would be stupid. That sort of extra overhead wastes valuable processor cycles that could be better used for telemetry or fine motor control. Why would you need to "open-source" analytical readings? A few letters and what-not and you can get access to lots of information from various probes over the past 50 years for "independant analysis".
Slashdot Mirror
Non-linear editing takes as much time if not more as old fashioned film painting. Even with digital video you've got to composite five layers of video, whether it be film or bytes. To do production quality digital editing you need a very expensive facility. Remember 35mm film is REALLY high resolution, so everything needs to be editing composed and rendered at that resolution. That takes forever even with really powerful computers. Digital equipment is initially expensive, very very expensive. Once it saturates sure you have your money's worth and can reuse a good deal of it but regular film can be alot cheaper because you can go back to cut and paste editing if need be not so with digital. Digital projectors are expensive as can be because theaters need to rip out all their old equipment. If you buy a newer film camaera you can still use your $75k light housing for it. Do some research rather than getting evangelical due to other people's posts.
One of the staed goals of the P4 though is to really whomp ass in terms of clock speed which is something I think it can be agreed thay've done. They needed to put something out that could do two things 1) make the home PC buying drool and drivel over the word gigahertz and 2) do things people wanted to do quickly. If you have a processor that really flies at video rendering and sound processing Dell and Gateway stick FireWire ports in it and classify it a multimedia editing PC to get eyes turned away from iMacs. Non-techies see the numerical difference between 1.2 and 1.5GHz and their eyesbrows go up because they understand that the higher number MUST be better. You're very correct though in saying render farms of these things will sink anyone. raytracing and picture editing use algorithms over and over again and are best kept in the cache until served with a cold soup. Coppermines and Willamettes have tiny caches that make it difficult to keep more complex (read larger) algorithms in the L2. When you buy an SGI Onyx system the processors have pretty heavy cache sizes which facilitate repeticious functions like this. If businesses want/need "high-performace" they ought to be buying Xeons if they want to stick with IA32 or maybe the UltraSPARC3. Quality hardware + quality software == justifiable cost.
Dude, what the fuck good would a "Linux" FPGA do for anyone? Ohhh, I get DSP performance out of my kernel? Big fucking deal. Its the 3D rendering or Quake3 I want DSP performance out of. And more to the guts of your post. Most modern RISC processors out now have a large amount of specialization either in their instructions or processing units. MIPS binaries don't run on SPARCs now do they? RISC is a good chip architecture but it is no reason to thrash CISC merely because one CISC implimenter is adding beacoup instructions to their chips. Extending your instructions is a good way for people to get high level optimization out of a processor. If certain functions that are popular among an entire class of products are turned into chip level instructions, anyone writing code in that catagory gets to replace a large chunk of code with a single instruction. This is good for programmers as they have finite time to complete a project and writing/debugging a chuck of code takes alot longer than using some hardware optimization.
Does you opinion affect tides and harvests also?
I think two very valuable architectural features would be 1) a well defined CORBA (insert more component interface protocols here) interface. It would be a really nice thing to be able to include features in the server program itself directly into office apps connected to it. If you've got a well structured interface I can write a small collaboration app that uses the DB parser of your server (and the server's processing power) to find messages relevant to my project. Secondly I think a good interface to web servers is also pretty important. I mean sure you can have CGI or some such communicate through abstraction (and extra processes) to the calendar server. Let Apache talk directly to the server and make queries from an apache module.
I don't know if this goes out of the bounds of a calendar server but I think a very good addition would be either collaboration functions (a la Domino) or just the ability to store information about your projects in the calendar as relational objects. Being able to stick a new project on my calendar and then store meetings, trips, ect in my calendar and have them related to the project I'm working on. Collaboration features are also a big plus in the office, especially if it's all OS so you can add the ability to link adendum notes or comments to a document that reside on an intranet server. This feature set can also be referenced to your calendar so you can keep track of everything in a related way. Above all else it ought to be able to make better coffee than the secretary.
It isn't Intel marketing fooling anybody. Its the fact that for some applications you really need more processors and a larger cache. Sun sells SPARC boxes with enormous cache sizes for a reason. If you're running a high powered web server with an efficiently compiled binary you can fit most if not all of the most frequently used instructions in the cache making CPU-to-memory retreival unimpoortant and making the full chip speed cache a big advantage. Big fast cache == benefit in performace but hinderance in cost.
Fighting over the newest and most bestest Pentium based system is getting a bit old. The P4 is DIFFERENT from the P3 it isn't merely a core modification like the P3 was to the P2 and the P2 was to the PPro. The P4 is something Intel wants to promote as a real solution until they are able to roll out IA64 in the mainstream. If it costs too much or doesn't do what you want, don't buy the fucking thing. I'm not going to buy one. Let the OEMs buy a shitload of them and package them in consumer PCs. In terms of actual core quality, I'd vie for a Xeon over a regular P3 or P4. I get the option of lots of memory, better pipelining, and better branch prediction. Along with that I get the option of an enormous L2 cache which is very important when you're using the same instructions over and over, say when I'm doing 3D rendering or have a web server running; it's also going to increase performance on large numbers of non-repeating instructions (i.e. a large number of individual apps running all doing their own thing, say on a large multi-user system). The P4 is in the same situation the K6-3 was, AMD haters hoped it would kill AMD while supports defended it as more of a testbed than anything else. The P4 is a production test of some new approaches to things. It won't sell well in retail sales but I bet OEMs will eat it up because their customers will want it. The Duron many of you bought is a combination of technologies used in the K6-3 and Athlon, if the K6-3 had never been released the Duron might cost you a lot more due to the fact that some techniques were untried which means unrefined. In production cases, anything you try to do without first perfecting will cost you alot of money.
Intel's offers neigh hold a candle to Sun's offerings. An E10000 can have 64 processors and 64 gigs of RAM (which is SDRAM by the way). P3s max out at two processors and ~6.8 gigs of RAM (IIRC) but due to practical hardware limitations usually max out at 4 gigs. Xeons on the other hand can page as much memory as SPARC chips but the GX chipset limits the number of processors to 8 meaning you have to build a cluster in order to get 64 chips. The E10000 scales to 64 processors with no clustering which is an inherent speed advantage. There are workarounds for everything of course, do theoretically you CAN hack together some P3 or P4 boxes to handle huge databases and do a bunch of fancy shit but it will still cost you alot of money.
Dude, don't even start. HTTP knows what sort of files it is sending out, where those files are going to, and who is downloading them. If 20% of your downloads are audio/mpeg rather than text/html they can run a parser over the log and find out. They can also tell when image/jpeg and image/gifs are downloaded in conjunction with text/html pages. This is not something an ISP won't be offended by. But if you've got a bunch of files archived on the server and have one HTML page that links to them all, it is in the ISP's interest to more closely examine your web logs.
Consider me this, Adobe gets pissed that Photoshop is one of the most pirated pieces of software on warez sites, they file an injunction with some major ISPs and hosts that says they ought to scan and/or delete files named ps6*.arj or .zip. Faced with an inunction the ISPs and hosts are going to have to do this or else face a legal battle that many web-based companies can neigh afford. If you're a CEO of a company that is in this sort of situation you would be hard pressed to satisfy a small handful of users with legitimate MP3s. You can either lose a handful of customers or get your company embroiled in a legal battle where you might end up losing your entire company. Hmmm, handful of users buying commodity bandwidth or going out of business...hmm indeed. You can get new users to replace the handful you lost, you cannot (or rarely) however get a management position if you ran a company to the ground.
If you are unhappy with the service provided by your/an/someone's ISP pay a little money for a co-location. Jesus, you're bitching about an ISP called Half-Priced Hosting. Thats just poetic justice not the heavy hand of totalinarianism. When you read your TOS with an ISP put yourself in the shoes of said ISP and then you'll realize that the stuff you have qualms about as a customer are things you yourself would put in a TOS to cover your technogeek ass. You fucking GPL commies think you own absolutely everything but luckily you don't. Because an ISP deletes MP3s doesn't mean you need to get out your pitchforks because they think it's illegal. ISPs own the hard drives and computers those files are hosted on you're merely renting that space from them. Let say Metallica decides they are going to sue anyone carrying a Metallica MP3, the ISP is now liable for your content. Its like me stashing my pot in your house and you getting caught with it. If you ran an ISP (which has an abyssmally slender profit margin) you usually can't afford to fight some legal battle or pay some fuckers damn royalty fees. Communists seem to think that because they work only in GroupThink that a corporation is a single thing which is evil. A corporation comprises of all levels of workers from secretaries to VPs, if an ISP has to pay 10% of its yearly earnings in legal fees and then loses and has to pay another 40% thats some people getting pink slips or investors having their portfolios crash. You might be one of those investors who just bought some stock with E-Trade by the way.
So you ask "Well what can I do oh sultan of sarcasm who has little patience for our GPL communist GroupThink RMS clone ideals?" Go to a hosting company and get a co-location box. Don't rent or lease this box though, build it yourself and have them plug it into their network. In this situation you're no longer renting the host's hardware, merely bandwidth. Hosts usually don't give a fuck what is inside the packets flooding out from their network they just want to make surer you're paying for the bytes you're transfering. Or even better, pay for your own fucking dedicated connection and run a host from your house. If you do either of these things you get to write your own damn TOS and then you can stop whining and try to get Apache running on Hurd.
The difference between your usage of your cable modem and the usage they take a dislike to is that you're not getting a DNS entry for your computer and running some huge server. They don't want you registering sanemind.org and running a webserver hundreds of people are going to access. You're paying 50$ for your cable modem but they are paying many thousands of dollars to keep your city hooked up with enough bandwidth to satisfy all of their customers of which you are but one of many. Many early bouts of cable modem service resulted in people paying 50$ for T1 speeds and running websites off of their home machines. If you download 1bigfile.tar.gz that might take an hour and be of little concequence but if everyone and their mother is downloading 1bigfile.tar.gz from you that is a tax on their service. Besides all of that justification, you don't own the fucking cable line or the network connections going anywhere. You can't make demands on their service like you're some Machiavellian ruler. You're paying for a service as dictacted by a contract you accepted. You are legally liable for the upholding of said contract. You can't get your panties in an uproar over a contract you're supposed to have read.
What do you think HTTP already does? You store (or link) the files you want people to access in your www directory. Your server receives a HTTP request looking for a file, the HTTP server either finds it and sends it back or doesn't find it and tells the person so. You can transfer anything you want over HTTP and it allows for contextual information to be transfered (size and MIME type) of the object requested. DNS doesn't touch your system's internal components it just helps computers find each other whcih is all it ought to do.
The handle approach wants to extend and replace URLs. Right now you type a URL into your browser and it goes to a DNS server and places a query. The DNS server matches up slashdot.org with an IP address which your browser sends an HTTP request to. The slashdot server then finds the file and sends it to you. With the handle approach the file and address are stored in a directory so you type in slashdot.org.index and your browser authomatically goes to the index file on the slashdot server. That directory entry to slashdot.org.index is dynamic though, if the file moves to a different computer or server the name still points to it.
Thats what the dudes who came up with EJB wanted to do. Objects pass around the net and the person accessing said object only becomes aware if there is some response lag. Lets say instead of just a reference to a static web page you accessed a web page object. You could then pass that object onto someone else. But you could also substitute the object's contents with your own content that someone will access. Even with a powerful encryption and validation system in place you could still spoof the contents of an object. You could also make the contents of an object interactive so you could make a BO-type spoofed object that totally fucks someone over if it is ever accessed. The host-based internet sidesteps some of the security problems of an all peer internet by centralizing files and objects. Besides, I don't want to share my comparitively paltry dialup bandwidth with some other dude because I have a web page object on my system he wants to see.
It seems to me this sort of discussion has been handled many times with the same results. Object descriptions and addresses ought to remain separate. DOI looks like a big directory structure for the net; your objects be they computers, printers, or individual files are given handles which are are then in turn given directory registrations. Am I following so far? It seems like this is just restructuring overhead without making it particularly more efficient or effective. TCP/IP packets can be run through a stack which pretty quickly gives the receiver information about the packet but leaves the content alone. This is very simple and amorphous which is why it caught on (you can even use different routing/addressing schemes as long as it follows the header-has-little-to-do-with-the-packet concept). Directory structures on the other hand need alot of overhead due to the fact something somewhere has to know exactly where something is. Lets say all of the DNS servers around today had to hold references for every file available on the internet. That is amazing overhead just to access a text file on a server someplace. Overhead that is distributed over the WHOLE network (the entire internet) as you've only got so many directory servers you can possibly access. TCP/IP combined with transfers that overhead to the computers that are actually talking rather than the entire network. Its easy to upgrade the speed of your hardware to handle an increased demand or whatnot which is generating the extra overhead but is truely hard to squeeze more umph out of a network that is forced to access a limited number of nodes to do absolutely everything.
Making development teams compete makes Jobs a jackass? Hmm it seems to me the people working for Apple are not only there voluntarily but they are getting paid regardless. Around the same time you had Microsoft's one development team trying like mad to get a GUI (read Xerox clone) out of a compiler and into a product box. The Macintosh made a huge splash when it was released and really did its part to advance personal computing, in ways moreso than Windows.
AmigaDE is a completely different beast. With AmigaDE Amiga does all the hard work for you, they build a bunch of libraries and a runtime interpreter for different architectures under a single API. You then write apps for DE which get compiled to a P-code sort of state, they will not run without the Amiga runtime. You can load up the runtime from any OS and then run your Amiga apps natively on Linux or Win32 or MIPS-64 or PPC or any other microprocessor.
It always amuses me to see the non-newbie Linux zealots get their panties in an uproar if there is ever any news about Microsoft. Yes Whistler has an option to only let certified applications be run by users or user groups. Big fucking deal. If you're a serious Unix admin you spend hours configuring your system so foreign binaries don't get executed on your box and fuck everything up. Certificates are the next logical step in system security. Say I have a bunch of Whistler workstations and I write a cool VB script to automate some tasks and keep the systems in top shape even when I'm not in the office. Wait a second, it's a bad idea to allow VB scripts to run on your system because of malicious code. Solution, sign MY script with a signature (anything I can import, not just VeriSign) and allow it to run unfettered because it is signed by me. When you write a shell script do you not make sure it runs with an SUID but can't be executed by unprivilaged users? Oh yeah huh. Microsoft with Whistler is going to start integration of the .NET concept, security is going to be an enormous concern in this area. NextBestThing.exe fires up and uses some ActiveX components somewhere on the internet if there is no handshaking and authorization going on in this transaction you've just opened a huge security hole in your system and network. The EJB 1.0 spec has a problem of this sort that crops up. It supports authorization as you can have a user have access to a certain set of objects but beyond that it is up to the developer how to impliment more security. If I was planning to start offering applications that resided or executed remotely I would want some proof they were the real shebang. I'd actually like to see a good certificate implimentation on Linux, especially in a professional environment where you need real reliability and authorization, not just strong crypto on your pr0n and mp3s.
finally ventured away from the PentiumPro core! Yay Intel, you go. The concept of SSE2 is interesting, I'm still waiting for software using SSE other than a small handful of Photoshop filters and a 3D model of the Soalr System. I guess to be a wet blanket I'll complain about lack of innovative architecture. The P4 and Athlon haven't inspired me at all. Wow more powerful scalar operations yay Intel, you go. Why hasn't anyone building x86 chips stuck a vector unit in one yet? Is there some architectural design limit I'm missing? Whatever happened to the insta-recompile code morphing the P4 was supposed to have (I haven't read all the articles posted yet, only Sharkey's)? Since they lowered the IPC and lock you into a go-nowhere processor line with extraordinarily priced memory and hardware who exactly is the tarket sucke..market for the P4?
Thanks for the tip, I'll have to remember that one myself.
the other guy said, is wireless broadband (or even broadband) the way to go? An ISDN or leased line can be had for a decent price and are usually pretty available. You can maximize bandwidth by putting a NAT server between you and the internet with cache'ing set so you only need to download the bare minimum of a page's components. Another option is to use something like DirecPC which has a fair downstream bandwidth but is limited to your modem upload speed (and severe latency). The drawbacks to all of these is a high initial cost (networking equipment and wiring along with ISDN/Leased line charges) but I think it is still more affordable than wireless broadband. Wireless just isn't "here" in quality and quantity that would make it a real option in my opinion.
I've found the best way to plan for future upgrades is to overcompensate, at least a little bit. Fiber and Wireless stuff is hip now but not really worth the overall cost you're incurring. Wireless is pretty unreliable and is exceptionally limited in bandwidth and overall functionality. Something like AirPort can only have a relatively small number of users and has a fraction of the bandwidth of wired options. Fiber is robust but is still an expensive proposition that doesn't have enough benefits (in your case) to merit use. So that leaves copper wiring. With Cat-5 wiring you have the option of a 10, 100 or 1000Mb network which allows you to save some money now with 100Mb equipment and then upgrade to 1000Mb when you really need it (and when it is a bit cheaper).
After you've got the wire run you need to manage it properly to get the most bang for your buck. I've seen pretty large networks with expensive equipment bottleneck because the original designers didn't plan things out well enough. Put a switch on every floor with computers on it, having a switched network makes peer-to-peer communication faster and more efficient. Connect the switches and any servers all into a set of central routers. Good routers are going to balance the load of the network and connect you to the outside world. Do not merely put powered hubs on every floor and connect them to a single server that is supposed to do routing in software and manage your network. Switches are going to speed up peer-to-peer stuff which is increasingly replacing client-server transaction (although servers are still needed) and plugging those switches into routers is going to let the routers balace the load of the entire network and help out client-server communication.
Aside from copper networking wire and the like, run some audio and video wire and leave plenty of open space for more wiring. I really wish my school has had the forsight a while ago to add the A/V cabling. A few cables inserted into labeled outlets and you've got a projector hooked up to a workstation in another room. Run some regular RJ-11 wire for a decent PBX, while it may not be glamorous the faculty will enjoy the ability to use a phone from a lab or lecture hall rather from only in their offices.
Well I hope that is at least helpful, if someone wants to point out errors on my part or suggestions feel free. Just don't flame.
Probes ARE reprogrammable, the problem with the Mars probe that ate atmosphere was a miscalculation of an orbital trajectory. Having a server you could update to wouldn't solve shit if your math is wrong. An internet server on a decive with a very small range communication channel would be stupid. That sort of extra overhead wastes valuable processor cycles that could be better used for telemetry or fine motor control. Why would you need to "open-source" analytical readings? A few letters and what-not and you can get access to lots of information from various probes over the past 50 years for "independant analysis".
Just ask Australia.