If something like Xen can switch OSes between OSX and Windows, the new mac might be a great package for people who need to run win32 as well. OSX could even do API translation like in WINEX to run win32 apps at near native speeds.
Like Be and RedHat, Apple is the new OS vendor out there. Be ran on 2 platforms and is dead now. RedHat runs on 3 (or more) platforms and has big community backing. Apple does too. They both have good application base, although Apple has more on one platform, the PPC.
Some say Apple has a good OS, so they'll have success. Others say their Intel hardware will be superior and people will buy more of it, since it will be cheaper and efficient, so Apple will be successful.
And yet for some reason, I'm also pessimistic here.
We had the evil wintel. And then we had the Apple, motorolla, IBM alliance. IBM is very busy pushing Linux-on-PowerPC, which means that hardware platform will have a future, and might just pull ahead of x86.
However, the AMD64 platform showed that the x86/x64 platform is the best thing out there and Apple is too moving to it. Less diversity. Just a bunch of OSes on the same chip on roughly the same motherboard (since the mem handler is built into the chip, theres less else on the AMD64 mobo). Thats now the entire desktop market of the world.
There was once a time when we had IRIX on MIPS, OpenVMS and Tru64 on Alpha and VMS, Solaris on Ultrasparc, HPUX on PARISC, Unixware on Intel, OS2, and all the BSDs plus Linux out there. It was a rich world. Lots to learn. Each one had a strength you could count on. All thats collapsed, Be was bought out, SCO was too, Alpha, Tru64, OpenVMS were too, Ultrasparc and Itanium and PARISC are dying, MIPS is dead, OS2 is dead, the diverse mainframes are dead, and we're seeing even more industry consolidation, and later the demise of some of the companies who couldnt differentiate enough.
I suppose I'll feel different when I'll see a cheaper macmini with an Athlon64 FX55 (or equiv) running OSX.
OSX had better be able to make me buy the whole deal now.
I personally think the analog is quite important, from experience. Many a times in my digital experiments, things didnt work for analog reasons.. busses too long, wrong metals, interference, voltage stuff, and it becomes more interesting at higher speeds and in DSPs and DACs and ADCs.
Even for the PICs they direct you to use RC oscillators, and that can have consequences from the change of temp, part inaccuracy etc.
Youre right about not buying stuff retail. I bought a soldering iron from radioshack and regretted it. For passive parts, its best to buy assorted bulks from active, jameco and digi* etc. Later on, more advanced parts become harder to buy in small quantities, like the EP7312, and they dont come in DIP format either. Thats when you know youre an engineer.
Dont know about books, I cant learn from books. I just take the equipment, manuals, and aim to build something and read along whatever I need to know as I go along. I setup pretty ambition projects some of which can be months, but it feels more interesting than books.
But thats me.
I cant come up with many books beside PDF datasheets of the devices, sample circuits online and in popular electronics, and quite possibly the textbooks of electronics courses of reputable universities like MIT.
For one, you should know about passive analog electronics. Build an AM radio or something.
Next step, is to go digital. Buy a couple of PIC or AVR microcontrollers and build some simple stuff. You'll get to write assembly code (or even in C) and upload the code to the chip and run it there. The pic can be interfaced with ethernet, audio chips, flash chips, LCD, camera CCD etc. Think of the possibilities.
If the PIC is tough, just use the simplest PIC16F54A initially, or even just use a BASIC stamp. Make a set of blinking lights to begin with, and download the test code first before writing code.
After the 8-bit level, you can buy the powerpc or ARM kits from olimex.com or ebay, and with enough flash, sram and boot code, try to boot netbsd, linux or something similar.
Why do you think the OS cannot benefit from the SPU? What about encryption? What about RAID and filesystem level API? And networking code (imagine the whole networking code of the kernel running on one SPU, filesystem on another SPU, and X with KDE on yet another.
I didnt think the SPU was specialized. They're general chips that are used as specialized chips in the PS3.
I'd like to see a Unisys machine of these with octa-CPUs with DDR3 memory. Next I'd like to see games that were developed with heavy threading, so they can use different processors.
To see how anything can be good or bad for Linux, we'd have to see what is 'good for Linux' anyway.
Something that increases the uptake and usage of Linux is good for Linux. The biggest problem today Linux faces is lack of application market. Well OK the biggest problem is lack of standards, so lack of applications is the second biggest problem. Any slashdotters who are currently on win32 right now know which apps have forced them to stay with win32.... games, ERP system etc.
So why are these apps not ported to Linux?? (1) Linux isnt a big enough market yet (2) Porting is a royal pain.
How can we fix that?
One way to solve this problem is to have diverse platforms. If the market is fragmented between platforms, vendors will be forced to sell different ports of apps. This means more vendors will move to toolkits like QT and wxwindows and opengl which are portable. If youre code is portable, porting to another platform is easy. If your code is based exclusively on VisualC and MFCs, its tough to port.
So anyone porting apps to win32 AND apple, might as well hire one more developer and maintain a Linux port as well. You wont find many apps that are just dual-platform. They're either just win32 (the kind of apps that are killing Linux), or they're released under at least 3 ports, win32, mac and Linux. In many cases the company says what the heck, and releases BeOS, FreeBSD, OS2, Solaris and other ports too, since porting further is easy.
Apple forces the move from the first type of apps to the latter type. That brings us all kinds of games and business apps on Linux. Probably the second best thing that can happen to Linux (the first thing is STANDARDS!!!)
Whats so artificial about it? It may not be a 'natural' diamond, but artificial? Like combining sodium and chlorine or burning hydrogen in the lab produces artificial salt or water.
You can manufacture diamonds and swear its real. The gold part, now that would be tough. Although many girls like silver colors and will settle for silver or other shinly substitute. If it is bling bling enough, they'll accept it as proof of your love. Just dont tell her the price.
I think DeBeers crushes all possibilities of manufacturing such diamonds legally in western countries. Thats why I'm interested in this business; I can make an upstart in my native Afghanistan. At least Central and South Asia could break free of the monopoly, no matter what any law says.
It certainly is happening. There are various seasonal cycles, glaciation cycles and others we dont know about because our history doesnt go back far enough. The neocene should be over by now, the next ice age is due. So we have no idea what to expect... its been 12000 years since the end of the pliestocene.
So all that means the Earth was never constant. There were forests in Saudia, a great desert where forests are currently, a lush ecosystem in the middle of the Antarctica. Surely humans didnt change that, we didnt exist. We only very recently became powerful enough to make big changes in the global system, but since its all constantly changing so much, our effected changes are lost. Think of the Tsunami. Could we do that? That was weather dear friend, it showed us how our effects are completely dwarfed by mother Natures'.
I dont think we can make lakes disappear yet. The greatest change we have been able to make is putting dams on rivers, and the heat of cities raising its temperatures by 2 celsius. Apart from that, things have always been changing, and will always be. My hometown used to get 10ft snow every winter 'back in the days', while now it didnt snow for a good 5 years straight. Legend has it that its always been getting warmer there, and such temperatures are to be expected.
The sky IS blue. But mac users certainly arent 16%. I'd be VERY surprised if they were more than 10%. You'll have great trouble finding a mac in most Asian countries; the mac is America-centric.
So here in N America, running into a mac is a rare event. The enormous bulk installations of colleges and companies are PC. Most of them are Dell, HP or IBM. A mac is more like for certain people with certain tastes, a few college libraries and graphic designing classes. Even the libraries I should say are more than 90% PC.
Mac users do spend $$$ on average more than PC users. Thats partly beause they HAVE to, and macs are more expensive anyway. Which means mac users will spend more on other things like software, monitor etc. Mac users are also more vocal; there arent many pro-PC people around.
It makes me wonder why shoot and kill in one location at all? Think of de_dust, which makes more sense... hunting down terrorists in a small village. Although de_dust has practically no houses or even shops or vehicles, and too many crates.
Look around you. There are apartments (great setting for a game, have never seen a complete apartment map, or a good one, there are villages, mountains, and certain places like airports, hospitals, downtown, offices, railway yard etc. The 747 map in CS was original too, but they overused airplanes in other maps.
I wouldnt mind seeing an underground parking lot map. Think of the parking lot scene in terminator2... I always thought that was a great setting for a game lots of glass to break and places to duck.
I was gonna say a school is good too, but I suppose its not.
And if youre gonna make a warehouse, add computer desks, trucks, weighing and wrapping machinery, forklifts and lifttrucks, piles of crates arranged properly to maximize space...
A library would be great if pieces of paper will fly if you shoot the books...
If the mantle is fluid enough and under pressure, wouldnt it gush out and form a volcano? We dont have volcanoes everywhere because the crust is thick enough to hold it down. Punch a hole in the waterbed and viola!
Volcanoes have a knack of closing themselves up, which is partly why this is being done in the ocean. Would be cool to have a build-a-volcano-mountain experiment on the surface somewhere, maybe close to Mt Fiji to augument the current one.
Why cant we just type 'yahoo' and 'google'? Many companies will register (eg mysql.org and mysql.com) multiple TLDs to cover their names online, but they wont have to...
Most people just use.com. Go ahead and register org and net, people will forget and type.com first to get to your site.
They can sell top level domains except 2-character top-level domains which should remain as country TLDs and the likes...
Lotus Domino Business Objects SQL 2000 win32-specific ERP software commercially supported antispam system Active Directory server and remote install server
Domino can run on Linux, BO too, SQL can be replaced with oracle, ERP run in WINEX, antispam replaced and Active Directory replaced with the Novell directory.
But thats way way too much money for little additional gain and the combination is not so well tested.
What OS do you use for the DEC? I havent been able to get Tru64 from anywhere, but efforts have been focused on trying to run OpenVMS. I did try the windows2000 beta with VisualC 6 and Office 2000 apps. Windows2000 is great with 164lx, but crashed a bit, not as stable as NT.
I did however find firefox for NT and ran that whole combination for a little while before returning to the OpenVMS attempts.
Generally, one big process requires the heaviest load, either a game engine, or gcc, or a video encoder, or an unthreaded database engine.
So given the current market, users will likely run a heavy single-threaded 32-bit app on this chip expecting it to be fast. In due time we'll have x64 everything, and if you run apps from good vendors, they'll be properly threaded in a balanced way and will take the maximum advantage of the chip you purchased.
That has more to do with GUI design than sheer power. When you click on an icon, it starts to piece together code from libraries and files everywhere, sometimes before the app window appears. That means disk bandwidth, and the concurrentness of disk reads. If it is read from a RAID 5 array, the program might start much faster than a single 5200RPM IDE disk with the head clicking back and forth while the CPU idles at 2%.
For this reason, OSX is quite snappy, and even WindowsXP is more snappy at starting most apps than the Linux desktops that I've seen. Linux is a general OS while XP is designed to feel snappy at other costs. I dont know where to place OSX
And I have alcohol-based stuff on my hands? I cant drive my car?
Or what if youre helping a drunk friend get home? As soon as you handle him/her, your hands are alcoholic and you cant drive.
Theres no deciding factor about a drunk person except the percentage of alcohol in the bloodstream, so the only reliable possible solution is to have an rfid device under the drivers skin which measures alcohol and relays the data to the car enabling it to start.
Thats right. Portability ensures correctness of code. Thats why WindowsNT was released for 3 platforms, to ensure the code's bugs are ironed out faster (didnt work there much eh?)
Think of gcc and glibc. We get tonnes of issues from all sorts of ports. After they're fixed, the code is really clean. Maybe thats a bad example since they also handle lots of exceptions, but its more true for the likes of X, apache, mozilla, KDE and such.
The same can be said of source code. If it can be compiled by multiple compilers, its true C99 (or something), otherwise if it depends on compiler specific hacks, it wont compile on other compilers, or some compiler will dump errors that others wont.
If you were a graphics card manufacturer, would you just plug it into a development board, and if it works, sell it? Or would you have a slew of systems to test and benchmark the card in before releasing it to the market?
And what happens when another user on your server wants to start another daemon listening to such packets? How do you differentiate two different daemons listening to two different types of traffic, and keep them seperate enough to be safe from one another?
That identifier became 'tcp ports' in tcpip, which is why we dont use ip alone. If a user's daemon is connected to a port, noone else uses that port, and external processes can get to that daemon through that port securely.
The SYN header data will work if the data is small enough, if you screw the TCPIP stack enough and if it is guaranteed that only one daemon will ever need to listen on each given IP address.
Slashdot Mirror
Seriously enough. This might be a great plan.
If something like Xen can switch OSes between OSX and Windows, the new mac might be a great package for people who need to run win32 as well. OSX could even do API translation like in WINEX to run win32 apps at near native speeds.
In that case, I'm buying a mac!
Like Be and RedHat, Apple is the new OS vendor out there. Be ran on 2 platforms and is dead now. RedHat runs on 3 (or more) platforms and has big community backing. Apple does too. They both have good application base, although Apple has more on one platform, the PPC.
Some say Apple has a good OS, so they'll have success. Others say their Intel hardware will be superior and people will buy more of it, since it will be cheaper and efficient, so Apple will be successful.
And yet for some reason, I'm also pessimistic here.
We had the evil wintel. And then we had the Apple, motorolla, IBM alliance. IBM is very busy pushing Linux-on-PowerPC, which means that hardware platform will have a future, and might just pull ahead of x86.
However, the AMD64 platform showed that the x86/x64 platform is the best thing out there and Apple is too moving to it. Less diversity. Just a bunch of OSes on the same chip on roughly the same motherboard (since the mem handler is built into the chip, theres less else on the AMD64 mobo). Thats now the entire desktop market of the world.
There was once a time when we had IRIX on MIPS, OpenVMS and Tru64 on Alpha and VMS, Solaris on Ultrasparc, HPUX on PARISC, Unixware on Intel, OS2, and all the BSDs plus Linux out there. It was a rich world. Lots to learn. Each one had a strength you could count on. All thats collapsed, Be was bought out, SCO was too, Alpha, Tru64, OpenVMS were too, Ultrasparc and Itanium and PARISC are dying, MIPS is dead, OS2 is dead, the diverse mainframes are dead, and we're seeing even more industry consolidation, and later the demise of some of the companies who couldnt differentiate enough.
I suppose I'll feel different when I'll see a cheaper macmini with an Athlon64 FX55 (or equiv) running OSX.
OSX had better be able to make me buy the whole deal now.
I personally think the analog is quite important, from experience. Many a times in my digital experiments, things didnt work for analog reasons.. busses too long, wrong metals, interference, voltage stuff, and it becomes more interesting at higher speeds and in DSPs and DACs and ADCs.
Even for the PICs they direct you to use RC oscillators, and that can have consequences from the change of temp, part inaccuracy etc.
Youre right about not buying stuff retail. I bought a soldering iron from radioshack and regretted it. For passive parts, its best to buy assorted bulks from active, jameco and digi* etc. Later on, more advanced parts become harder to buy in small quantities, like the EP7312, and they dont come in DIP format either. Thats when you know youre an engineer.
Dont know about books, I cant learn from books. I just take the equipment, manuals, and aim to build something and read along whatever I need to know as I go along. I setup pretty ambition projects some of which can be months, but it feels more interesting than books.
But thats me.
I cant come up with many books beside PDF datasheets of the devices, sample circuits online and in popular electronics, and quite possibly the textbooks of electronics courses of reputable universities like MIT.
For one, you should know about passive analog electronics. Build an AM radio or something.
Next step, is to go digital. Buy a couple of PIC or AVR microcontrollers and build some simple stuff. You'll get to write assembly code (or even in C) and upload the code to the chip and run it there. The pic can be interfaced with ethernet, audio chips, flash chips, LCD, camera CCD etc. Think of the possibilities.
If the PIC is tough, just use the simplest PIC16F54A initially, or even just use a BASIC stamp. Make a set of blinking lights to begin with, and download the test code first before writing code.
After the 8-bit level, you can buy the powerpc or ARM kits from olimex.com or ebay, and with enough flash, sram and boot code, try to boot netbsd, linux or something similar.
Thats interesting isnt it? I thought GPL didnt allow it.
By law the coin slot should be broken.
Why do you think the OS cannot benefit from the SPU? What about encryption? What about RAID and filesystem level API? And networking code (imagine the whole networking code of the kernel running on one SPU, filesystem on another SPU, and X with KDE on yet another.
I didnt think the SPU was specialized. They're general chips that are used as specialized chips in the PS3.
I'd like to see a Unisys machine of these with octa-CPUs with DDR3 memory. Next I'd like to see games that were developed with heavy threading, so they can use different processors.
And you suddenly have a PS3 killer.
To see how anything can be good or bad for Linux, we'd have to see what is 'good for Linux' anyway.
Something that increases the uptake and usage of Linux is good for Linux. The biggest problem today Linux faces is lack of application market. Well OK the biggest problem is lack of standards, so lack of applications is the second biggest problem. Any slashdotters who are currently on win32 right now know which apps have forced them to stay with win32.... games, ERP system etc.
So why are these apps not ported to Linux??
(1) Linux isnt a big enough market yet
(2) Porting is a royal pain.
How can we fix that?
One way to solve this problem is to have diverse platforms. If the market is fragmented between platforms, vendors will be forced to sell different ports of apps. This means more vendors will move to toolkits like QT and wxwindows and opengl which are portable. If youre code is portable, porting to another platform is easy. If your code is based exclusively on VisualC and MFCs, its tough to port.
So anyone porting apps to win32 AND apple, might as well hire one more developer and maintain a Linux port as well. You wont find many apps that are just dual-platform. They're either just win32 (the kind of apps that are killing Linux), or they're released under at least 3 ports, win32, mac and Linux. In many cases the company says what the heck, and releases BeOS, FreeBSD, OS2, Solaris and other ports too, since porting further is easy.
Apple forces the move from the first type of apps to the latter type. That brings us all kinds of games and business apps on Linux. Probably the second best thing that can happen to Linux (the first thing is STANDARDS!!!)
Whats so artificial about it? It may not be a 'natural' diamond, but artificial? Like combining sodium and chlorine or burning hydrogen in the lab produces artificial salt or water.
You can manufacture diamonds and swear its real. The gold part, now that would be tough. Although many girls like silver colors and will settle for silver or other shinly substitute. If it is bling bling enough, they'll accept it as proof of your love. Just dont tell her the price.
I think DeBeers crushes all possibilities of manufacturing such diamonds legally in western countries. Thats why I'm interested in this business; I can make an upstart in my native Afghanistan. At least Central and South Asia could break free of the monopoly, no matter what any law says.
Not all female types are wives. Slashdotters have mothers you know
Maybe youre the only one. I hate the eraser touchpoints. Touchpads are bearable once you get used to it, they both have steep learning curves.
Of course I'd prefer a wireless bluetooth trackball over both
It certainly is happening. There are various seasonal cycles, glaciation cycles and others we dont know about because our history doesnt go back far enough. The neocene should be over by now, the next ice age is due. So we have no idea what to expect... its been 12000 years since the end of the pliestocene.
So all that means the Earth was never constant. There were forests in Saudia, a great desert where forests are currently, a lush ecosystem in the middle of the Antarctica. Surely humans didnt change that, we didnt exist. We only very recently became powerful enough to make big changes in the global system, but since its all constantly changing so much, our effected changes are lost. Think of the Tsunami. Could we do that? That was weather dear friend, it showed us how our effects are completely dwarfed by mother Natures'.
I dont think we can make lakes disappear yet. The greatest change we have been able to make is putting dams on rivers, and the heat of cities raising its temperatures by 2 celsius. Apart from that, things have always been changing, and will always be. My hometown used to get 10ft snow every winter 'back in the days', while now it didnt snow for a good 5 years straight. Legend has it that its always been getting warmer there, and such temperatures are to be expected.
Its certainly Global Warming, but not manmade.
The sky IS blue. But mac users certainly arent 16%. I'd be VERY surprised if they were more than 10%. You'll have great trouble finding a mac in most Asian countries; the mac is America-centric.
So here in N America, running into a mac is a rare event. The enormous bulk installations of colleges and companies are PC. Most of them are Dell, HP or IBM. A mac is more like for certain people with certain tastes, a few college libraries and graphic designing classes. Even the libraries I should say are more than 90% PC.
Mac users do spend $$$ on average more than PC users. Thats partly beause they HAVE to, and macs are more expensive anyway. Which means mac users will spend more on other things like software, monitor etc. Mac users are also more vocal; there arent many pro-PC people around.
It makes me wonder why shoot and kill in one location at all? Think of de_dust, which makes more sense... hunting down terrorists in a small village. Although de_dust has practically no houses or even shops or vehicles, and too many crates.
Look around you. There are apartments (great setting for a game, have never seen a complete apartment map, or a good one, there are villages, mountains, and certain places like airports, hospitals, downtown, offices, railway yard etc. The 747 map in CS was original too, but they overused airplanes in other maps.
I wouldnt mind seeing an underground parking lot map. Think of the parking lot scene in terminator2... I always thought that was a great setting for a game lots of glass to break and places to duck.
I was gonna say a school is good too, but I suppose its not.
And if youre gonna make a warehouse, add computer desks, trucks, weighing and wrapping machinery, forklifts and lifttrucks, piles of crates arranged properly to maximize space...
A library would be great if pieces of paper will fly if you shoot the books...
If the mantle is fluid enough and under pressure, wouldnt it gush out and form a volcano? We dont have volcanoes everywhere because the crust is thick enough to hold it down. Punch a hole in the waterbed and viola!
Volcanoes have a knack of closing themselves up, which is partly why this is being done in the ocean. Would be cool to have a build-a-volcano-mountain experiment on the surface somewhere, maybe close to Mt Fiji to augument the current one.
Why cant we just type 'yahoo' and 'google'? Many companies will register (eg mysql.org and mysql.com) multiple TLDs to cover their names online, but they wont have to...
.com. Go ahead and register org and net, people will forget and type .com first to get to your site.
Most people just use
They can sell top level domains except 2-character top-level domains which should remain as country TLDs and the likes...
"Microsofts plans to ship a Windows 2000 Update Rollup"
"The Update Rollup comes just one month before mainstream support for Windows 2000 client and server releases expires on June 30."
So have they released it or not? Those statements are contradictory. I can go check, but that means one of the statements will be proven false.
Lotus Domino
Business Objects
SQL 2000
win32-specific ERP software
commercially supported antispam system
Active Directory server and remote install server
Domino can run on Linux, BO too, SQL can be replaced with oracle, ERP run in WINEX, antispam replaced and Active Directory replaced with the Novell directory.
But thats way way too much money for little additional gain and the combination is not so well tested.
What OS do you use for the DEC? I havent been able to get Tru64 from anywhere, but efforts have been focused on trying to run OpenVMS. I did try the windows2000 beta with VisualC 6 and Office 2000 apps. Windows2000 is great with 164lx, but crashed a bit, not as stable as NT.
I did however find firefox for NT and ran that whole combination for a little while before returning to the OpenVMS attempts.
Generally, one big process requires the heaviest load, either a game engine, or gcc, or a video encoder, or an unthreaded database engine.
So given the current market, users will likely run a heavy single-threaded 32-bit app on this chip expecting it to be fast. In due time we'll have x64 everything, and if you run apps from good vendors, they'll be properly threaded in a balanced way and will take the maximum advantage of the chip you purchased.
That has more to do with GUI design than sheer power. When you click on an icon, it starts to piece together code from libraries and files everywhere, sometimes before the app window appears. That means disk bandwidth, and the concurrentness of disk reads. If it is read from a RAID 5 array, the program might start much faster than a single 5200RPM IDE disk with the head clicking back and forth while the CPU idles at 2%.
For this reason, OSX is quite snappy, and even WindowsXP is more snappy at starting most apps than the Linux desktops that I've seen. Linux is a general OS while XP is designed to feel snappy at other costs. I dont know where to place OSX
And I have alcohol-based stuff on my hands? I cant drive my car?
Or what if youre helping a drunk friend get home? As soon as you handle him/her, your hands are alcoholic and you cant drive.
Theres no deciding factor about a drunk person except the percentage of alcohol in the bloodstream, so the only reliable possible solution is to have an rfid device under the drivers skin which measures alcohol and relays the data to the car enabling it to start.
Now theres an idea.
Thats right. Portability ensures correctness of code. Thats why WindowsNT was released for 3 platforms, to ensure the code's bugs are ironed out faster (didnt work there much eh?)
Think of gcc and glibc. We get tonnes of issues from all sorts of ports. After they're fixed, the code is really clean. Maybe thats a bad example since they also handle lots of exceptions, but its more true for the likes of X, apache, mozilla, KDE and such.
The same can be said of source code. If it can be compiled by multiple compilers, its true C99 (or something), otherwise if it depends on compiler specific hacks, it wont compile on other compilers, or some compiler will dump errors that others wont.
If you were a graphics card manufacturer, would you just plug it into a development board, and if it works, sell it? Or would you have a slew of systems to test and benchmark the card in before releasing it to the market?
And what happens when another user on your server wants to start another daemon listening to such packets? How do you differentiate two different daemons listening to two different types of traffic, and keep them seperate enough to be safe from one another?
That identifier became 'tcp ports' in tcpip, which is why we dont use ip alone. If a user's daemon is connected to a port, noone else uses that port, and external processes can get to that daemon through that port securely.
The SYN header data will work if the data is small enough, if you screw the TCPIP stack enough and if it is guaranteed that only one daemon will ever need to listen on each given IP address.