Am I the only one who would buy one of those nifty new stamp printers _IF_ they didn't charge you extra? What a bunch of crap. I'm paying for the gear and saving them time/money, but they are charging me more... and they want to boost it even further. Bah!
Remember that cat vision article a while back? Looks like the technology to talk both ways is there... well kind of. This is still very promising.
My question is how useful would it be to someone blind from birth? How would you calibrate, etc. This guy was a least able to see till age 36, so had some idea of what to expect.
You need to understand that realtime doesn't have to mean "realfast". Realtime only means that the system needs to meet a deadline. (though this deadline is frequently a short amount of time from when some event happens...)
The reason this isn't really applicable for desktop stuff is that you have no true deadlines. I guess you could kludge some stuff like "the system shall redraw my widget within n seconds of a mouse click" or something, but generally there is no need.
On an mostly unrelated note, one thing that makes rt-linux cool is that you CAN run desktop apps on the same box as RT stuff. So you can be controlling a robot arm or something from the same box that your UI is running on. The "rest" of linux runs in the RT kernels idle time.
As much as I enjoy launching things at people, there are a LOT of other things you can do with a hard RT os.
What I always use as an example of a hard RT system is high-speed check sorter. You have a rigid deadline imposed by the mechanical hardware... that paper is coming down the slot NOW! (and let me tell you these things haul ass!)
If you miss one single deadline you usually (like 99% of the time) end up in a miss-sort situation or something and get to start all over.
Embedding a whole PC with linux on it is an ideal solution to lots of things on sorter. like imaging (Arghh!) and acting as a interface layer between the mainframe controlling the job and the sorter, etc. all the way dow to running the sorter itself.
I was more replying to the note that DNA couldn't/wouldn't occur in other life forms that live in a similar environment. You are of course way way correct on the evolution front...:-)
I wish I could recall the guy's name who ran this experiment, but I can't. This was quite a while back though. Asimov wrote up a good brief on it if I recall correctly...
Anyway, some people jammed a bunch of inorganic junk into a tube and let it cook for a while. (they were trying to simulate primordial "soup") Nothing significant came out of it so they added a device that made electric arcs. They were suddenly creating all sorts of interesting compounds and... here's the part i seem to remember but am not sure sbout... among them were some lower to medium level DNA building blocks.
Seems as if those types of compounds form so easily that DNA could be a bit more universal than you suggest.
food for thought anyway. If I weren't so darn lazy I would go look it up on the web I guess...
The systems people are SCARED about crashing are mostly embedded.
While not all embedded systems operate in the manner I'm about to describe, lots do.
Basically they use a clock that is just a counter (like the UNIX system clock) and reset at convient times. For example heavy machines. You have to PM the machine every so often. When you perform the maintenance procedure you reset the clock. When the machine detects a certain clock value, it shuts itself down to prevent damage due to lack of maintenance.
Note that the date doesn't matter. It's a trivial example, but very very common. More so than using a data cause it's easier (read as SMALLER) to code .
Sounds similar to how the kernel level MPI stuff for linux works...
The TCP/IP stack for clustered computing has long been a pet hate of mine... if this does what you say I think it's a great stride forward for distributd computing on Linux.
your def and my def of an STB must be different. When your saying set top box, I'm assumming MPEG I decode, etc., not just a IR decode setup...
For what I think your talking about now, hell yeah your right, but I don't think that kind of application even warrants a OS of any type. More like an hc11 or PIC with a coupla hundred lines of asm.
No I'm not that person. I always post as me. sorry.:-)
I'm also not confused, i do do this for a living. I really shouldn't respond to this cause i don't want to start a thread on the "definition" of real time, but I will anyway.
Real time is a term that is very very hard to make a fixed definition for. Generally a real time OS is deterministic. predictable. For things like set top boxes (a very vague term in itself) this is important. In a plam top,you are correct, you probably don't need a rtos (unless of course you are doing something cool like running a software modem in the back geound or something, then it could be nice).
a nice side effect of being predictable is usually being small and fast. Linux is fairly bloated for embedded apps. so is CE.
If you want put this back into the context of the original post, I was only refering to rtos stuff cause this is the class of product that CE claims to compete in. if you want to look at free NON rt micro kernels and executives, then CE is really toast in the free software arena and I stand by my original argument. if i ever had one.
and it's really quite simple. Ignoring Linux completely, there are a LOT of really really good RTOS and non real time embedded OSes out there. For free. And they are already mature.
Take OAR's RTEMS for example. It's small. It's quick. It has an IP stack. It runs on everything. It supports Ada. It supports multiple procs of different types. It's been around a long time.
CE has only 1 "feature" to make it stand out from the pack (IMHO) and that's the fact that the API tracks another non-RT os. Big deal. I don't think this has been enough to really launch it into it's target market yet.
I mean, good grief, look at the number of COMMERCIAL real time OSes out on the market right now that have been around for longer than CE. They picked a super competitive market, no wonder they are worried.
This is political survival. The winner will be a person who is a good manipulator and should be able to get elected to high office with no trouble whatsoever. Even get free exposure from CBS on the way.
YOU sir have no idea about what you are talking about. The original poster does actually know what he's talking aboute. You are arguing about the definition of the word clasified, when it's pretty obvious what the intent of the first poster was.
The security people happen to error on the side of EXTREME paranoia when clasifying data. What does this mean? It means that they will slap a "confidential" label on anything that could be considered even remotely interesting. And then it's usually not very usefull data anyway. If you want something interesting you usually end up dealing with S/SAR or TS material. (needless to say, this makes engineering life a royal PAIN in the arse.)
Let's talk networks now, shall we?
Networks that deal with different class levels have to be isolated. In between class levels you can SOMETIMES get away with firewall or crypto type isolation. sometimes. (like you can install a 1 way gate between a lower level and higher level to allow data be moved from a low level on up). To go from higher to lower level you have this god awfull procedure to follow. very "unfun".
From unclassified to ANY classification level ye olde NISPOM (gov security manual) says you must have physical media isolation. Period. An they really want you to go a step further and separate the wires by at least 1 meter. Fiber is prefered.
So, to conclude my rambling so I can go do something usefull, if someone stole truly "classified" data from a system that was accessiable via the Internet, somebody broke the rules on the DoD side. While that is possible, security folks tend to be really really paranoid and I doubt any data they stole was really classified.
Ummm, builder has a counter-intuitive interface? Just a sec here, i truly mean no offense here, but what DO you consider intuitive?
I would really like to know what you mean by "More often than not, the "reqirements" for CB4 to make and run my program get in the way of the actual design of what I want."
I used builder to make the win32 gui front end for a real time product. I had written the original app in VC++ and it was, quite frankly, a bitch to create and maintain. I downloaded the BCB demo from borland and recreated my entire product front end (a week of work in vc++) in a day. (granted I reused a LOT of code that wasn't gui related...)
I don't recall that builder ever got in my way. (this was a couple of years ago, and I've slept a few times since...:-) I was using all sorts of obscure crap too, serial drivers, custom (unisys poll/select woohoo!) serial protocol stacks, interface to a custom RT kernel, custom drivers for high speed DAQ boards, opengl, directx, lots of really messy crap.
As far as the GUI goes, it's not the greatest (I like tornado, but its in a different world i guess...) but I would really hesitate to call it counter intuitive.
I really would like to here from you on this one...
Ever try to get a passport? You are required BY the IRS to supply your SSN.
How about a PO box?
What about "selective" service?
SF-86?
FAFSA?
Sorry, you are just a number. Whether you like it or not doesn't matter. The SSN is no longer a private number. Just one of those things that happened over time. Started out private, then more obscure government uses started, and now every company or institution thinks you are required to supply it.
Your SSN is likely in every database on planet earth anyway.. ever look at those little labels on your junk mail? (what do they say, cartsort or something?) anyway they are covered with all sorts of screwy numbers along with your name and address. In quite a few of those number you will find you SSN encoded.
Anyway, I really do believe that it is already a universal federal government issued ID. I truly respect those who are trying to stop it from occurring, but I think your too late.
OK people, here's the deal. Banks don't roll their own ATMs. Several companies that make ATMs and the software that drives them use OS/2. Last I checked both Unisys and NCR did, I think their were some others too...
Fot those who have neve rdealt eith the financial market place, legacy hardware/software is more or less the rule. These machines will be running OS/2 for quite some time.
Besides ATMs many popular credit card imprint machines run OS/2 as well.
Hmm. Someone hit me if I'm way out in left field here, having not worked with IP stuff THIS big before, but I would think you could drop this into a Xilinx or Altera pretty easily.
Might not run quite as fast as silicon but it would still be plenty interesting for students, etc.
Maybe i'll have to dust off some hardware and try it out... using some of that spare time I don't have.:-)
I used to work for Digital, and I met some of the compiler team once. They are very smart people, and their compilers are very efficient and standards compliant to a point.
Sorry can't agree. The digital VAX C compiler is pretty bad. It has some bugs that are so blatant it should have never been released. (for exmaple it has an ugly habit of not ctacthing missing braces, and then generating bogus code. Ick.) I generally hesitate to use the word "garbage" in describing someone elses software, but that's what first comes to mind here. (Course I'm using an OLD version on OpenVMS 6.2, so that could be my problem I guess... )
Their ADA compiler rocks though, (very nice. a lot nicer than the ada we use for our cross compiler) so I guess I can't rip em to bad.:-)
Upgrade manufacturers would need a license to clone the ROM. (Can't just stick a new cPu into board and screw with firmware--whole new rom chip with valid Apple firmware required!)
Why? It's perfectly legal to recreate a boot ROM. I'm talking re-engineering here, not copying verbatim. You could even use the clasic "one engineering takes it a apart, another (clean) engineer puts it back together based on the report" method or something.
Slashdot Mirror
Am I the only one who would buy one of those nifty new stamp printers _IF_ they didn't charge you extra? What a bunch of crap. I'm paying for the gear and saving them time/money, but they are charging me more ... and they want to boost it even further. Bah!
dv
Remember that cat vision article a while back? Looks like the technology to talk both ways is there ... well kind of. This is still very promising.
My question is how useful would it be to someone blind from birth? How would you calibrate, etc. This guy was a least able to see till age 36, so had some idea of what to expect.
dv
Hmmmm ... just like WordPerfect, huh?
You need to understand that realtime doesn't have to mean "realfast". Realtime only means that the system needs to meet a deadline. (though this deadline is frequently a short amount of time from when some event happens ...)
The reason this isn't really applicable for desktop stuff is that you have no true deadlines. I guess you could kludge some stuff like "the system shall redraw my widget within n seconds of a mouse click" or something, but generally there is no need.
On an mostly unrelated note, one thing that makes rt-linux cool is that you CAN run desktop apps on the same box as RT stuff. So you can be controlling a robot arm or something from the same box that your UI is running on. The "rest" of linux runs in the RT kernels idle time.
dv
As much as I enjoy launching things at people, there are a LOT of other things you can do with a hard RT os.
... that paper is coming down the slot NOW! (and let me tell you these things haul ass!)
... :-)
What I always use as an example of a hard RT system is high-speed check sorter. You have a rigid deadline imposed by the mechanical hardware
If you miss one single deadline you usually (like 99% of the time) end up in a miss-sort situation or something and get to start all over.
Embedding a whole PC with linux on it is an ideal solution to lots of things on sorter. like imaging (Arghh!) and acting as a interface layer between the mainframe controlling the job and the sorter, etc. all the way dow to running the sorter itself.
back to launching things
dv
Last I checked you could get this on a Hummer! Far cooler, and safer to drive, and only slightly *grin* more expensive.
dv
Congrats.
I was more replying to the note that DNA couldn't/wouldn't occur in other life forms that live in a similar environment. You are of course way way correct on the evolution front ... :-)
I wish I could recall the guy's name who ran this experiment, but I can't. This was quite a while back though. Asimov wrote up a good brief on it if I recall correctly ...
... here's the part i seem to remember but am not sure sbout ... among them were some lower to medium level DNA building blocks.
...
Anyway, some people jammed a bunch of inorganic junk into a tube and let it cook for a while. (they were trying to simulate primordial "soup") Nothing significant came out of it so they added a device that made electric arcs. They were suddenly creating all sorts of interesting compounds and
Seems as if those types of compounds form so easily that DNA could be a bit more universal than you suggest.
food for thought anyway. If I weren't so darn lazy I would go look it up on the web I guess
dv
The systems people are SCARED about crashing are mostly embedded.
While not all embedded systems operate in the manner I'm about to describe, lots do.
Basically they use a clock that is just a counter (like the UNIX system clock) and reset at convient times. For example heavy machines. You have to PM the machine every so often. When you perform the maintenance procedure you reset the clock. When the machine detects a certain clock value, it shuts itself down to prevent damage due to lack of maintenance.
Note that the date doesn't matter. It's a trivial example, but very very common. More so than using a data cause it's easier (read as SMALLER) to code .
dv
Sounds similar to how the kernel level MPI stuff for linux works ...
... if this does what you say I think it's a great stride forward for distributd computing on Linux.
The TCP/IP stack for clustered computing has long been a pet hate of mine
Does anybody know what they are changing?
(another flavor of kernel support for cluster wide resoures/pids/etc. or something?)
The article basically said nothing about what their system does to make it better/different than any existing clustering setup.
dv
your def and my def of an STB must be different. When your saying set top box, I'm assumming MPEG I decode, etc., not just a IR decode setup ...
For what I think your talking about now, hell yeah your right, but I don't think that kind of application even warrants a OS of any type. More like an hc11 or PIC with a coupla hundred lines of asm.
The difference is that a real time OS will allow you to know what the hard bound of the response will be. The worst case is PREDICTABLE.
Normal Linux scheduling will not do this.
This is really the essence of what real time is.
No I'm not that person. I always post as me. sorry. :-)
I'm also not confused, i do do this for a living. I really shouldn't respond to this cause i don't want to start a thread on the "definition" of real time, but I will anyway.
Real time is a term that is very very hard to make a fixed definition for. Generally a real time OS is deterministic. predictable. For things like set top boxes (a very vague term in itself) this is important. In a plam top,you are correct, you probably don't need a rtos (unless of course you are doing something cool like running a software modem in the back geound or something, then it could be nice).
a nice side effect of being predictable is usually being small and fast. Linux is fairly bloated for embedded apps. so is CE.
If you want put this back into the context of the original post, I was only refering to rtos stuff cause this is the class of product that CE claims to compete in. if you want to look at free NON rt micro kernels and executives, then CE is really toast in the free software arena and I stand by my original argument. if i ever had one.
later,
/dev
and it's really quite simple. Ignoring Linux completely, there are a LOT of really really good RTOS and non real time embedded OSes out there. For free. And they are already mature.
... back to work. argh.
Take OAR's RTEMS for example. It's small. It's quick. It has an IP stack. It runs on everything. It supports Ada. It supports multiple procs of different types. It's been around a long time.
CE has only 1 "feature" to make it stand out from the pack (IMHO) and that's the fact that the API tracks another non-RT os. Big deal. I don't think this has been enough to really launch it into it's target market yet.
I mean, good grief, look at the number of COMMERCIAL real time OSes out on the market right now that have been around for longer than CE. They picked a super competitive market, no wonder they are worried.
speaking of embedded
dv
Yeah, I thought "Well when the hell did they BUY them whem I read it ..." too.
... :)
Thought I may have missed an investment op
glad I wasn't the only one.
This is political survival. The winner will be a person who is a good manipulator and should be able to get elected to high office with no trouble whatsoever. Even get free exposure from CBS on the way.
dv
YOU sir have no idea about what you are talking about. The original poster does actually know what he's talking aboute. You are arguing about the definition of the word clasified, when it's pretty obvious what the intent of the first poster was.
The security people happen to error on the side of EXTREME paranoia when clasifying data. What does this mean? It means that they will slap a "confidential" label on anything that could be considered even remotely interesting. And then it's usually not very usefull data anyway. If you want something interesting you usually end up dealing with S/SAR or TS material. (needless to say, this makes engineering life a royal PAIN in the arse.)
Let's talk networks now, shall we?
Networks that deal with different class levels have to be isolated. In between class levels you can SOMETIMES get away with firewall or crypto type isolation. sometimes. (like you can install a 1 way gate between a lower level and higher level to allow data be moved from a low level on up). To go from higher to lower level you have this god awfull procedure to follow. very "unfun".
From unclassified to ANY classification level ye olde NISPOM (gov security manual) says you must have physical media isolation. Period. An they really want you to go a step further and separate the wires by at least 1 meter. Fiber is prefered.
So, to conclude my rambling so I can go do something usefull, if someone stole truly "classified" data from a system that was accessiable via the Internet, somebody broke the rules on the DoD side. While that is possible, security folks tend to be really really paranoid and I doubt any data they stole was really classified.
later,
dv
Ummm, builder has a counter-intuitive interface? Just a sec here, i truly mean no offense here, but what DO you consider intuitive?
...)
... :-) I was using all sorts of obscure crap too, serial drivers, custom (unisys poll/select woohoo!) serial protocol stacks, interface to a custom RT kernel, custom drivers for high speed DAQ boards, opengl, directx, lots of really messy crap.
...) but I would really hesitate to call it counter intuitive.
...
I would really like to know what you mean by "More often than not, the "reqirements" for CB4 to make and run my program get in the way of the actual design of what I want."
I used builder to make the win32 gui front end for a real time product. I had written the original app in VC++ and it was, quite frankly, a bitch to create and maintain. I downloaded the BCB demo from borland and recreated my entire product front end (a week of work in vc++) in a day. (granted I reused a LOT of code that wasn't gui related
I don't recall that builder ever got in my way. (this was a couple of years ago, and I've slept a few times since
As far as the GUI goes, it's not the greatest (I like tornado, but its in a different world i guess
I really would like to here from you on this one
/dev
Ever try to get a passport? You are required BY the IRS to supply your SSN.
.. ever look at those little labels on your junk mail? (what do they say, cartsort or something?) anyway they are covered with all sorts of screwy numbers along with your name and address. In quite a few of those number you will find you SSN encoded.
How about a PO box?
What about "selective" service?
SF-86?
FAFSA?
Sorry, you are just a number. Whether you like it or not doesn't matter. The SSN is no longer a private number. Just one of those things that happened over time. Started out private, then more obscure government uses started, and now every company or institution thinks you are required to supply it.
Your SSN is likely in every database on planet earth anyway
Anyway, I really do believe that it is already a universal federal government issued ID. I truly respect those who are trying to stop it from occurring, but I think your too late.
/dev
OK people, here's the deal. Banks don't roll their own ATMs. Several companies that make ATMs and the software that drives them use OS/2. Last I checked both Unisys and NCR did, I think their were some others too ...
Fot those who have neve rdealt eith the financial market place, legacy hardware/software is more or less the rule. These machines will be running OS/2 for quite some time.
Besides ATMs many popular credit card imprint machines run OS/2 as well.
/dev
Hmm. Someone hit me if I'm way out in left field here, having not worked with IP stuff THIS big before, but I would think you could drop this into a Xilinx or Altera pretty easily.
... using some of that spare time I don't have. :-)
Might not run quite as fast as silicon but it would still be plenty interesting for students, etc.
Maybe i'll have to dust off some hardware and try it out
/dev
I used to work for Digital, and I met some of the compiler team once. They are very smart people, and their compilers are very efficient and standards compliant to a point.
... )
:-)
Sorry can't agree. The digital VAX C compiler is pretty bad. It has some bugs that are so blatant it should have never been released. (for exmaple it has an ugly habit of not ctacthing missing braces, and then generating bogus code. Ick.) I generally hesitate to use the word "garbage" in describing someone elses software, but that's what first comes to mind here. (Course I'm using an OLD version on OpenVMS 6.2, so that could be my problem I guess
Their ADA compiler rocks though, (very nice. a lot nicer than the ada we use for our cross compiler) so I guess I can't rip em to bad.
enough rambling. back to the salt mines.
/dev
Why? It's perfectly legal to recreate a boot ROM. I'm talking re-engineering here, not copying verbatim. You could even use the clasic "one engineering takes it a apart, another (clean) engineer puts it back together based on the report" method or something.
/dev