I think this is a good thing, but until there is sufficient infrastructure in place to verify electronic signatures, we're all in for a shock.
What we need is a public-key registrar that can be used to verify the authenticity of the signature. At the moment the only folks that are key-registrars are for fee, and I'm not going to pay an ongoing fee so someone else can vouch for my signature. Unfortunately this means we'll need the fed to step in a be a public, public-key registrar.
I realy HATE having to invoke ANOTHER federal program, but here I think it makes sense.
Does anyone know if the 2.2 or 2.4 kernels are 64bit clean? It would be ashame to run a 450MHz Ultra Sparc II processor in 32bit mode. Anyone know if the lack of a 64bit Linux version had any bearing on this or Sun Soft's actions of late surounding Linux?
No, he's serious. I did a quick look at his previous postings, and this guy seems to get off talking down to folks about all there various oppinions - the mark of a true M$ zealot.
Perhaps we can get the powers that be to have everyone with a negative karma to automatically post at -1, so the rest of us can get on with our lives...just a thought
There is a diference between a user-level service, like SmartDrive, wherein they hook the INT13 calls to redirect them to an in-memory cache - and Unix. Unix uses the memory as part of every read-write operation within the kernel. That saves context switches and page faults.
Okay, Context switch: when one process foregoes using the CPU so that another process can have time. This involves at a minimum a switch in the stack frame pointer and a fixup to the program-counter and instruction pointer
Page fault: When a page of memory, typically 4096 bytes but can be as small as 512 bytes or as large as the hardware allows, is referenced (read or write) by a running program. When the memory management unit converts the logical address from the program address register to a physical memory address, and faults - meaing there is no active translation, the running process is halted, put into the WAIT queue, and the PAGING process is sent a message (or called directly) to bring back the page from disk (SWAP). This is Demand-Paging. If there are no unused pages in memory, then when the paging process runs to fulfill the request(s) it will determine which pages can be safely removed from physical memory and written to disk, to make room for the pages that need to be brought back.
In a memory starved system a great deal of time is spent by the paging process to move chunks of memory, constantly between disk and physical memory. This is called "thrashing".
Getting back to the original statement, the reason unix is fundamentally different from all the examples of disk caching, is that it is part of the kernel routine for ALL DISK I/O, not just certain files that the caching program likes. Smartdrive has to determine whether the current read is for a regular file, for a paging request or for a system file and acts differently depending on the decision. This all takes TIME. Since in Unix the buffering is part of the kernel, and not a separate process it saves two (2) context-switches on every operation. Since Unix has no distinctions in file types, ALL files are buffered. Since in Unix the buffering is handled at a higher level than the hardware (BIOS) interrupt service routines, which is what SMartdrive has to do, there is no ambiguity between a paging request - no buffering, and a file I/O - buffered. If you ever write a program, in C, on a Unix platform, one of the things that becomes apparent is the use of libary calls - section 3 in the man pages, and system calls - section 2. The library calls are for buffered file i/o. The system calls are for unbuffered i/o. So if you want syncronous writes you can have them. In Windows as in other OS's there is no distinction. You have write() and read(). Even if you have an s_write() (syncronous) call available, if there is a caching program running, since it has hooked the hardware interrupt service routine, it is wholly up to that program to allow your syncronous write through, or not. The upper layer file I/O routines can't tell if the hardware has been hooked.
As far as read/write performance, that is the essence of the differences in the various file systems. UFS assumed a sequential read and so performs a limited read-ahead, typical just the next block. Other file systems use other read-ahead algorithms - of which I am not specifically privy. Perhaps some of you Linux kernel hacks could expound on e2fs and its read-ahead algorithm.
In any case the disk meta-data, the inode table and the superblock, are held in memory at all times and are marked as locked, in systems that allow that level of control in their mmu. That is what the FSCK process rebuilds when you fail to do a clean shutdown - it rebuilds the metadata from the files.
In conclusion, Unix is better because it buffers on every file I/O, because it handles the buffering higher in the call chain, because it buffers without doing an extra context switch, because it performs paging at a lower level in the call chain than the buffering, and because it has the tools in place to do true syncronous writes - if you need them.
One of the Fundamental differences between all Unixes and every other OS ever invented is the use of memory to buffer the filesystem. There are no direct writes to a disk, ever! All file IO is done to the buffers in memory, and then eventually the bdflush daemon runs and syncs the disk to the image in memory. Notwithstanding the recent journaling file-systems, and the sync-write IO calls, Unix today still does all its file-IO in this manner...which is why a RAMdisk is redundant. You already HAVE a RAMdisk.
Mac, Windows, VMS, MVS, Amiga, et.al all do direct and/or syncronous writes to the disks. That's why a RAMdisk has such an effect.
Linux boot floopies use a RAMdisk because they can't put all the needed files onto a 1.44MB floppy without compressing the image. The RAMdisk is simply the "disk" to which the decompression runs its output. If the root filesystem could fit entirely onto a floopy, there'd be no need for a RAMdisk upon install. See Redhat verion 3.x
Quick2C: Name any hacked site that uses MVS...
You do know what MVS is don't you...
Okay how about Guardian...
Comodore Basic is certainly cute, but it's not relevant. I want to know what I can run my Business on, not my 13-year-old son's web-page.
As far as security is concerned, try this on for size;
1. Build one of the micro-Linux distro's into a bootable image with all your fav services, http, ftp, sendmail - the works.
2. Burn a CD with the ROM image of the distro including the static directories like/bin - all you guru's do know how to make gcc build to a foreign file system, right?
3. Bring up your fav AMD-K6 (I HATE Intel) based system, with 128M of memory, and/var and/tmp as RAM disks.
Lets see some idiot script-kiddie crack THAT box! I think/bin/ls would be a little tough to overwrite given its on Write-once media...
Oh, and if it does somehow get compromised, I just power-cycle the whole thing....
Let me start by saying a work for a company that is currently putting 6-billion, note the 'b", into improving the last mile-experience of customers. The idea is simple, get glass closer to the customer.
The idea of glass to the curb, or glass to the home is great, but it ignores the simple fact of all the copper that is currently in place, not to mention a few minor details - like how do I provide power over glass?
In case you weren't aware, the phone company is required by law to provide power to phones within a building so that in the event of a power failure the occupants can still call 911 for help. To put glass up to the house, or even the curb, requires that a device be in place at that point to povide 48v DC to the telephone equipment.
So, what is my company doing? We're putting glass into the remote terminals, which are the large cabinet-looking devices you see along the roads that service a few blocks, or a buisness park. These are powered already and have elecrtronics in them currently. Thus the only thing realy needed is a DSLAM in the cabinet, and glass pulled from there back to the CO. Very clean, very effective. Once that happens, nearly anyone in a Metro-service area is within 3000 cable-feet of the DSLAM, making even VDSL possible.
As far as some of the other issues flying around here, like who deserves fast access, and how will it be controlled - Is this the same community that wants to abolish all controls by anyone? Is this the same community that goes on forever about the evil FBI, and corporate security outfits that snoop lines and emails? Why is it that you folks want no fetters on YOUR activity but want all the non-techy Joe/Jane Users out there to be hammered? Out we being just a little bit hipocritical here?
Having had to administer MacOS 6,7 and 8, as well as Novel 3,4; NT 3.5/4; Windows 3/9x; I can honestly say that for the most part there are two (2) kinds of people that use a Mac:
People that have a specific piece of software that only runs on a Mac;
People that are too stupid to use any other kind of computer.
Please don't misunderstand me. I'm NOT saying that everyone that use a Mac is less than intelligent. Rather the Mac has so much built into the UI that even a flaming idiot can use the computer sucessfully.
If your not in one of those two groups, then you need a different OS. I have several friends that are very much Mac bigots (one's an edu-crat) and even they concede that the Mac is not the Best answer to all questions.
NEITHER IS LINUX!!!
It is very good at tasks that involve lots of:
File I/O
Interprocess communication
IP Networking
It stinks as a compute server, or a GUI based workstation - apologies to the Gnome community but as long as the GUI runs as a user process it will have to compete with all the other user processes and thus it will not ever be responsive to all my clicking. That's why the Windows and Mac OS programmers buried the UI in the kernel - and why the kernel is therefore much less stable.
Everything is a trade off.
I say, use the best tool for the job. Got a pointy-haired boss that wants to read his own email - get him a Mac. Got a secretary that types 120 words/minute - get her DOS Word Perfect. Got an airline reservation system that takes 120,000 transactions per minute - get an S390. No one tool is always correct.
I'm suprised SCO lasted as long as it did. I STILL run into doctor's offices and other such places that run there whole business on a single app, with Wyse60 terminals (9600 baud of course) into a small PC running one of the above.
I actually had the pleasure? of working with OpenServer before it was called that, back in the days when SVR3 was new, and TCP/IP was an OPTION! What I discovered was that Openserver and Unixware STILL (i.e. today) have TCP/IP as an add on. That along with the ridiculously expensive licensing fees associated with the SCO product lines are what killed the company.
Its kind of like the old token-ring versus ethernet discussion: Sure token is "better", deterministic, rant-rant-rant, but when it comes down to signing on the dotted line everybody wanted to buy $150 ethernet cards (1992) not $600 token cards. The same is true for the SCO product line. The only reason ANYONE is running SCO or Unixware at this point is because the VAR developers still get a cut of the licensing fees from SCO, so that's the platform they sell to their customers; even though the same App could run unmodified on BSD or Linux.
Personally I'm glad to see SCO gasping its last breath. I'm just sorry that Caldera picked up those Unices. It seems that anyone that handles Unixware in particular gets hammered.
Re:If all that stuff is OS..
on
Is UNIX An OS?
·
· Score: 1
I think the folks from Adobe would take exception to our definition. So would a large number of other folks that sell any kind of Application.
There are two main reasons to get education beyond what you are curious about. 1. Self-directed learning will leave you with significant gaps in your knowledge. Since you only pursue those areas that interest you, you fail to grasp any of the "mundane" concepts or parts of the system outside your personal tastes. You may know procmail because that's what you set up, but you don't know sendmail, or appreciate the vaguarities of writing a general purpose parser. 2. You only learn by being stretched and it takes someone that knows more than you to stretch you. If you think there is no one that knows more than you....
The original post dealt with one aspect of this greater issue. That is how do we authenticate what we know. Demonstration is not an option, unless you plan on working for free, forever.
Certification is ONE answer, not the only answer or even the best answer. I know of many bad lawers, doctors, and CPA's that managed to pass their respective certification processes. Nevertheless our industry will continue to suffer both internally and in relation to other industries as long as there are unqualified, unaware and self-taught individuals making major decisions. To put this into as much perspective as possible, would you visit a self-taught doctor, or see a self-taught lawyer for the subsequent malpratice suit? Would you want a self-taught CPA to go with you to the IRS audit? If you answered NO to any of these, than why are we still trying to validate the efforts of those unwilling to submit to outside validation?
If you taught yourself Linux, Redhat, then can I call on you to disentagle a terminal server on the end of a frame-relay link? How about a Novell server that has a recurring NLM failure? What about a problem with a Visual Basic program used to run my business? I could go on, but the point is you don't know everything. You don't even know all you could know, and because of that you are a liability to yourself, your employer/customer and to our industry.
Wake up guys. It's not a hobby, its a career. Treat it as such.
The reason the IT industry has certifications is primarily a reaction to a problem that I don't see in any other industry. We have a demand that far outpaces the supply and because of that we have a LOT of people who are in the industry in various positions that don't know the elbow from the...w ell from there foot. What's worse is we have even more people who no enough to be mildly succesful wihtin the narrow field of experience but are lost when they leave their comfort zone. I'd very much like to see the same kind of review and certification process that CPA's, lawyers, doctors and other professions enjoy (endure?). I'm tired of dealing with people who THINK they know technology when all they realy know is the three things they've worked on since they left school.
Second rant: (if you're already slighted leave now) I agree with the two other writers here about the tauted "experience" versus "book knowledge" argument. If you taught yourself everything you know then you had a fool for a student. The way to inovation is to build on the efforts, knowledge and experience of others. The experience benefit only occurs when you have the necessary foundation of book knowledge. Otherwise you MAY figure out HOW to accomplish something, but you'll never understand WHY it works or WHAT the ramifications are of your actions. It is all of you self-taught folks I'd like to weed out of this field. You are the main thing that prevents the rest of us from moving forward. Feel free to email me personally with your flames. I've heard them all before.
I looked at the wiring diagram posted above and noticed a few things: 1. There is no feed-back capability, thus the machine in question will not receive its own packets. This is a violation of all the IEEE rules and will likely reek havoc on all but the most stable IP stacks. 2. The collision detection pin is diliberately grounded, thus this is a FULL DUPLEX connection.
Both points can be handled IF AND ONLY IF you can configure the device driver to accept a full-duplex configuration, which is unlikely for a driver meant to work across an AUI. You may find that it works, sort of, under light load.
I agree with the other posts, the transceiver is the best alternative. The microtransceivers are running $35 or so, retail.
I think the bridge metaphor breaks down pretty quickly...a bridge is a physical thing that can be objectively measured. A bridge goes through several "inspections" by disinterested or hostile others. A bridge is expected to last 50 to 100 years.
Anyone that ever looked at some of the theoretical implications of say proving the "correctness" of a program knows that the objective inspection is basically impossible. The others are obviously outside the realm of progrmaming. So what is the idea behind the certification/license issue?
Are we trying to prevent studip people from writing code? Then we would need to prevent all the various software warehouses from selling a copy of Visual Stupid (you pick) to any and all suits and programmer-wanna-be's. That would be a good thing for several reasons.:)
If the idea is to give some malcontent reason for a lawsuit, well we already have enough of them to keep the courts busy for the next several millenia.
I like the idea of the guild though...self directed, self fulfilling, able to bring along the journeyman to the level of master with a known process and a mentoring methodology...Yeah I like that idea
Clearly you are a part of the HR organization. Cooperation is demonstrated by being available during weekends, wearing a pager continuously, working >40 hours week (w/o OT compensation) and being willing to hop on a plane at a moments notice. These are all clearly described in a simple resume and easily verifiable (Is the candidate wearing the ubiquitous pager during the interview?) Divulging other personal information, such as marital status, is clearly prohibited by law. Why do the HR and placement (read head-hunter) forces feel they are entitled to this particular piece of VERY personal information. I think countering with a request for the current salaries of other members of the team and/or the budget would put this kind of talk in its proper place. I've been down this road a few times myself in recent (read last 3 years) history. The times I divulged my current salary I typically got offers that were 5-7% higher than that figure. When I did not, there was no correlation whatsoever. The job I currently hold I put a stake in the ground at a particular figure and refused to mention any other figure in any other context. I got what I asked for, and isn't success more compelling an argument than making the interviewer "feel better" about you.
Slashdot Mirror
All well and good, but you can't put the astronauts, themselves, into an autoclave! The spores are caried on ->US- not in boxes in the cargo hold.
I think this is a good thing, but until there is sufficient infrastructure in place to verify electronic signatures, we're all in for a shock.
What we need is a public-key registrar that can be used to verify the authenticity of the signature. At the moment the only folks that are key-registrars are for fee, and I'm not going to pay an ongoing fee so someone else can vouch for my signature. Unfortunately this means we'll need the fed to step in a be a public, public-key registrar.
I realy HATE having to invoke ANOTHER federal program, but here I think it makes sense.
Does anyone know if the 2.2 or 2.4 kernels are 64bit clean? It would be ashame to run a 450MHz Ultra Sparc II processor in 32bit mode. Anyone know if the lack of a 64bit Linux version had any bearing on this or Sun Soft's actions of late surounding Linux?
:-)
Enquiring minds want to know
No, he's serious. I did a quick look at his previous postings, and this guy seems to get off talking down to folks about all there various oppinions - the mark of a true M$ zealot.
Perhaps we can get the powers that be to have everyone with a negative karma to automatically post at -1, so the rest of us can get on with our lives...just a thought
Okay, lesson in Operating Systems...
There is a diference between a user-level service, like SmartDrive, wherein they hook the INT13 calls to redirect them to an in-memory cache - and Unix. Unix uses the memory as part of every read-write operation within the kernel. That saves context switches and page faults.
Okay, Context switch: when one process foregoes using the CPU so that another process can have time. This involves at a minimum a switch in the stack frame pointer and a fixup to the program-counter and instruction pointer
Page fault: When a page of memory, typically 4096 bytes but can be as small as 512 bytes or as large as the hardware allows, is referenced (read or write) by a running program. When the memory management unit converts the logical address from the program address register to a physical memory address, and faults - meaing there is no active translation, the running process is halted, put into the WAIT queue, and the PAGING process is sent a message (or called directly) to bring back the page from disk (SWAP). This is Demand-Paging. If there are no unused pages in memory, then when the paging process runs to fulfill the request(s) it will determine which pages can be safely removed from physical memory and written to disk, to make room for the pages that need to be brought back.
In a memory starved system a great deal of time is spent by the paging process to move chunks of memory, constantly between disk and physical memory. This is called "thrashing".
Getting back to the original statement, the reason unix is fundamentally different from all the examples of disk caching, is that it is part of the kernel routine for ALL DISK I/O, not just certain files that the caching program likes. Smartdrive has to determine whether the current read is for a regular file, for a paging request or for a system file and acts differently depending on the decision. This all takes TIME. Since in Unix the buffering is part of the kernel, and not a separate process it saves two (2) context-switches on every operation. Since Unix has no distinctions in file types, ALL files are buffered. Since in Unix the buffering is handled at a higher level than the hardware (BIOS) interrupt service routines, which is what SMartdrive has to do, there is no ambiguity between a paging request - no buffering, and a file I/O - buffered. If you ever write a program, in C, on a Unix platform, one of the things that becomes apparent is the use of libary calls - section 3 in the man pages, and system calls - section 2. The library calls are for buffered file i/o. The system calls are for unbuffered i/o. So if you want syncronous writes you can have them. In Windows as in other OS's there is no distinction. You have write() and read(). Even if you have an s_write() (syncronous) call available, if there is a caching program running, since it has hooked the hardware interrupt service routine, it is wholly up to that program to allow your syncronous write through, or not. The upper layer file I/O routines can't tell if the hardware has been hooked.
As far as read/write performance, that is the essence of the differences in the various file systems. UFS assumed a sequential read and so performs a limited read-ahead, typical just the next block. Other file systems use other read-ahead algorithms - of which I am not specifically privy. Perhaps some of you Linux kernel hacks could expound on e2fs and its read-ahead algorithm.
In any case the disk meta-data, the inode table and the superblock, are held in memory at all times and are marked as locked, in systems that allow that level of control in their mmu. That is what the FSCK process rebuilds when you fail to do a clean shutdown - it rebuilds the metadata from the files.
In conclusion, Unix is better because it buffers on every file I/O, because it handles the buffering higher in the call chain, because it buffers without doing an extra context switch, because it performs paging at a lower level in the call chain than the buffering, and because it has the tools in place to do true syncronous writes - if you need them.
One of the Fundamental differences between all Unixes and every other OS ever invented is the use of memory to buffer the filesystem. There are no direct writes to a disk, ever! All file IO is done to the buffers in memory, and then eventually the bdflush daemon runs and syncs the disk to the image in memory. Notwithstanding the recent journaling file-systems, and the sync-write IO calls, Unix today still does all its file-IO in this manner...which is why a RAMdisk is redundant. You already HAVE a RAMdisk.
Mac, Windows, VMS, MVS, Amiga, et.al all do direct and/or syncronous writes to the disks. That's why a RAMdisk has such an effect.
Linux boot floopies use a RAMdisk because they can't put all the needed files onto a 1.44MB floppy without compressing the image. The RAMdisk is simply the "disk" to which the decompression runs its output. If the root filesystem could fit entirely onto a floopy, there'd be no need for a RAMdisk upon install. See Redhat verion 3.x
Quick2C: Name any hacked site that uses MVS...
/bin - all you guru's do know how to make gcc build to a foreign file system, right?
/var and /tmp as RAM disks.
/bin/ls would be a little tough to overwrite given its on Write-once media...
:-)
You do know what MVS is don't you...
Okay how about Guardian...
Comodore Basic is certainly cute, but it's not relevant. I want to know what I can run my Business on, not my 13-year-old son's web-page.
As far as security is concerned, try this on for size;
1. Build one of the micro-Linux distro's into a bootable image with all your fav services, http, ftp, sendmail - the works.
2. Burn a CD with the ROM image of the distro including the static directories like
3. Bring up your fav AMD-K6 (I HATE Intel) based system, with 128M of memory, and
Lets see some idiot script-kiddie crack THAT box! I think
Oh, and if it does somehow get compromised, I just power-cycle the whole thing....
Just a thought
Let me start by saying a work for a company that is currently putting 6-billion, note the 'b", into improving the last mile-experience of customers. The idea is simple, get glass closer to the customer.
The idea of glass to the curb, or glass to the home is great, but it ignores the simple fact of all the copper that is currently in place, not to mention a few minor details - like how do I provide power over glass?
In case you weren't aware, the phone company is required by law to provide power to phones within a building so that in the event of a power failure the occupants can still call 911 for help. To put glass up to the house, or even the curb, requires that a device be in place at that point to povide 48v DC to the telephone equipment.
So, what is my company doing? We're putting glass into the remote terminals, which are the large cabinet-looking devices you see along the roads that service a few blocks, or a buisness park. These are powered already and have elecrtronics in them currently. Thus the only thing realy needed is a DSLAM in the cabinet, and glass pulled from there back to the CO. Very clean, very effective. Once that happens, nearly anyone in a Metro-service area is within 3000 cable-feet of the DSLAM, making even VDSL possible.
As far as some of the other issues flying around here, like who deserves fast access, and how will it be controlled - Is this the same community that wants to abolish all controls by anyone? Is this the same community that goes on forever about the evil FBI, and corporate security outfits that snoop lines and emails? Why is it that you folks want no fetters on YOUR activity but want all the non-techy Joe/Jane Users out there to be hammered? Out we being just a little bit hipocritical here?
People that have a specific piece of software that only runs on a Mac;
People that are too stupid to use any other kind of computer.
Please don't misunderstand me. I'm NOT saying that everyone that use a Mac is less than intelligent. Rather the Mac has so much built into the UI that even a flaming idiot can use the computer sucessfully.
If your not in one of those two groups, then you need a different OS. I have several friends that are very much Mac bigots (one's an edu-crat) and even they concede that the Mac is not the Best answer to all questions.
NEITHER IS LINUX!!!
It is very good at tasks that involve lots of:
File I/O
Interprocess communication
IP Networking
It stinks as a compute server, or a GUI based workstation - apologies to the Gnome community but as long as the GUI runs as a user process it will have to compete with all the other user processes and thus it will not ever be responsive to all my clicking. That's why the Windows and Mac OS programmers buried the UI in the kernel - and why the kernel is therefore much less stable.
Everything is a trade off.
I say, use the best tool for the job. Got a pointy-haired boss that wants to read his own email - get him a Mac. Got a secretary that types 120 words/minute - get her DOS Word Perfect. Got an airline reservation system that takes 120,000 transactions per minute - get an S390.
No one tool is always correct.
I'm suprised SCO lasted as long as it did. I STILL run into doctor's offices and other such places that run there whole business on a single app, with Wyse60 terminals (9600 baud of course) into a small PC running one of the above.
I actually had the pleasure? of working with OpenServer before it was called that, back in the days when SVR3 was new, and TCP/IP was an OPTION! What I discovered was that Openserver and Unixware STILL (i.e. today) have TCP/IP as an add on. That along with the ridiculously expensive licensing fees associated with the SCO product lines are what killed the company.
Its kind of like the old token-ring versus ethernet discussion: Sure token is "better", deterministic, rant-rant-rant, but when it comes down to signing on the dotted line everybody wanted to buy $150 ethernet cards (1992) not $600 token cards. The same is true for the SCO product line. The only reason ANYONE is running SCO or Unixware at this point is because the VAR developers still get a cut of the licensing fees from SCO, so that's the platform they sell to their customers; even though the same App could run unmodified on BSD or Linux.
Personally I'm glad to see SCO gasping its last breath. I'm just sorry that Caldera picked up those Unices. It seems that anyone that handles Unixware in particular gets hammered.
I think the folks from Adobe would take exception to our definition. So would a large number of other folks that sell any kind of Application.
This is just a little TOO close to what's been happening over at Columbia Internet.
0 4&mode=
http://ars.userfriendly.org/cartoons/?id=200008
I'll bet that there is some way that the LAIM allows VBA macro processing.
There are two main reasons to get education beyond what you are curious about.
1. Self-directed learning will leave you with significant gaps in your knowledge. Since you only pursue those areas that interest you, you fail to grasp any of the "mundane" concepts or parts of the system outside your personal tastes. You may know procmail because that's what you set up, but you don't know sendmail, or appreciate the vaguarities of writing a general purpose parser.
2. You only learn by being stretched and it takes someone that knows more than you to stretch you. If you think there is no one that knows more than you....
The original post dealt with one aspect of this greater issue. That is how do we authenticate what we know. Demonstration is not an option, unless you plan on working for free, forever.
Certification is ONE answer, not the only answer or even the best answer. I know of many bad lawers, doctors, and CPA's that managed to pass their respective certification processes. Nevertheless our industry will continue to suffer both internally and in relation to other industries as long as there are unqualified, unaware and self-taught individuals making major decisions. To put this into as much perspective as possible, would you visit a self-taught doctor, or see a self-taught lawyer for the subsequent malpratice suit? Would you want a self-taught CPA to go with you to the IRS audit? If you answered NO to any of these, than why are we still trying to validate the efforts of those unwilling to submit to outside validation?
If you taught yourself Linux, Redhat, then can I call on you to disentagle a terminal server on the end of a frame-relay link? How about a Novell server that has a recurring NLM failure? What about a problem with a Visual Basic program used to run my business? I could go on, but the point is you don't know everything. You don't even know all you could know, and because of that you are a liability to yourself, your employer/customer and to our industry.
Wake up guys. It's not a hobby, its a career. Treat it as such.
The reason the IT industry has certifications is primarily a reaction to a problem that I don't see in any other industry. We have a demand that far outpaces the supply and because of that we have a LOT of people who are in the industry in various positions that don't know the elbow from the ...w ell from there foot. What's worse is we have even more people who no enough to be mildly succesful wihtin the narrow field of experience but are lost when they leave their comfort zone.
I'd very much like to see the same kind of review and certification process that CPA's, lawyers, doctors and other professions enjoy (endure?). I'm tired of dealing with people who THINK they know technology when all they realy know is the three things they've worked on since they left school.
Second rant: (if you're already slighted leave now)
I agree with the two other writers here about the tauted "experience" versus "book knowledge" argument. If you taught yourself everything you know then you had a fool for a student. The way to inovation is to build on the efforts, knowledge and experience of others. The experience benefit only occurs when you have the necessary foundation of book knowledge. Otherwise you MAY figure out HOW to accomplish something, but you'll never understand WHY it works or WHAT the ramifications are of your actions.
It is all of you self-taught folks I'd like to weed out of this field. You are the main thing that prevents the rest of us from moving forward.
Feel free to email me personally with your flames. I've heard them all before.
I looked at the wiring diagram posted above and noticed a few things:
1. There is no feed-back capability, thus the machine in question will not receive its own packets. This is a violation of all the IEEE rules and will likely reek havoc on all but the most stable IP stacks.
2. The collision detection pin is diliberately grounded, thus this is a FULL DUPLEX connection.
Both points can be handled IF AND ONLY IF you can configure the device driver to accept a full-duplex configuration, which is unlikely for a driver meant to work across an AUI. You may find that it works, sort of, under light load.
I agree with the other posts, the transceiver is the best alternative. The microtransceivers are running $35 or so, retail.
I think the bridge metaphor breaks down pretty quickly...a bridge is a physical thing that can be objectively measured. A bridge goes through several "inspections" by disinterested or hostile others. A bridge is expected to last 50 to 100 years.
:)
Anyone that ever looked at some of the theoretical implications of say proving the "correctness" of a program knows that the objective inspection is basically impossible. The others are obviously outside the realm of progrmaming. So what is the idea behind the certification/license issue?
Are we trying to prevent studip people from writing code? Then we would need to prevent all the various software warehouses from selling a copy of Visual Stupid (you pick) to any and all suits and programmer-wanna-be's. That would be a good thing for several reasons.
If the idea is to give some malcontent reason for a lawsuit, well we already have enough of them to keep the courts busy for the next several millenia.
I like the idea of the guild though...self directed, self fulfilling, able to bring along the journeyman to the level of master with a known process and a mentoring methodology...Yeah I like that idea
Clearly you are a part of the HR organization.
Cooperation is demonstrated by being available during weekends, wearing a pager continuously, working >40 hours week (w/o OT compensation) and being willing to hop on a plane at a moments notice. These are all clearly described in a simple resume and easily verifiable (Is the candidate wearing the ubiquitous pager during the interview?)
Divulging other personal information, such as marital status, is clearly prohibited by law. Why do the HR and placement (read head-hunter) forces feel they are entitled to this particular piece of VERY personal information. I think countering with a request for the current salaries of other members of the team and/or the budget would put this kind of talk in its proper place.
I've been down this road a few times myself in recent (read last 3 years) history. The times I divulged my current salary I typically got offers that were 5-7% higher than that figure. When I did not, there was no correlation whatsoever. The job I currently hold I put a stake in the ground at a particular figure and refused to mention any other figure in any other context. I got what I asked for, and isn't success more compelling an argument than making the interviewer "feel better" about you.