It cracks me up that MS got its footing in 95+ by being "good enough", and cheap. Now that Open Source is "good enough", and, um, "cheaper", there's a problem? Puhleeeeze! Same curve, different "vendor". Give it a think.
I think you guys are missing the obvious. Try writing some asp.net code - you end up with variables declarations that look like this:
System.Configuration.AppSettingsReader appSettingsReader = new AppSettingsReader();
That can be shortened to just:
AppSettingsReader appSettingsReader = new AppSettingsReader();
if you include a 'using System.Configuration' at the top of your module. But still, I'd much prefer to write:
appSettingsReader = new AppSettingsReader();
and let the compiler do the obvious.
It's darn hard sometimes to pick out variable names in those longer declarations because there's text on both sides of the variable name. Implicity typing means variable names can be leftmost on the line, which improves the readability of the code noticably.
What an astounding lack common sense! Who the f*ck chose this configuration? Does the notion "it has to work" every come into play in the bidding process? Management could do everyone a huge favor by firing themselves.
Google Talk... and all that dark fibre Goggle has been buying up? This isn't just about instant messenger - Google is building the next voice communications network! With their new WiFi hotspots - it could be wireless voice communications (at least if you're in a major center).
Bee-bee-boo-boop "Picard to all phone companies: You are being replaced."
In my humble opinion, the quality of these managers is far from what it used to be. Ten, maybe fifteen years ago, IT managers knew their shit. They had already been a DBA, a programmer, and team leader, etc. So they knew how to run the show - having been actors in it prior. And they knew when a vendor was bullshitting them, having been burnt often enough in the past as a techie.
These days I find the IT management ranks swelling with inexperienced idiots who prefer "relationships" with their knee-pad toting vendors rather than an accurate engineering analysis from on-board staff. If a salesman says it, it must be true. If a staff engineer says it, it must be legacy.
How these smarmy, self-interested psuedo-managers got into place with a BA in arts and zip for tech experience I'll never know. But I'm not suprised you're seeing them avoid knowledge accumulation at all costs - it certainly fits the narcissistic model. I suspect these are the children that grew up during the "praise the child at all costs - even when they suck" phase of our educational policy here in North America.
A good stiff spanking and banishment to the lower ranks (where expereience is earned) might be the only cure.
Reliability: Only Windows and Linux are players?
on
Linux Feels Growing Pains
·
· Score: 2, Insightful
'it's got to get the final tier of reliability and predictability that I'm going to bet a multi-billion dollar corporation's future on.'
It sounds like this guy thinks the only choice in "high reliability" is between Linux and Windows. A better education in the computing landscape would serve him well.
If it's really reliability and predicatability he's after, he should explore older, more mature operating systems. Obviously this is not his goal, so it makes you wonder which MS salesperson stuffed those words in his mouth for convenient regurgitation.
Am I the only one that thinks 2.5 meter resolution is pretty sloppy? Or is this a case where 2.5 meters is the "advertised" resolution, and special customers get 0.25 meters?
Not all of the companies involved will be able to make enough money out of these products to stay in business. That means that at some point in the future the market will consolidate and a number of these products will disappear.
Complete bullshit. The companies will disappear, but the product will live on in sourceforge (or where ever), exactly oppositite of what this inexperienced author says. Every customer of the product will have a copy of the source, which at least allows them the option of continuing development and support internally. This simply cannot be said for closed, commericial software.
If I remember correctly, Sun's approach can migrate a thread to another (local) core with almost (or exactly) zero latency.
Intel's HT (which is a Intel-ization of DEC's Simultaneous Multithreading aka SMT), requires time to switch between one "virtual processor" and the other.
True multi-core chips (as opposed to HT stuff), hate to migrate a thread because it's so expensive in terms of time.
But let us not forget the CELL processor (discussed here recently) - which has a collection of special-purpose "cores" - a small array of which will no doubt blow the doors off of the methods described above when it comes to aggregate throughput.
Indeed I am not. For those like me, please explain how the "(declare (optimize (safety 3)))" statement affects the stability of the OS kernel hosting the lisp run-time environment.
You might be on to something here. The pesky IP lawyer I refered to in my previous post also mentioned "submarine patents". Actually, I read it somewhere else and asked him about it. He said "of course".
A submarine patent is a patent on an almost-obvious item that is bound to cause *someone*else* problems in the future. To collect the $$, one would re-surface in the future and demand payment for usurping (or grazing) the patent. Apparently this is ongoing in the genome world, and leaking into the software world.
IBM, despite their significant contributions to the FOSS world lately, led the software world in patents in 2004 (please correct me if I'm wrong here), as well as many recent years. Some say they invented this game (or perhaps the game was thrust upon them?).
So as bani suggests, is the FOSS world at a significant disadvantage because they don't have a "holding company" in which to deposit patents - a company that can threaten IP-retaliation when needed?
It's unclear to me what "optimized for stability" means. Optimization typically refers to the efforts the compiler goes to to either increase the speed of execution, or reduce the memory requirements of the executable. "Stability" is not a switch on any compiler I'm aware of.
However, software engineers, when deciding exactly how to design their kernel (or other software component), have decisions to make - and some of them are a trade-offs between speed, size, and stability. So it might be fair to say engineers can optimize for stability, but really only compilers optimize for the speed/size equation. And here is where some alert reader will point out that much of the speed of any software comes from smart software design - and attribute that could be called "optimization". But of course, this is design optimization, not compile-time optimization.
With that said, I can't agree that "OS kernels are optimized purely for speed". Some kernels are not - small memory footprint counts in embedded applications. And that's why the GCC compiler has switches for size and speed - it's the engineer's choice when compiling.
According to an IP lawyer I once had beer with, and since confirmed numerous times in the press, most large software companies have violated each others' patents - numerous times. But since everyone has "injured" everyone, most no one sues - since they will simply get counter-sued on a different patent.
That said, it seems obvious that the MS kernel, like probably all other kernels, contains IP violations. The difference is that only the FOSS kernels are open to public scrutiny.
So commerical companies can keep their IP violations largely secret, while FOSS ones get vetted publicly. Anyone wanna bet whether a closed kernel or open kernel has more IP violations? Does anyone believe that the MS kernel is completely free of IP violations? Anyone wanna buy a bridge?
In my experience, the manager's ego issues determine his effectiveness. Some think they're god's gift to the technical team - and they eventually (none too soon!) drown in their own bullshit. The good ones realize they know shit about the details, and listen carefully to their top technical people. They tend to do well, get promoted, and leave a vacancy soon to be filled with an idiot from the first category.
I initially tried writing PHP in emacs, which I use on Unix systems anyway. But after using ASP for a couple of years, my head got pretty addicted to two features: syntax color coding, and auto-complete. These rather modern features make the code (much) easier to read, and reduce typing and typos rather significantly. Why they never put in macro key abilities I'll never figure out.
Exepting vi (which I'm pretty sure has neither feature), I haven't used the other editors you mentioned - but thanks for the suggestions!
Even if the desired optimizations are not obtained on each platform this mythical code runs on, the readability of the code has value.
Since processors seem to double in speed each time my "next" project starts, the expenditure seems affortable, and in my opinion, well worth the potential performance loss on any particular platform.
The real-time guys with their special compilers have other ways to deal with optimization.
Perhaps a simpler way of putting this: Gigahertz are cheap - fill your boot!
I've written applications in both - and here's a difference no one talks about. When you open up MS's ASP environment, all that great GUI stuff is there and it's pretty easy to get going. Then as often happens in a development environment, you need a quick script to munge a long list of field names. Is ASP your first choice? It wasn't mine, because I couldn't find a way to get input into/out of it from the command line. So I whipped up a temporary web page with a text box to do it. More overhead than I wanted to spend for what should be a 2 minute job given an editor with macro key abilities.
Then a couple of years later I built my first app in PHP. The first thing I noticed was how easy it was to script from the command line. Since I'm not a perl junkie, it was real useful for small scripting jobs. I'd use a shell language for this, but fankly, I'd rather poke a fork in my eyes.
The next thing I noticed in PHP was I needed an modern editor (the free download doesn't come with an IDE), so I bought one from zend.com for a couple of hundred bucks. It's getting better, but like ASP, it too has no macro key ability (maybe I'm wrong and someone will tell me?), and other nits I'd pick given the chance.
But the big discovery in PHP was that all my ASP data-type problems magically went away. Hours and freaking hours I spent debugging situations where an int was returned from a DLL and ASP string'ed it, or vice versa. There were byref/byval issues I recall as well. We had to build test local harnesses for all our middle tier ASP components because these problems rendered ASP too lame for a debugging platform.
But my original point is really that PHP is useful along a continium of the problem space. Need a quick script? Need a nightly job that cleans up your app? Need web pages? PHP works well for all. ASP, from my experience, hits one for three.
1) Protected mode Windows consumes significantly more processor than DOS/Windows 3.1. Protected mode costs cycles to resolve virtual memory addresses. I seem to recall that loading a physical address into a 80286 (back in the day) register took a cyle or two. Loading a "selector" (a protected-mode memory reference) took 20 (or was it 50?) cycles on the 286. We took a huge hit on cycles for this one. And those selectors live on in today's x86 line.
2) Bad design has gotta account for many slowdowns in today's Windows UI. Get on any modern Windows box, open multiple applications that use the network, and you'll see one app block as it waits for some network resource that's been locked by the other. Obviously locks are needed, but fine-grained locking is critical any code (like a tcp/ip stack) that services multiple threads of exectution. My (non-scientific) observations indicate that there's work to be done on the tcp/ip stack in MS's products. The same is likely true on other platforms, but Windows seems to be a consistent mal-performer.
Some of us in NYC order our groceries from freshdirect.com. Their ordering web site is where I keep my grocery list - no need for the paper-refridgerator-magnet thing anymore. But I'm a lazy bastard who hates to walk to the other room to put butter on the list. Gimme a browser in my kitchen and my life would be complete! (A beer fridge in the living room wouldn't hurt either).
Everyone else has talked up the legal options, so allow me to point this out: The author of the software package (that recommended the rip-off firm) probably has more of a reputation to protect. Ask him to make this situation right - either by fixing the bug, or applying pressure (loss of future referrals?) to the consulting company. Threaten to name the HIM here in Slashdot if he won't help you - he may respond better than a consulting company an abstract name.
If service providers offered configuration management services, would they then pressure Microsoft to, when adding features, make those features more amenable to provider-based management rather than end user based management?
Does Microsoft invest in any bandwidth providers? Should future investments in this direction make us nervous?
What about the mono-culture problem? When a provider applies the latest patch and clicks the wrong button, will a million PC's get trashed? Who's going to visit all those homes? Will grandpa have to wait until Microsoft and the provider duke it out in court before someone drops by to re-install Windows?
When it comes to stability, the big OS's have tools the little ones don't - the ability isolate the misbehavior of one application from others. Within big isolation containers (partitions), quota limits are important to prevent one process from affecting others.
Maybe part of reviewing and testing should include tests to ensure a second, mis-having application can't affect the one under the microscope (and vice versa?).
Time to re-read the thread. I've never mentioned Linux or Open Source. I've only bashed Microsoft's one-click facilitation of the trashing an entire network of pc's. As demonstrated in last week (links elsewhere in this thread), it's a hole someone will fall into eventually. It's a matter of probability, nothing more. You should really stop trying to bait me into a Linux vrs. MS argument now, as the repetoire of some of us includes numerous other operating systems, some of which could be put to good work "layering" a corporation's resources to prevent monoculture-itis. (this is my actual "point")
Slashdot Mirror
It cracks me up that MS got its footing in 95+ by being "good enough", and cheap. Now that Open Source is "good enough", and, um, "cheaper", there's a problem? Puhleeeeze! Same curve, different "vendor". Give it a think.
I think you guys are missing the obvious. Try writing some asp.net code - you end up with variables declarations that look like this:
System.Configuration.AppSettingsReader appSettingsReader = new AppSettingsReader();
That can be shortened to just:
AppSettingsReader appSettingsReader = new AppSettingsReader();
if you include a 'using System.Configuration' at the top of your module. But still, I'd much prefer to write:
appSettingsReader = new AppSettingsReader();
and let the compiler do the obvious.
It's darn hard sometimes to pick out variable names in those longer declarations because there's text on both sides of the variable name. Implicity typing means variable names can be leftmost on the line, which improves the readability of the code noticably.
What an astounding lack common sense! Who the f*ck chose this configuration? Does the notion "it has to work" every come into play in the bidding process? Management could do everyone a huge favor by firing themselves.
Google Talk ... and all that dark fibre Goggle has been buying up? This isn't just about instant messenger - Google is building the next voice communications network! With their new WiFi hotspots - it could be wireless voice communications (at least if you're in a major center).
Bee-bee-boo-boop "Picard to all phone companies: You are being replaced."
In my humble opinion, the quality of these managers is far from what it used to be. Ten, maybe fifteen years ago, IT managers knew their shit. They had already been a DBA, a programmer, and team leader, etc. So they knew how to run the show - having been actors in it prior. And they knew when a vendor was bullshitting them, having been burnt often enough in the past as a techie.
These days I find the IT management ranks swelling with inexperienced idiots who prefer "relationships" with their knee-pad toting vendors rather than an accurate engineering analysis from on-board staff. If a salesman says it, it must be true. If a staff engineer says it, it must be legacy.
How these smarmy, self-interested psuedo-managers got into place with a BA in arts and zip for tech experience I'll never know. But I'm not suprised you're seeing them avoid knowledge accumulation at all costs - it certainly fits the narcissistic model. I suspect these are the children that grew up during the "praise the child at all costs - even when they suck" phase of our educational policy here in North America. A good stiff spanking and banishment to the lower ranks (where expereience is earned) might be the only cure.
'it's got to get the final tier of reliability and predictability that I'm going to bet a multi-billion dollar corporation's future on.'
It sounds like this guy thinks the only choice in "high reliability" is between Linux and Windows. A better education in the computing landscape would serve him well.
If it's really reliability and predicatability he's after, he should explore older, more mature operating systems. Obviously this is not his goal, so it makes you wonder which MS salesperson stuffed those words in his mouth for convenient regurgitation.
Am I the only one that thinks 2.5 meter resolution is pretty sloppy? Or is this a case where 2.5 meters is the "advertised" resolution, and special customers get 0.25 meters?
Not all of the companies involved will be able to make enough money out of these products to stay in business. That means that at some point in the future the market will consolidate and a number of these products will disappear.
Complete bullshit. The companies will disappear, but the product will live on in sourceforge (or where ever), exactly oppositite of what this inexperienced author says. Every customer of the product will have a copy of the source, which at least allows them the option of continuing development and support internally. This simply cannot be said for closed, commericial software.
If I remember correctly, Sun's approach can migrate a thread to another (local) core with almost (or exactly) zero latency.
Intel's HT (which is a Intel-ization of DEC's Simultaneous Multithreading aka SMT), requires time to switch between one "virtual processor" and the other.
True multi-core chips (as opposed to HT stuff), hate to migrate a thread because it's so expensive in terms of time.
But let us not forget the CELL processor (discussed here recently) - which has a collection of special-purpose "cores" - a small array of which will no doubt blow the doors off of the methods described above when it comes to aggregate throughput.
-spinLock
Indeed I am not. For those like me, please explain how the "(declare (optimize (safety 3)))" statement affects the stability of the OS kernel hosting the lisp run-time environment.
You might be on to something here. The pesky IP lawyer I refered to in my previous post also mentioned "submarine patents". Actually, I read it somewhere else and asked him about it. He said "of course".
A submarine patent is a patent on an almost-obvious item that is bound to cause *someone*else* problems in the future. To collect the $$, one would re-surface in the future and demand payment for usurping (or grazing) the patent. Apparently this is ongoing in the genome world, and leaking into the software world.
IBM, despite their significant contributions to the FOSS world lately, led the software world in patents in 2004 (please correct me if I'm wrong here), as well as many recent years. Some say they invented this game (or perhaps the game was thrust upon them?).
So as bani suggests, is the FOSS world at a significant disadvantage because they don't have a "holding company" in which to deposit patents - a company that can threaten IP-retaliation when needed?
- spinLock
Thanks for the tip - much appreciated! I hunted around, and found a few sites that talked about the PHP Eclipse plug in, but nowhere could I find:
1) Feature List
2) Download
Got any pointers?
Many thanks,
- spinLock
It's unclear to me what "optimized for stability" means. Optimization typically refers to the efforts the compiler goes to to either increase the speed of execution, or reduce the memory requirements of the executable. "Stability" is not a switch on any compiler I'm aware of.
However, software engineers, when deciding exactly how to design their kernel (or other software component), have decisions to make - and some of them are a trade-offs between speed, size, and stability. So it might be fair to say engineers can optimize for stability, but really only compilers optimize for the speed/size equation. And here is where some alert reader will point out that much of the speed of any software comes from smart software design - and attribute that could be called "optimization". But of course, this is design optimization, not compile-time optimization.
With that said, I can't agree that "OS kernels are optimized purely for speed". Some kernels are not - small memory footprint counts in embedded applications. And that's why the GCC compiler has switches for size and speed - it's the engineer's choice when compiling.
- spinLock
According to an IP lawyer I once had beer with, and since confirmed numerous times in the press, most large software companies have violated each others' patents - numerous times. But since everyone has "injured" everyone, most no one sues - since they will simply get counter-sued on a different patent.
That said, it seems obvious that the MS kernel, like probably all other kernels, contains IP violations. The difference is that only the FOSS kernels are open to public scrutiny.
So commerical companies can keep their IP violations largely secret, while FOSS ones get vetted publicly. Anyone wanna bet whether a closed kernel or open kernel has more IP violations? Does anyone believe that the MS kernel is completely free of IP violations? Anyone wanna buy a bridge?
In my experience, the manager's ego issues determine his effectiveness. Some think they're god's gift to the technical team - and they eventually (none too soon!) drown in their own bullshit. The good ones realize they know shit about the details, and listen carefully to their top technical people. They tend to do well, get promoted, and leave a vacancy soon to be filled with an idiot from the first category.
I initially tried writing PHP in emacs, which I use on Unix systems anyway. But after using ASP for a couple of years, my head got pretty addicted to two features: syntax color coding, and auto-complete. These rather modern features make the code (much) easier to read, and reduce typing and typos rather significantly. Why they never put in macro key abilities I'll never figure out.
Exepting vi (which I'm pretty sure has neither feature), I haven't used the other editors you mentioned - but thanks for the suggestions!
Even if the desired optimizations are not obtained on each platform this mythical code runs on, the readability of the code has value.
Since processors seem to double in speed each time my "next" project starts, the expenditure seems affortable, and in my opinion, well worth the potential performance loss on any particular platform.
The real-time guys with their special compilers have other ways to deal with optimization. Perhaps a simpler way of putting this: Gigahertz are cheap - fill your boot!
I've written applications in both - and here's a difference no one talks about. When you open up MS's ASP environment, all that great GUI stuff is there and it's pretty easy to get going. Then as often happens in a development environment, you need a quick script to munge a long list of field names. Is ASP your first choice? It wasn't mine, because I couldn't find a way to get input into/out of it from the command line. So I whipped up a temporary web page with a text box to do it. More overhead than I wanted to spend for what should be a 2 minute job given an editor with macro key abilities.
Then a couple of years later I built my first app in PHP. The first thing I noticed was how easy it was to script from the command line. Since I'm not a perl junkie, it was real useful for small scripting jobs. I'd use a shell language for this, but fankly, I'd rather poke a fork in my eyes.
The next thing I noticed in PHP was I needed an modern editor (the free download doesn't come with an IDE), so I bought one from zend.com for a couple of hundred bucks. It's getting better, but like ASP, it too has no macro key ability (maybe I'm wrong and someone will tell me?), and other nits I'd pick given the chance.
But the big discovery in PHP was that all my ASP data-type problems magically went away. Hours and freaking hours I spent debugging situations where an int was returned from a DLL and ASP string'ed it, or vice versa. There were byref/byval issues I recall as well. We had to build test local harnesses for all our middle tier ASP components because these problems rendered ASP too lame for a debugging platform.
But my original point is really that PHP is useful along a continium of the problem space. Need a quick script? Need a nightly job that cleans up your app? Need web pages? PHP works well for all. ASP, from my experience, hits one for three.
I think the author missed some biggies:
1) Protected mode Windows consumes significantly more processor than DOS/Windows 3.1. Protected mode costs cycles to resolve virtual memory addresses. I seem to recall that loading a physical address into a 80286 (back in the day) register took a cyle or two. Loading a "selector" (a protected-mode memory reference) took 20 (or was it 50?) cycles on the 286. We took a huge hit on cycles for this one. And those selectors live on in today's x86 line.
2) Bad design has gotta account for many slowdowns in today's Windows UI. Get on any modern Windows box, open multiple applications that use the network, and you'll see one app block as it waits for some network resource that's been locked by the other. Obviously locks are needed, but fine-grained locking is critical any code (like a tcp/ip stack) that services multiple threads of exectution. My (non-scientific) observations indicate that there's work to be done on the tcp/ip stack in MS's products. The same is likely true on other platforms, but Windows seems to be a consistent mal-performer.
Some of us in NYC order our groceries from freshdirect.com. Their ordering web site is where I keep my grocery list - no need for the paper-refridgerator-magnet thing anymore. But I'm a lazy bastard who hates to walk to the other room to put butter on the list. Gimme a browser in my kitchen and my life would be complete! (A beer fridge in the living room wouldn't hurt either).
o-sl-o
Everyone else has talked up the legal options, so allow me to point this out: The author of the software package (that recommended the rip-off firm) probably has more of a reputation to protect. Ask him to make this situation right - either by fixing the bug, or applying pressure (loss of future referrals?) to the consulting company. Threaten to name the HIM here in Slashdot if he won't help you - he may respond better than a consulting company an abstract name.
If service providers offered configuration management services, would they then pressure Microsoft to, when adding features, make those features more amenable to provider-based management rather than end user based management?
Does Microsoft invest in any bandwidth providers? Should future investments in this direction make us nervous?
What about the mono-culture problem? When a provider applies the latest patch and clicks the wrong button, will a million PC's get trashed? Who's going to visit all those homes? Will grandpa have to wait until Microsoft and the provider duke it out in court before someone drops by to re-install Windows?
Relying on Windows for security is like fighting for peace, or screwing for virginity. 'Nuff said.
When it comes to stability, the big OS's have tools the little ones don't - the ability isolate the misbehavior of one application from others. Within big isolation containers (partitions), quota limits are important to prevent one process from affecting others.
Maybe part of reviewing and testing should include tests to ensure a second, mis-having application can't affect the one under the microscope (and vice versa?).
Time to re-read the thread. I've never mentioned Linux or Open Source. I've only bashed Microsoft's one-click facilitation of the trashing an entire network of pc's. As demonstrated in last week (links elsewhere in this thread), it's a hole someone will fall into eventually. It's a matter of probability, nothing more. You should really stop trying to bait me into a Linux vrs. MS argument now, as the repetoire of some of us includes numerous other operating systems, some of which could be put to good work "layering" a corporation's resources to prevent monoculture-itis. (this is my actual "point")