The paradox with parent is that any new engineer, no matter how qualified, will have to do things "the Microsoft Way". This because of limitations that Microsoft itself created: Patents, licensing and over-marketing all built around a dysfunctional software core immobilised by the same.
It took some real b*lls for apple to scrap their own proprietary core for Unix's - that is what making computers that work better is all about.
Microsoft, on the other hand, owes its fortune to the fact that it managed to "guinea-train" an entire generation of first-time computer users - with an OS core that wasn't even of their own making - by their deal to having it shipped "for free" in every new computer. Turn a computer on, first-time user, and what's the first thing you'll "learn" to use? What you see in front of you.
Even though Microsoft could use their massive profits for researching something better or even new, they've spent so much time on protecting a system based on patents and marketing techniques that they've basically stifled any means for real innovation. Their error is refusing to change from their present path.
Because of Microsoft's history, I can't hide that my first impression was one of doubt (about the efficiency of the Bing algorithm over Google's), but you never know. The thing about search engines is that 99% of what happens in a search isn't visible to the searcher - it is not a function-laden gui - and the chain of operation is simpler, so who knows? If the algorithm *is* better and users get better results, it will be the better product.
My work conditions match EXACTLY those of the author: 1.5 years of service, mid-sized company, and I'm the only IT guy.
Parent is right especially in the last part, but don't always expect a coherent and logical response to even perfectly valid explanations.
Today I'm in developer hell at the end of a one-year project. Also a photographer, I developed the idea (with my boss, the owner of a photo-equipment rental company) of creating a secure-storage service for photographers (who were constantly coming to us with stories of crashed hard disks/computers/media). At the outset I was to be the manager of a new 'project development' division of our company, with a staff of 2-3, but then the recession hit us and I ended up having to build, configure, program and manage the whole server/interface service myself.
My mistake was making promises too early: pressed by the boss for a deadline date, my estimate was based on the time that it would take me to finish the project if I was working only on that; of course I couldn't enter into the calculation all the other stuff I do for the company, so I ended up bypassing deadline after deadline.
After 10 months of 16-hour days and ~200,000â later, the project is complete to the debugging stage, but around one month ago I had developed such an aversion to pressure from the boss (who was constantly telling me that the future of the entire company was on my shoulders) and the size of the task in front of me that I lost all ability to program, test and debug efficiently. At that point even the sound of the boss' voice brought a sinking feeling, and I ended up having something similar to a nervous breakdown.
One week of vacation helped things a bit, and now I'm back at it in a more objective manner.
The most important word in the above is "objective" â" when you're doing a job that no-one else understands, communication tends to lean towards the personal (emotional) and one can get stuck in an almost prejudiced and reactionary attitude rut if things aren't going well.
So keep your cool and just think of what's being asked of you, and avoid reacting to the way it was asked. If you can provide a reasonable answer or solution in a calm, detached way, things will get much cooler/more respectful over time.
Oh, and don't forget the reason most people speak rudely to IT guys: embarrassment at their own ignorance. People often try to mask ignorance under a mask of aggression.
In order to get a complete model of what costs what economically/environment-wise, one must include in their calculation every aspect of a mode of transportation, everything from the energy/cost/pollution needed to create the transportation through its maintenance and management, and not only the energy/pollution needed for the completed mode of transport per se.
For example, most all trains here (France) run on electric power, but most electric power is generated in nuclear power plants, but the creation of the latter required X amount of fossil fuels (mining, construction equipment, other forms of transport for materials and nuclear fuel). So if I wanted to compare this model to, say, air travel, I would have to measure the consumption/pollution created by plane production and plane fuel, and study not just the consumption/transport capabilites of the plane itself.
...with their web-crawling keyword-sculling technology, but it's only normal that someone else was researching what to ~do~ with all the data. IE (data analysis for human comprehension) and Google would make one fierce - and useful - blend.
Exactly. What's the point of having 24million pixels in the camera when you only have a crappy lens? This is why I like Nikon over Canon - consistent, never-changing-nor-degrading glass lens quality. I'm still using some of my F3-era lenses with my D3x with amazing results.
Another question: at what point will a camera's captor pixel resolution surpass lens quality? A 135-format lens has grain; once pixel capacity grows fine enough to capture this, there's not much point in refining further captor-wise. The next step up would be a middle-format captor with a middle-format lens (more glass).
IMHO they're missing the point. How about going all the way: what about shipping computers offering other OS's? Especially if the computer maker and the OS maker are not the same.
The above suggestion is much like the browser issue is to windows - the EU is ignoring the forest in favour of a few trees.
The Cisco 'bug' is an oversight - with its own configuration system (where the actual AS path is written out, not an algorithm treating the same set earlier in a variable), there can be no problem. Cisco does not take into account possible errors (garbage) created by the configuration of other-type routers, thus the problem. True, this also reveals a laziness on the behalf of network engineers who assume that all routers use the dominant Cisco-ish configuration language - not. So what is needed is a means of filtering errored garbage from all platforms and sources, and this job would be most efficient were it undertaken by ISP's.
Did you RTFA? The problem was due to a router misconfiguration - a human error - and a worldwide ISP tendency of not reading/filtering garbage from what they pass along. Not bugs, not upgrades.
IMHO, this the Firefox bundling is kind of missing the point; what about bundling OS's with other-company hardware? Is this not also an imposition of 'choice'?
A bit of reasoning here. Before even beginning to try to lump the question of MS' future into a global "relevance", one must determine what that relevance ~is~. Better still, to better analyse what future market impact their products will have, it is better to examine their product directly vis-Ã-vis a) market shares and sales techniques compared to b) product quality and c) other existing products (themselves analysed through the first two criteria). Sooooo...
a) & c) Microsoft is for sure the major player in the proprietary desktop market. They got there not through promoting/developing product quality, but through marketing techniques (hooking first-time users by shipping pre-installed in most every PC shipped) and its incompatibility with other OS's/media. Although innovative, Mac only became popular through the sale of Mac computers - important, because the market 'staying power' of a product is directly related to its quality - but it only recently began to deal with incompatibility issues. Lastly, "free" *nix has the majority of the Server market, its ground gained through a long evolution and accessible development - again a market share maintained by the quality of the product itself - but these are still largely incompatible with other OS's/media.
b) & c) The quality of their product is improving but negligible compared to other products out there (ease of use, compatibility, security, etc.). For the time being, the only widely-accessible "better" desktop product out there is the proprietary Mac OS, but for the time being it is usable only Mac computers/emulators. This leaves the various *nix distros - for the desktop market, there are many available *nix solutions (Ubuntu, etc.), but these fail on ease of use (installation, etc.) compared to other proprietary OS's.
The direct opposites in the above are MS and *nix - the former gained/holds its share through marketing techniques before quality, and the latter, even if it is free and more 'difficult' to use, depends totally on its quality in maintaining its market share. Mac seems to be the middle ground between the two.
My conclusion: MS will be around as long as a) they are not surpassed in "ease of use" by other other-OS/media compatible PC-installable OS's, b) they continue their 'pre-installed' deal with PC manufacturers and c) the Mac OS continues to be (easily) installable only on Mac computers. If any of these conditions should change, MS is going to be losing market-share big-time.
Agreed ! What percentage of the world's servers run Linux - ~90%? - without the public even being aware of it? Ad = public attention, but the geeks running things are quite aware of Linux already.
I also don't see the point in making an "I'm linux" ad when linux already has the 'market majority' - the 'irony' quotes are there because Linux doesn't 'sell' anything. Yet another reason why no ad is needed.
I see you, and thanks for the tip. Fortunately though, we are not using SAN - each server in a duo-cluster is directly attached to its own storage space, and only the server itself (and its dependencies) are mirrored through DRDB - should either a server or RAID array fail (enter recovery mode), DRDB will activate the failover. This method takes few CPU resources as far as I know, but I'm still in development myself ; )
As far as my case is concerned, your comment is surprisingly on-topic - Have you tried HA or DRDB? You can cluster servers even over remote locations with an instant fail-over and fallback - zero downtime.
The fact that I'm running 64-bit Linux is the very reason I'm not using it yet: Each of our (soon to be paired) servers are managing 14-terra RAID arrays, a capacity perfectly manageable by a 64-bit system, yet DRBD can only manage a disks/partitions no larger than 4 terra. This should change with the next version.
It's all about the software; no point in buying a hydrogen-fuelled car until hydrogen fuel is available.
Yes, you can, through 'mirroring' - either through a second level software raid (say, making a raid-1 out of two raid-5 arrays (or call it raid-51)), or you can do it on the server level by 'mirroring' two computers each managing their own raid array. The first method assures simultaneous synchronisation (but counts on the managing computer's uptime and health); the latter provides a "HA" factor (High Availability), as should either an array or a server go down (even for maintenance/recuperation), the other server will take over, thus giving you uptime.
HA is open source, as are more elaborate HA programs such as DRDB.
Slashdot Mirror
The paradox with parent is that any new engineer, no matter how qualified, will have to do things "the Microsoft Way". This because of limitations that Microsoft itself created: Patents, licensing and over-marketing all built around a dysfunctional software core immobilised by the same.
It took some real b*lls for apple to scrap their own proprietary core for Unix's - that is what making computers that work better is all about.
Microsoft, on the other hand, owes its fortune to the fact that it managed to "guinea-train" an entire generation of first-time computer users - with an OS core that wasn't even of their own making - by their deal to having it shipped "for free" in every new computer. Turn a computer on, first-time user, and what's the first thing you'll "learn" to use? What you see in front of you.
Even though Microsoft could use their massive profits for researching something better or even new, they've spent so much time on protecting a system based on patents and marketing techniques that they've basically stifled any means for real innovation. Their error is refusing to change from their present path.
Because of Microsoft's history, I can't hide that my first impression was one of doubt (about the efficiency of the Bing algorithm over Google's), but you never know. The thing about search engines is that 99% of what happens in a search isn't visible to the searcher - it is not a function-laden gui - and the chain of operation is simpler, so who knows? If the algorithm *is* better and users get better results, it will be the better product.
My work conditions match EXACTLY those of the author: 1.5 years of service, mid-sized company, and I'm the only IT guy.
Parent is right especially in the last part, but don't always expect a coherent and logical response to even perfectly valid explanations.
Today I'm in developer hell at the end of a one-year project. Also a photographer, I developed the idea (with my boss, the owner of a photo-equipment rental company) of creating a secure-storage service for photographers (who were constantly coming to us with stories of crashed hard disks/computers/media). At the outset I was to be the manager of a new 'project development' division of our company, with a staff of 2-3, but then the recession hit us and I ended up having to build, configure, program and manage the whole server/interface service myself.
My mistake was making promises too early: pressed by the boss for a deadline date, my estimate was based on the time that it would take me to finish the project if I was working only on that; of course I couldn't enter into the calculation all the other stuff I do for the company, so I ended up bypassing deadline after deadline.
After 10 months of 16-hour days and ~200,000â later, the project is complete to the debugging stage, but around one month ago I had developed such an aversion to pressure from the boss (who was constantly telling me that the future of the entire company was on my shoulders) and the size of the task in front of me that I lost all ability to program, test and debug efficiently. At that point even the sound of the boss' voice brought a sinking feeling, and I ended up having something similar to a nervous breakdown.
One week of vacation helped things a bit, and now I'm back at it in a more objective manner.
The most important word in the above is "objective" â" when you're doing a job that no-one else understands, communication tends to lean towards the personal (emotional) and one can get stuck in an almost prejudiced and reactionary attitude rut if things aren't going well.
So keep your cool and just think of what's being asked of you, and avoid reacting to the way it was asked. If you can provide a reasonable answer or solution in a calm, detached way, things will get much cooler/more respectful over time.
Oh, and don't forget the reason most people speak rudely to IT guys: embarrassment at their own ignorance. People often try to mask ignorance under a mask of aggression.
In order to get a complete model of what costs what economically/environment-wise, one must include in their calculation every aspect of a mode of transportation, everything from the energy/cost/pollution needed to create the transportation through its maintenance and management, and not only the energy/pollution needed for the completed mode of transport per se.
For example, most all trains here (France) run on electric power, but most electric power is generated in nuclear power plants, but the creation of the latter required X amount of fossil fuels (mining, construction equipment, other forms of transport for materials and nuclear fuel). So if I wanted to compare this model to, say, air travel, I would have to measure the consumption/pollution created by plane production and plane fuel, and study not just the consumption/transport capabilites of the plane itself.
Cobol? Fortran? ; )
...with their web-crawling keyword-sculling technology, but it's only normal that someone else was researching what to ~do~ with all the data. IE (data analysis for human comprehension) and Google would make one fierce - and useful - blend.
...or WebNeanderthal?
PeopleWithTooMuchTimeOnTheirHands - or PeoopleWithMoneyToLose ?
Exactly. What's the point of having 24million pixels in the camera when you only have a crappy lens? This is why I like Nikon over Canon - consistent, never-changing-nor-degrading glass lens quality. I'm still using some of my F3-era lenses with my D3x with amazing results.
Another question: at what point will a camera's captor pixel resolution surpass lens quality? A 135-format lens has grain; once pixel capacity grows fine enough to capture this, there's not much point in refining further captor-wise. The next step up would be a middle-format captor with a middle-format lens (more glass).
ImageMagick can be a RAW-development tool - if you install the ufraw libraries (that in turn use dcraw) and build imagemagick on that.
IMHO they're missing the point. How about going all the way: what about shipping computers offering other OS's? Especially if the computer maker and the OS maker are not the same.
The above suggestion is much like the browser issue is to windows - the EU is ignoring the forest in favour of a few trees.
Er, what's that? ...Does it like "sports"? Does it like "pic-tchaaas"? Nudgenudge winkwink?
...talk about what?
The Cisco 'bug' is an oversight - with its own configuration system (where the actual AS path is written out, not an algorithm treating the same set earlier in a variable), there can be no problem. Cisco does not take into account possible errors (garbage) created by the configuration of other-type routers, thus the problem. True, this also reveals a laziness on the behalf of network engineers who assume that all routers use the dominant Cisco-ish configuration language - not. So what is needed is a means of filtering errored garbage from all platforms and sources, and this job would be most efficient were it undertaken by ISP's.
Did you RTFA? The problem was due to a router misconfiguration - a human error - and a worldwide ISP tendency of not reading/filtering garbage from what they pass along. Not bugs, not upgrades.
...so ISP's should filter AS paths!
They're going to need a few more bits for the root password.
IMHO, this the Firefox bundling is kind of missing the point; what about bundling OS's with other-company hardware? Is this not also an imposition of 'choice'?
A bit of reasoning here. Before even beginning to try to lump the question of MS' future into a global "relevance", one must determine what that relevance ~is~. Better still, to better analyse what future market impact their products will have, it is better to examine their product directly vis-Ã-vis a) market shares and sales techniques compared to b) product quality and c) other existing products (themselves analysed through the first two criteria). Sooooo...
a) & c) Microsoft is for sure the major player in the proprietary desktop market. They got there not through promoting/developing product quality, but through marketing techniques (hooking first-time users by shipping pre-installed in most every PC shipped) and its incompatibility with other OS's/media. Although innovative, Mac only became popular through the sale of Mac computers - important, because the market 'staying power' of a product is directly related to its quality - but it only recently began to deal with incompatibility issues. Lastly, "free" *nix has the majority of the Server market, its ground gained through a long evolution and accessible development - again a market share maintained by the quality of the product itself - but these are still largely incompatible with other OS's/media.
b) & c) The quality of their product is improving but negligible compared to other products out there (ease of use, compatibility, security, etc.). For the time being, the only widely-accessible "better" desktop product out there is the proprietary Mac OS, but for the time being it is usable only Mac computers/emulators. This leaves the various *nix distros - for the desktop market, there are many available *nix solutions (Ubuntu, etc.), but these fail on ease of use (installation, etc.) compared to other proprietary OS's.
The direct opposites in the above are MS and *nix - the former gained/holds its share through marketing techniques before quality, and the latter, even if it is free and more 'difficult' to use, depends totally on its quality in maintaining its market share. Mac seems to be the middle ground between the two.
My conclusion: MS will be around as long as a) they are not surpassed in "ease of use" by other other-OS/media compatible PC-installable OS's, b) they continue their 'pre-installed' deal with PC manufacturers and c) the Mac OS continues to be (easily) installable only on Mac computers. If any of these conditions should change, MS is going to be losing market-share big-time.
Doh! Magmafique. Oh, forget it...
Magnafique!
Agreed ! What percentage of the world's servers run Linux - ~90%? - without the public even being aware of it? Ad = public attention, but the geeks running things are quite aware of Linux already.
I also don't see the point in making an "I'm linux" ad when linux already has the 'market majority' - the 'irony' quotes are there because Linux doesn't 'sell' anything. Yet another reason why no ad is needed.
Well, at least he doesn't throw sh*t. Er, wait... does Vista fall into that category?
I see you, and thanks for the tip. Fortunately though, we are not using SAN - each server in a duo-cluster is directly attached to its own storage space, and only the server itself (and its dependencies) are mirrored through DRDB - should either a server or RAID array fail (enter recovery mode), DRDB will activate the failover. This method takes few CPU resources as far as I know, but I'm still in development myself ; )
As far as my case is concerned, your comment is surprisingly on-topic - Have you tried HA or DRDB? You can cluster servers even over remote locations with an instant fail-over and fallback - zero downtime.
The fact that I'm running 64-bit Linux is the very reason I'm not using it yet: Each of our (soon to be paired) servers are managing 14-terra RAID arrays, a capacity perfectly manageable by a 64-bit system, yet DRBD can only manage a disks/partitions no larger than 4 terra. This should change with the next version.
It's all about the software; no point in buying a hydrogen-fuelled car until hydrogen fuel is available.
Yes, you can, through 'mirroring' - either through a second level software raid (say, making a raid-1 out of two raid-5 arrays (or call it raid-51)), or you can do it on the server level by 'mirroring' two computers each managing their own raid array. The first method assures simultaneous synchronisation (but counts on the managing computer's uptime and health); the latter provides a "HA" factor (High Availability), as should either an array or a server go down (even for maintenance/recuperation), the other server will take over, thus giving you uptime.
HA is open source, as are more elaborate HA programs such as DRDB.