Check out ExIf data. Digital Cameras have been storing tons of meta data for years. Shutter Speed, Lens Speed, tons of other little things.... including date/time.
Intelligent usage of the System.gc() method can result in collection and cleanup that results in optimal usage.
However it is not really valid to say that calling System.gc() is the same as deleting ONE indiviudal object in C++. Calling System.gc() has a lot of overhead that the garbage collector can better deal with. Direct client calling of the garbage collector is not a good method of testing performance. Better to pass in some parameters to the Java VM telling it the max heap size allowed and it will constarin itself to that memory limit. Much better than just calling System.gc() and saying its the same as calling delete in C++.
Yeah, you want to maintain the golden air/fuel mixture always. But, generally speaking, lugging the engine )IE a real low gear at low speed (like 4th at 20) with the accelerator at WOT will use more fuel. You can sit with a laptop hooked up to your OBD II connection and watch your fuel injector load go up while your still putzing along if you do a low gear at low speeds. (I mean low by the numbers, not bu the 1-5 numberings).
Ont he other hand if you optimize your shift points so that the engine output is at its best with a relatively constant amount of fuel consumption you can accelerate more efficiently. I know all this is true by observing the sensors and such, but im not precisly sure of the phsyics involved.:)
Well, I think in an absolute since you are right, but it oversimplifies the physics to a childish level.
For example, here are two scenarios:
1. I shift right below the peak torque output for my vehicle (2800 rpms). I shift this way consistently and I accelerate at a modest and average pace.
2. I shift at about 1500 RPMS where my vehicle barely has any real torque (relative to the rest of the RPM range).
Which scenario is more likely to be fuel efficient in city driving? Scenario 1. Why? Because in scenario 2 I am lugging my engine. I am so underutilizing my engine that my vehicle must lug the engine injecting a LOT more fuel just to keep it going and compensate for what I am doing. The gas usage goes way up. At 2800 RPMS my vehicle is accomplishing more work per unit of gas burned than it was at 1500 RPMS. So for NORMAL driving and average city driving it becomes more fuel efficient if I shift in the powerband of my vehicle.
Now, in interstate driving, once you have overcome the initial acceleration part the goal is to have your vehicle at a nice comfie RPM spot that will still give you decent power (enough to accelerate a little) without having to down shift while still maintaining really good fuel efficiency. So it makes since then, to stay at 1500 RPMS since we aren't really lugging the engine as we aren't pressing the accelerator down much at all. In my car staying around 1800-2000 RPMS in the interstate affords me optimal gas mileage with reasonable acceleration. To maintain optimal low end acceleration I use a 6spd manual transmission. The gearing has a really wide range which still gives good gearing onthe low and high ends. To much gear and you overrev on the interstate and waste fuel. Not enough gear and you underrev everywhere but the interstate and you lose a lot of acceleration potential. Finding the sweet spot in gearing and designing an engine with the appropriate torque and horsepower curves is a real trick.
Oh one other neat thing those new 300+ HP hemi's do while cruising on the interstate. They shut down some of the cylinders in the engine so they use only 4 cyldiners improving fuel economy by 20%. (they are still big cyldiners after all).
P.S. I have a 415 HP V8 with a turbo that gets 18city/23highway mpg:)
Not if its purely over the Internet. I think as long as it stays digital it should stay unregulated.
THat would be like regulating instant messaging, they do something similar to what phones do, right? Allow for communication with other people. I guess we should regulate video conferencing too, since it does what phones do and more. I think its silly to regulate purely digital communication mediums.
VoIP that hits a physical resource such as the standard phone system, yes, they can be regulated. Not all VoIP is moving into an existing market, but they are also creating a market that never really existed before.
Whatever, I bet the sales guy is financially secure from here out after a few years at the beast. Do your time there, walk out with some cash!:)
Integrity or not having to worry about money for the rest of your life? Especially when your a sales person anyway! If it were an influential Linux developer like Linux, or Alan cox, I would raise an eyebrow or two.
I suppose we should get into one of those circular arguments here.
Here I'll start it, "but oh wait, I just said that was a Micro level issue. Talking about one person when we are viewing this whole thing from a more macro perspective".
"You reply, it doesn't matter! Think about the children"
I read some articles that debunked that. Outsourcing has allowed margins on wholesale goods and such to drop lower. If goods and services can be purchased for less it stimulates the economy and helps it grow. It actually makes sense at a macro level. I think oursourcing is a macro level economic issue and looking at it in a micro level can lead to misperceptions about where the money really is.
So people can still spend money buying toys for their kids because they are less expensive while still having a good quality level, so people can still spend.
Well, lets assume on a graphics heavy site we save 50KB just doing some simple image optimizations and squeezing the images here and there spending maybe 4 hours just optimizing the graphics. That is more like saving 50 GB per day, or 1200 dollars. I doubt you would pay 1200 dollars for the optimization.
However, in a more real world scenario sites raking in that many hits generally save money by going to a "real" hosting solution and just paying for a pipe size with unlimited data transfer. Generally having a 1-2Mb(it) pipe with unlimited transfer is just easier to deal with than a limited amount of bandwidth. Most sites with a lot of traffic have an unlimited transfer with burstable capability to handle spikes in traffic.
That isn't a very creative approach to dealing with problems in modern web applications.
Whether its one record of twenty I don't see how it really affects the data layer. If your application layer is written properly it won't matter about the data layer so much. Your application layer expects a recordset with certain data. Of course having n records instead of 1 you will have to change the way the application layer deals with it. That doesn't mean you still can't view the data from a SQL server in a very generic way (a recordset, or multiple recordsets). Your application layer should be smart enough to convert SQL data types into native types. And if your application is properly normalized and you aren't using any properitary column types it will be that much easier. You can take an OO view to your recordsets or a procedural view.. whatever fits best with the application layer's current design model.
Yes, abstraction has a cost. However in a practical sense there are platforms out there that already provide this level of abstraction and they support some incredibly busy sites. (Like ColdFusion powering the Macromedia site). So I think for all intents and purposes the abstraction is just fine for all but the most intense applications. In the real world having a flexible architecture is often more important than raw performance. I stand by my statements that abstracting SQL database access down to recordsets that have only native types to the language dealing with the data is the way to go with most applications. There are always exceptions. I am not really talking about those specific exceptions, just in general.
One final point of yours I would like to address is that you say it will take more time when you change the way the data needs to be accessed. It depends, if we are talking changing SQL platforms and optimizing for another SQL platform.. then no it is much easier to have a specific data access layer that you change. That is the ONLY place you should have to change code when changing SQL servers (save things like stored procedures and triggers). Your application layer should see the SQL server as a black box and have no inherent knowledge about the data access layer. If we are talking a complete paradigm shift over to something like an XML database or some other different data paradigm, yes you would be talking about changing your data access paradigms in the application layer as well.
I know its not quite so clear cut in a lot of cases, but simple SQL server abstraction is good. If you don't abstract your data access layer you are making your code that much more DIFFICULT to modify when you have to change database platforms.
The one big area of optimization that this kind of misses is stored procedures and triggers and such. You just can't abstract away something that is IN the database layer of your app. When you need the raw performance of a stored proc that is an optimization choice you must make. When you move logic away from your application layer and towards your data layer I think you need to have specific performance reasons to do so, because.. again.. most apps don't really need that much power. You should definitely be aware of the full range of such optimization techniques, because if you don't know about them you won't have knowledge of what they are and how they work in your application design phase.
I think we both agree on the pragmatic aspects of this, and that is such a heavily architected approach is not always required, especially for small projects that need to be done quick and fast. As your requirements grow larger so, typically, does your need for better and more flexible architecture. The bigger the app the more important it is to perform risk analysis. The more important it is to have a more defined and developed architecture. Most apps, especially open source apps that aim to have borad appeal, can benefit from an abstracted data access layer with very few data access platform specific optimizations. (And if you DO have such optimizations it makes sense to provide an un-optimized application layer only version of code in stored procs and such).
Sorry, I don't buy it. Even with the SQL hacks in any application of even moderate size you can hide your data layer from your application well enough that you would still only be rewriting the data access layer of your app. That is what abstraction is all about.
I know that you will invariable have to optimize for your current DB platform to take advantage of unique optimizations and features of that platform, but that still does not mean you can't do it in a way that keeps all of the optimization details hidden from the rest of the system. This is true in almost all cases where you are using a SQL server as your data storage layer.
Trivial apps will often not even bother with abstraction. It just seems silly to me given the handy abstraction tools via PEAR and Native libraries in PHP that anyone would not use them in a new project (versus the native calls). The only reasons (performance reasons) can still be architected away with just a little extra effort and some careful planning.
Jabber.. free server.. free clients.. all code available to ensure no gremlins if you are in a tight environment (Like financial companies that trade stocks for folks etc) and require high deals of confidence/confidientality.)
Uhm, so.. you mean to tell me that you expect a daily driven engine to last for 7+ years with the piston rings perfectly intact?
I drove a 10+ year old ford and my piston rings and pistons and heads were all in great shape. (Seals went bad from the car sitting to long).. but I don't buy this myth that Ford makes especially poor engine designs etc.
There aren't many 7 year old cars OUT there that don't leak a little patch of oil or burn a little off if they are on the original engine. The ones that DONT leak oil are the ones that had their filters serviced regularly, all the TSBs were followed for that car, and the oil was changed regularly. Most people NEVER do that. If they did you would find that the engines really last about the same length of time.
I read a study once that took a general health check of a group of cars over a 10 year period. It was comparing foreign cars (ALA: BMW/Mercedes) to domestics.. mostly fords. It found that the people with the more expensive cars were more likely to do the required maintenance. It also found that the domestic owners that followed the required maint schedules had cars in good shape, just like the foreign cars. Hummmmn. Now lets apply occams razor? Which seems more likely.. that there is some majorly defective thing about fords engineering that causes them to "smell like oil??" or that the users of said cars who paid much less for them simply don't care as much about the cars and thus dont do the maint.
Something to think about
For the record I have a 1987 2.3L 4cyl mustang with nary an oil leak and minimal oil consumption on completely original stuff. The trick? It was well taken care of with regular maint.
It's pretty easy for me. I spend my time writing code and am compensated appropriately for my time spent working on said code.
An employer puts a value on how skilled you are and compensates you for your time in monetary means. The more your perceived value the more you make. I like my perceived value. I like eating and having nice things, thus I will work. My job just happens to be programming? So? It is still a job. Its not some mystery.. as an employer if you want something you use the ultimate motivator: money. You can't just go post something to a bulletin board and watch as the teeming hoard of OSS geeks magically code it for you.. if it doesn't exist you have to pay. If it exists and you want it modified you have to pay. Notice I said if *YOU* want someones time. And who would simply volounteer their time to someone for no pay? Most OSS folks do what interests them:)
Sure its all nice if some group of OSS geeks happened to have coded something that meets 70% of your requirements, but it just doesn't change the way the world works. If you want someone's time it costs money usually. The more skilled the person the more that time costs. Case closed. No one can live without money.
You can cook up all kinds of ideas for how to make money with OSS, but I just don't see it working for the masses, yet. Anyway.
What about a program used to create a video of programmers used to portray actors as programmers who are... and what if we had a beowulf cluster os said programers working in perfect harmony...
Ahh, yes but one programmers representation of an orderly set of bytes to accomplish a task can be quite different from another programmers set of bytes that accomplish the same task. What if the end result is not the art? What if the art is in the creation and representation of what creates the end result?
So, even though in the end you ARE bound by a structure of some sort I think that creative thought processes lead to better representations of said orderly structure (code). It most definitely is NOT a traditional form of art, but when I sit down and look a particularly well written function that is simple and "elegant" it inspires me in a great way. Almost the same way as when I see a great photograph that has several complex "artistic" elements involved (IE: Not just a snapshot, a serious well done picture meant to cause you to think).
Since I also enjoy photography and auto mechanics quite a bit I see parallels in ways most people would just discard. Photography is closer to "traditional" art. And I feel inspired the same way when I see a great photograph. Or when I see someone who is doing serious non factory work on their 800hp car and they come up with an "out of the box" solution. Or when I see someone lay some code out so well and so ingeniously that I think, "how would you do it any other way?". It is all inspiring at some level.
So what this boils down to is that art is in the eye of the beholder, truly. Just because it is not traditional art does not mean it isn't art to others. I look at some paintings considered masterpieces, even with the prerequisite knowledge on why this picture is great and have a hard time accepting the creative genius in certain aspects of a painting. The creative genius in a well done piece of code seems much more obvious to me, thus much more artful.
Its all a matter of perspective, as so much is in life.
It is nice that we are talking about a DICTIONARY word here. The legal definition of a word and the common use definition often vary quite wildly. There is also a legal definition of terrorism. Ticking off three or four things from ONE dictionaries interpretation of the word terrorism just doesn't even work.. save for maybe casual discussion like this, but I wouldn't get to upset, especially if you are trying to equate a dictionary definition to a legal definition.
Jeremy
Re:PostFix + MYSQL + Cyrus Rocks!!!
on
Postfix
·
· Score: 2, Informative
One better...:)
I had also known Sendmail was a little tedious to learn. My main job is software development, but we are a small company so I multi task as system admin for about 10 systems, mostly Linux with a couple of internal windows systems.
None of the email systems are REAL painful to get working, even Sendmail. I can learn and understand these types of things easily with the years of experience I have. The thing is in a small company I have to squeeze every minute I can out of a day because my time means a lot. If I can spend half the time learning the ins and outs of a particular server app I am doing a good thing. So when I had to set up our first non shared systems in a data hosting facility I was very happy.
I decided QMail had enough of a popular following and a reasonable enough featureset and great security track record that we could live with QMail. So I installed QMail, read books, tweaked and eventually installed. Got everything working: SMTP+POP3 with selective relaying based on who had just popped. Worked great. Then it came time to set QMail up on a different server as we were shuffling services around to free up the server we currently used for mail. I was not looking forward to doing QMail again. I didn't like managing it the way we had it set up and didn't really feel I had the time to mess around with it, so I looked for alternatives.
I found Postfix. I downloaded it, compiled it, installed it and read through all of the configs and docs.Within an hour I was so amazed at the simplicity of Postfix and how much sense the configs made, I was in shock. Then I read a little more and searched around and found how easy IMAP was to get working as well. I then found an IMAP server that supported MySQL and found that Postfix also supported MySQL for domains/user configurations. That is all it took for me to be completely sold. I would have preferred more options for the RDBMS, but I wasn't going to complain much.. MySQL isn't to resource hungry. Performance wasn't a concern as I only needed mail for a pretty small group of people. As it is this setup is very READ heavy with the data being almost static (just needing a nice easy programmatic way to be updated when required) so it plays nicely into MySQLs performance forte.
In one day of configuration and code writing I had Postfix doing all of our SMTP, Courier IMAP (which includes POP3) all on the same backend. I had web mail setup and working. All of our mail storage was now centralized which only makes it easier given some of us travel around a lot and still need our mail.The best part, I wrote a web based management tool to manage domains, accounts, and aliases for our mail system. No more unix command line, log onto the application fill in a form or two and its done. Setup a new domain for email? Fill in a few forms and voila it's all done. I think it was one of the most pleasant experiences setting up a network service I administer I have ever had. (I realize that I am a programmer and not everyone will have the knowledge to hammer out a management tool in a small timeframe like that, but for us it simply works).
Thats my story for Postfix!
(P.S. we just have our system send a welcome email and the Maildir folders automaticaly get created after the first incoming message. Postfix handles this for you.:) )
Slashdot Mirror
*sigh*, militarily Iraq was in essence conquered and the the last few minor (on a large scale) pockets of resistance are being eliminated.
Nothing at all like the scale and scope of the loss and conflict in Vietnam.
Jeremy
Check out ExIf data. Digital Cameras have been storing tons of meta data for years. Shutter Speed, Lens Speed, tons of other little things.... including date/time.
Jeremy
There is something to be said for a controlled chaos ;)
Jeremy
Intelligent usage of the System.gc() method can result in collection and cleanup that results in optimal usage.
However it is not really valid to say that calling System.gc() is the same as deleting ONE indiviudal object in C++. Calling System.gc() has a lot of overhead that the garbage collector can better deal with. Direct client calling of the garbage collector is not a good method of testing performance. Better to pass in some parameters to the Java VM telling it the max heap size allowed and it will constarin itself to that memory limit. Much better than just calling System.gc() and saying its the same as calling delete in C++.
Jeremy
Really?
Id love one
email: jallen@elliptiq.com
Jeremy
Yeah, you want to maintain the golden air/fuel mixture always. But, generally speaking, lugging the engine )IE a real low gear at low speed (like 4th at 20) with the accelerator at WOT will use more fuel. You can sit with a laptop hooked up to your OBD II connection and watch your fuel injector load go up while your still putzing along if you do a low gear at low speeds. (I mean low by the numbers, not bu the 1-5 numberings).
:)
Ont he other hand if you optimize your shift points so that the engine output is at its best with a relatively constant amount of fuel consumption you can accelerate more efficiently. I know all this is true by observing the sensors and such, but im not precisly sure of the phsyics involved.
Jeremy
Well, I think in an absolute since you are right, but it oversimplifies the physics to a childish level.
:)
For example, here are two scenarios:
1. I shift right below the peak torque output for my vehicle (2800 rpms). I shift this way consistently and I accelerate at a modest and average pace.
2. I shift at about 1500 RPMS where my vehicle barely has any real torque (relative to the rest of the RPM range).
Which scenario is more likely to be fuel efficient in city driving? Scenario 1. Why? Because in scenario 2 I am lugging my engine. I am so underutilizing my engine that my vehicle must lug the engine injecting a LOT more fuel just to keep it going and compensate for what I am doing. The gas usage goes way up. At 2800 RPMS my vehicle is accomplishing more work per unit of gas burned than it was at 1500 RPMS. So for NORMAL driving and average city driving it becomes more fuel efficient if I shift in the powerband of my vehicle.
Now, in interstate driving, once you have overcome the initial acceleration part the goal is to have your vehicle at a nice comfie RPM spot that will still give you decent power (enough to accelerate a little) without having to down shift while still maintaining really good fuel efficiency. So it makes since then, to stay at 1500 RPMS since we aren't really lugging the engine as we aren't pressing the accelerator down much at all. In my car staying around 1800-2000 RPMS in the interstate affords me optimal gas mileage with reasonable acceleration. To maintain optimal low end acceleration I use a 6spd manual transmission. The gearing has a really wide range which still gives good gearing onthe low and high ends. To much gear and you overrev on the interstate and waste fuel. Not enough gear and you underrev everywhere but the interstate and you lose a lot of acceleration potential. Finding the sweet spot in gearing and designing an engine with the appropriate torque and horsepower curves is a real trick.
Oh one other neat thing those new 300+ HP hemi's do while cruising on the interstate. They shut down some of the cylinders in the engine so they use only 4 cyldiners improving fuel economy by 20%. (they are still big cyldiners after all).
P.S. I have a 415 HP V8 with a turbo that gets 18city/23highway mpg
Jeremy
Not if its purely over the Internet. I think as long as it stays digital it should stay unregulated.
THat would be like regulating instant messaging, they do something similar to what phones do, right? Allow for communication with other people. I guess we should regulate video conferencing too, since it does what phones do and more. I think its silly to regulate purely digital communication mediums.
VoIP that hits a physical resource such as the standard phone system, yes, they can be regulated. Not all VoIP is moving into an existing market, but they are also creating a market that never really existed before.
Jeremy
Yeah, but what if all companies did this?
1% of 100 medium sized companies would be like an entire medium sized company doing nothing but charitable works. Every % counts.
Jeremy
But.. in the context of the stock market it's all legal like as long as everything is done appropriately. Heh. Cute no?
The stock market has its ups and downs, why not play it if you think you can predict the future?
Jeremy
Whatever, I bet the sales guy is financially secure from here out after a few years at the beast. Do your time there, walk out with some cash! :)
Integrity or not having to worry about money for the rest of your life? Especially when your a sales person anyway! If it were an influential Linux developer like Linux, or Alan cox, I would raise an eyebrow or two.
Jeremy
I suppose we should get into one of those circular arguments here.
:)
Here I'll start it, "but oh wait, I just said that was a Micro level issue. Talking about one person when we are viewing this whole thing from a more macro perspective".
"You reply, it doesn't matter! Think about the children"
We do this until someone mentions Nazis..
Just kidding..
Jeremy
I read some articles that debunked that. Outsourcing has allowed margins on wholesale goods and such to drop lower. If goods and services can be purchased for less it stimulates the economy and helps it grow. It actually makes sense at a macro level. I think oursourcing is a macro level economic issue and looking at it in a micro level can lead to misperceptions about where the money really is.
So people can still spend money buying toys for their kids because they are less expensive while still having a good quality level, so people can still spend.
Jeremy
Well, lets assume on a graphics heavy site we save 50KB just doing some simple image optimizations and squeezing the images here and there spending maybe 4 hours just optimizing the graphics. That is more like saving 50 GB per day, or 1200 dollars. I doubt you would pay 1200 dollars for the optimization.
However, in a more real world scenario sites raking in that many hits generally save money by going to a "real" hosting solution and just paying for a pipe size with unlimited data transfer. Generally having a 1-2Mb(it) pipe with unlimited transfer is just easier to deal with than a limited amount of bandwidth. Most sites with a lot of traffic have an unlimited transfer with burstable capability to handle spikes in traffic.
Jeremy
That isn't a very creative approach to dealing with problems in modern web applications.
Whether its one record of twenty I don't see how it really affects the data layer. If your application layer is written properly it won't matter about the data layer so much. Your application layer expects a recordset with certain data. Of course having n records instead of 1 you will have to change the way the application layer deals with it. That doesn't mean you still can't view the data from a SQL server in a very generic way (a recordset, or multiple recordsets). Your application layer should be smart enough to convert SQL data types into native types. And if your application is properly normalized and you aren't using any properitary column types it will be that much easier. You can take an OO view to your recordsets or a procedural view.. whatever fits best with the application layer's current design model.
Yes, abstraction has a cost. However in a practical sense there are platforms out there that already provide this level of abstraction and they support some incredibly busy sites. (Like ColdFusion powering the Macromedia site). So I think for all intents and purposes the abstraction is just fine for all but the most intense applications. In the real world having a flexible architecture is often more important than raw performance. I stand by my statements that abstracting SQL database access down to recordsets that have only native types to the language dealing with the data is the way to go with most applications. There are always exceptions. I am not really talking about those specific exceptions, just in general.
One final point of yours I would like to address is that you say it will take more time when you change the way the data needs to be accessed. It depends, if we are talking changing SQL platforms and optimizing for another SQL platform.. then no it is much easier to have a specific data access layer that you change. That is the ONLY place you should have to change code when changing SQL servers (save things like stored procedures and triggers). Your application layer should see the SQL server as a black box and have no inherent knowledge about the data access layer. If we are talking a complete paradigm shift over to something like an XML database or some other different data paradigm, yes you would be talking about changing your data access paradigms in the application layer as well.
I know its not quite so clear cut in a lot of cases, but simple SQL server abstraction is good. If you don't abstract your data access layer you are making your code that much more DIFFICULT to modify when you have to change database platforms.
The one big area of optimization that this kind of misses is stored procedures and triggers and such. You just can't abstract away something that is IN the database layer of your app. When you need the raw performance of a stored proc that is an optimization choice you must make. When you move logic away from your application layer and towards your data layer I think you need to have specific performance reasons to do so, because.. again.. most apps don't really need that much power. You should definitely be aware of the full range of such optimization techniques, because if you don't know about them you won't have knowledge of what they are and how they work in your application design phase.
I think we both agree on the pragmatic aspects of this, and that is such a heavily architected approach is not always required, especially for small projects that need to be done quick and fast. As your requirements grow larger so, typically, does your need for better and more flexible architecture. The bigger the app the more important it is to perform risk analysis. The more important it is to have a more defined and developed architecture. Most apps, especially open source apps that aim to have borad appeal, can benefit from an abstracted data access layer with very few data access platform specific optimizations. (And if you DO have such optimizations it makes sense to provide an un-optimized application layer only version of code in stored procs and such).
Jeremy
Sorry, I don't buy it. Even with the SQL hacks in any application of even moderate size you can hide your data layer from your application well enough that you would still only be rewriting the data access layer of your app. That is what abstraction is all about.
I know that you will invariable have to optimize for your current DB platform to take advantage of unique optimizations and features of that platform, but that still does not mean you can't do it in a way that keeps all of the optimization details hidden from the rest of the system. This is true in almost all cases where you are using a SQL server as your data storage layer.
Trivial apps will often not even bother with abstraction. It just seems silly to me given the handy abstraction tools via PEAR and Native libraries in PHP that anyone would not use them in a new project (versus the native calls). The only reasons (performance reasons) can still be architected away with just a little extra effort and some careful planning.
Jeremy
Jabber.. free server.. free clients.. all code available to ensure no gremlins if you are in a tight environment (Like financial companies that trade stocks for folks etc) and require high deals of confidence/confidientality.)
Jeremy
Uhm, so.. you mean to tell me that you expect a daily driven engine to last for 7+ years with the piston rings perfectly intact?
I drove a 10+ year old ford and my piston rings and pistons and heads were all in great shape. (Seals went bad from the car sitting to long).. but I don't buy this myth that Ford makes especially poor engine designs etc.
There aren't many 7 year old cars OUT there that don't leak a little patch of oil or burn a little off if they are on the original engine. The ones that DONT leak oil are the ones that had their filters serviced regularly, all the TSBs were followed for that car, and the oil was changed regularly. Most people NEVER do that. If they did you would find that the engines really last about the same length of time.
I read a study once that took a general health check of a group of cars over a 10 year period. It was comparing foreign cars (ALA: BMW/Mercedes) to domestics.. mostly fords. It found that the people with the more expensive cars were more likely to do the required maintenance. It also found that the domestic owners that followed the required maint schedules had cars in good shape, just like the foreign cars. Hummmmn. Now lets apply occams razor? Which seems more likely.. that there is some majorly defective thing about fords engineering that causes them to "smell like oil??" or that the users of said cars who paid much less for them simply don't care as much about the cars and thus dont do the maint.
Something to think about
For the record I have a 1987 2.3L 4cyl mustang with nary an oil leak and minimal oil consumption on completely original stuff. The trick? It was well taken care of with regular maint.
Jeremy
Nice shot with DNF! :-D
Jeremy
It's pretty easy for me. I spend my time writing code and am compensated appropriately for my time spent working on said code.
:)
An employer puts a value on how skilled you are and compensates you for your time in monetary means. The more your perceived value the more you make. I like my perceived value. I like eating and having nice things, thus I will work. My job just happens to be programming? So? It is still a job. Its not some mystery.. as an employer if you want something you use the ultimate motivator: money. You can't just go post something to a bulletin board and watch as the teeming hoard of OSS geeks magically code it for you.. if it doesn't exist you have to pay. If it exists and you want it modified you have to pay. Notice I said if *YOU* want someones time. And who would simply volounteer their time to someone for no pay? Most OSS folks do what interests them
Sure its all nice if some group of OSS geeks happened to have coded something that meets 70% of your requirements, but it just doesn't change the way the world works. If you want someone's time it costs money usually. The more skilled the person the more that time costs. Case closed. No one can live without money.
You can cook up all kinds of ideas for how to make money with OSS, but I just don't see it working for the masses, yet. Anyway.
Jeremy
What about a program used to create a video of programmers used to portray actors as programmers who are... and what if we had a beowulf cluster os said programers working in perfect harmony...
Ahh, yes but one programmers representation of an orderly set of bytes to accomplish a task can be quite different from another programmers set of bytes that accomplish the same task. What if the end result is not the art? What if the art is in the creation and representation of what creates the end result?
So, even though in the end you ARE bound by a structure of some sort I think that creative thought processes lead to better representations of said orderly structure (code). It most definitely is NOT a traditional form of art, but when I sit down and look a particularly well written function that is simple and "elegant" it inspires me in a great way. Almost the same way as when I see a great photograph that has several complex "artistic" elements involved (IE: Not just a snapshot, a serious well done picture meant to cause you to think).
Since I also enjoy photography and auto mechanics quite a bit I see parallels in ways most people would just discard. Photography is closer to "traditional" art. And I feel inspired the same way when I see a great photograph. Or when I see someone who is doing serious non factory work on their 800hp car and they come up with an "out of the box" solution. Or when I see someone lay some code out so well and so ingeniously that I think, "how would you do it any other way?". It is all inspiring at some level.
So what this boils down to is that art is in the eye of the beholder, truly. Just because it is not traditional art does not mean it isn't art to others. I look at some paintings considered masterpieces, even with the prerequisite knowledge on why this picture is great and have a hard time accepting the creative genius in certain aspects of a painting. The creative genius in a well done piece of code seems much more obvious to me, thus much more artful.
Its all a matter of perspective, as so much is in life.
Jeremy
It is nice that we are talking about a DICTIONARY word here. The legal definition of a word and the common use definition often vary quite wildly. There is also a legal definition of terrorism. Ticking off three or four things from ONE dictionaries interpretation of the word terrorism just doesn't even work.. save for maybe casual discussion like this, but I wouldn't get to upset, especially if you are trying to equate a dictionary definition to a legal definition.
Jeremy
One better... :)
:) )
I had also known Sendmail was a little tedious to learn. My main job is software development, but we are a small company so I multi task as system admin for about 10 systems, mostly Linux with a couple of internal windows systems.
None of the email systems are REAL painful to get working, even Sendmail. I can learn and understand these types of things easily with the years of experience I have. The thing is in a small company I have to squeeze every minute I can out of a day because my time means a lot. If I can spend half the time learning the ins and outs of a particular server app I am doing a good thing. So when I had to set up our first non shared systems in a data hosting facility I was very happy.
I decided QMail had enough of a popular following and a reasonable enough featureset and great security track record that we could live with QMail. So I installed QMail, read books, tweaked and eventually installed. Got everything working: SMTP+POP3 with selective relaying based on who had just popped. Worked great. Then it came time to set QMail up on a different server as we were shuffling services around to free up the server we currently used for mail. I was not looking forward to doing QMail again. I didn't like managing it the way we had it set up and didn't really feel I had the time to mess around with it, so I looked for alternatives.
I found Postfix. I downloaded it, compiled it, installed it and read through all of the configs and docs.Within an hour I was so amazed at the simplicity of Postfix and how much sense the configs made, I was in shock. Then I read a little more and searched around and found how easy IMAP was to get working as well. I then found an IMAP server that supported MySQL and found that Postfix also supported MySQL for domains/user configurations. That is all it took for me to be completely sold. I would have preferred more options for the RDBMS, but I wasn't going to complain much.. MySQL isn't to resource hungry. Performance wasn't a concern as I only needed mail for a pretty small group of people. As it is this setup is very READ heavy with the data being almost static (just needing a nice easy programmatic way to be updated when required) so it plays nicely into MySQLs performance forte.
In one day of configuration and code writing I had Postfix doing all of our SMTP, Courier IMAP (which includes POP3) all on the same backend. I had web mail setup and working. All of our mail storage was now centralized which only makes it easier given some of us travel around a lot and still need our mail.The best part, I wrote a web based management tool to manage domains, accounts, and aliases for our mail system. No more unix command line, log onto the application fill in a form or two and its done. Setup a new domain for email? Fill in a few forms and voila it's all done. I think it was one of the most pleasant experiences setting up a network service I administer I have ever had. (I realize that I am a programmer and not everyone will have the knowledge to hammer out a management tool in a small timeframe like that, but for us it simply works).
Thats my story for Postfix!
(P.S. we just have our system send a welcome email and the Maildir folders automaticaly get created after the first incoming message. Postfix handles this for you.
Jeremy
Minus the development and component costs and support personnel.. I wonder what the real profit is?
Jeremy