My wife teaches 4th grade and is a PhD. candidate in education...so I listen to lots of her remarks about computers in the classroom. They have 4 or 5 computers in the class room and they have a shared computer lab. All machines run Windows. The students have structured lessons on the computer where they learn about word processing and graphics, and they have internet time where they use the web for education and fun - a browser-only experience. It does not make much difference TO THEM what OS they run...kids love figuring things out and messing with new things. The seeds of M$ world domination are not being planted by this obvious PR stunt. Look at the bright side...thats one billion down, thirty-five billion to go...
The threat to Microsoft is that they think they are creating a product with strategic implications to business. Server operating systems are a commodity now - that's why Linux is so successful and why it can be developed so successfully using a distributed development model. Improvements in this area are now incremental. They are evolutionary not revolutionary. Incremental means building on existing proven components. In a commodity market, stability and low price are the objectives for manufacturers and the expectation of customers.
The funny part of this is that the memo indicates that business are opting for commodity (Intel) servers. What Microsoft does not seem to understand is that the reasons that businesses are migrating in this direction is the same reason they're migrating to Linux.
Lastly, I recall back in the early '90's seeing TCO industry projections for Mac's as being 1/2 that of a PC. You can see the result. They don't call it "sticker-shock" for nothing.
...in my opinion. Whatever benefit capitalism has for society...it is and "emergent" one. Certainly corporations, under capitalism, only have an obligation to benefit shareholders. Customers benefit only as long as there is competition for their business. When that competition is gone, there is nothing left but shareholder profit. Look at XP...that says it all.
Did I not read on this site that M$ paid no federal taxes last year? Maybe Bill could help out Uncle Sam a little and pay his fair share..or under a Republican administration is his fair share zero!!!!
Congrats to the poster in pointing out something really critical. In order to design well, and communicate that design well, we need to emphasize abstractions. In the old days, this would have been known as the "conceptual design phase" of a project. I've come in on lots of projects in the middle and invariably, when I ask for the system design documents., somebody hands me a generated dump (pun intended) of all the objects/entities - whatever in the system (often, the proud owner points to a massive paper mosaic pinned to their wall.) I always reply...no let me see the conceptual design documents. Usually there ain't none. And I mean that in every way...there aint no conceptual design documents and there aint no conceptual design.
You gonna date this twinkie-suckin bit-pusher? I don't think so. What this guy needs to realize is that this is as good as it gets. Send him a "Slashdot" pocket protector and lets move on. Face it kid, WE are your only friends.
This is why I USED to be a libertarian. Not anymore. This solution is mandated from the White House. This is George W. in your face. Same deal with airport security. Most of us want police officers doing the screening. Bush wants security guards. Why? Money. Unbridled lust for more AT ANY COST. And behind that, a deep belief they have it coming to them. They friggin DESERVE IT. Bullshit. And I only want to hear back from posters who got into Yale with a C average.
Well denomalization is done all the time in data warehouses to eliminate joins - especially in marketing-oriented systems where a single index alone can be 50 gig. I worked on systems like this for a number of large drug companies. And object databases that hold things like engine designs and aircraft electronic system designs are navigationally oriented. Queries that perform what-if engineering analysis on a database holding an engine design implemented in a 5th normal form is more than a nightmare - it's virtually impossible. I worked on systems like this for the Boeing company. I guess you just have to live in the real world and don't get your panties in a bunch when sombody challenges your beloved theories. Sometimes they hold up and sometimes they don't.
As for calling you...well you'll have to send me the number of the YMCA you're staying at first...
Yea, yea... I read all those books while completing my Masters degree in computer science. Normalization trades off performance because of the join processing that must be done. Nothing wrong with that trade but it's a fact. I suggest you expand your reading to include how database management system actually work to produce results.
Lets clear some things up. Highly normalized databases can NEVER be quicker than denormalized ones. When you normalize you are INTENTIONALLY slowing down retrieval so that you can, amoung other things, store a piece of data only once.
When you normalize, you force the database server to recombine an object from it's spread-out parts at runtime. This is why database companies have spent the last 20 years working on transaction optimization...so they could match the transaction performance of databases based upon a navigational data model.
Relational databases where originally conceived to help with decision-support processing - like data warehouses where queries are the majority of processing. Navigational data models (like object models) are optimal for on-line transactions because the records usually have pointers that take you directly to the data you want to access - rather than have the DBMS server perform join-processing.
Please, don't get me wrong, I got Normalization-fever just like the poster. You just don't do it to speed things up... Normalization generally has the opposite effect....except perhaps for updates on highly denormalized systems.
Gates told the Associated Press: "It's a value for consumers. Why are there headlights in cars? Why don't they make you go and buy those things separately?
What a staggering analogy! Lets continue. If I do not like my current headlights, I can go to the auto parts store and pick from a wide variety of different headlights and easily put them in myself. And by the way, one thing I would never do is go back to the dealership and have them install new headlights - cause as George Kostanza says,...you get the "scroogie"
Don't like the service? Don't sign up for it. Don't like the levels of service you got last month? Don't pay for them.
The assumption here is that the cost of changing is low. Microsoft's challenge will be to make the cost of switching too high. This will be easy from the perspective that they will (unlawfully I believe) leverage their OS to tie you into their services. From the other side, this will be difficult since most companies should know by now that if their service is sufficiently lucrative, Microsoft will copy and brand it - putting the originator and all other competitors out of business.
The poster also discusses relationships. Remember that it will not just be users that they have to pamper but they will also have to cut deals with other companies to establish a web presence that exploits Microsoft's proprietary (or I should say propriaterily implemented) features.
Right on... I do not mind entering my credit card information. I prefer to do that almost always. My address is a bit of a pain but not so much that I want to be taxed by Microsoft to avoid it. My household buys maybe one thing a week over the web. Finally, my login. Would it not have been a lot simpler to associate login management with my client-side browser software rather than have Microsoft keep it for me for a fee? Really, this has to be a simple programming problem. Like I come up on a web page that wants me to login (maybe the browser detects the input tag with the password attribute) and a window pops up with all of my user names and passwords. Then its cut and past or whatever. Really. These guys but the creep in feature-creep. Jeeze.
The key to being a great architect is knowing how to enter designs into as many different design tools as possible. Don't make the mistake of thinking you can bluff your way through a project by coming up with salient concepts and defining relationships. Top management knows that everybody can "think." But it takes something special to be able to translate "ideas" into something you can print-out and pass around.
I suggest you get certified on at least 5 to 10 new design tools every year. If you can get your company to pay for your training - even better.
Architect loves himself too much to give advice
on
Coder or Architect?
·
· Score: 1
There are many different kinds of organizations and many different types of projects where software gets developed - and the role of an architect is different for each one. On defense projects, the planning and development of software is very formal if for no other reason than the government requires it. On the other end of the spectrum, you could get called in on a business-oriented development project where it is a mad rush to complete some web functionality before an upcoming conference. In those circumstances there are no coders or architects...there are just developers (hopefully good ones) pulling from their past experience and development expertise to help these people make their deadline.
And there is everything else in between.
The BEST software developers are those who LOVE to write programs. They enjoy watching it come together, tracking down bugs, and especially watching someone use the final product to do something that they could never do before. They are there for the end-game on a project - when temper's flair and time is short. Also, they are lifetime learners...they study new programming languages and techniques long before the majority of their peers - and don't wait for their company to offically send them to training before cracking something open and finding out what makes it tick. The best developers are counted on to suggest a better way to implement something and never miss a chance to pick-up and improve their own approach when shown a better way.
The best architects that I've known are those who are dragged into the position after being top developers. A top developer will be acquiring the skills of an architect because this is how the concepts of design patterns, analysis patterns, data modelling, object modelling etc. are presented. Top software developers learn architectural principles because they realize at some point that allowing somone else to screw this up affects their work. Acquiring these skills allows them to review proposed "architecures" for flaws and "buildability."
My last remarks on this are to look in the paper and see how many jobs there are for software developers vs. architects. To be hired as an architect (as opposed to being "promoted" into the position internally) will require you to have significant education (MS in computer science - which you should have anyway) as well as significant software development experience and architecture experience.
And one last thing, on the interview, don't tell them that you got out of programming because you couldn't hack it anymore. Even if it's true, you should probably keep that to yourself.
I wish I could write like this guy. He expresses my feelings exactly. These security problems are the result of a flawed business strategy that tries to exploit an installed base of software without creating a layered architcture to work from. The motivation for not building a layered architecture is not a technical one it is a selfish one. What is truly astoundingly ironic is that they have critically harmed themselves by doing this. Bad karma Bill. You harm no one but yourself.
There's gold in them thar bugs...and somebody's gotta stomp 'em...
Re:A great example of open-source at work.
on
Five Years of KDE
·
· Score: 1
Microsoft software intentionally violates the well known software engineering principle of avoiding module coupling in system design. This is done so that you can not replace one layer of functionality with another. The only exception I can think of right now is ODBC which was required because Oracle had the advantage as client-server emerged and they needed to play nice to play at all. To say that comparing Microsoft's GUI work to KDE GUI work is unfair seems to punish KDE because they are working on a system developed with sound software engineering principles. As most software developers will tell you, if you build the foundation of your system correctly in the beginning, everything else you do down the road will be much easier and robust. I think the comparison is fair and intructional.
I would like to rise up in support of this opinion as well. Look at how Micropatch has tried to associate the various "open" movements with communist sympathizing hippies. The CLEAR motive here is to form an image in the "collective minds" (sorry!) of IT management that open software is not "corporate" software. I'm not suggesting that we stoop to the level of dishonesty, but after all, it is a Microsoft Virus.
You know I think M$ really missed the market here. Perhaps they should start giving their software away for free and sell the patches. They could change their name from Microsoft to Micropatch.
I was listening to NPR in the car about 1/2 hour ago. The scientist being interviewed said this winds up being a ban on future research. This is because many of the 60 lines are in Europe. He said we should not expect to get access to them. Of the 5 to 10 lines that are available here he said that Mr. Bush got it wrong when he said that these lines could be extended on and on. He said that after a finite number of extensions, they are worthless. Lastly, he said that data gathered on stem-cells from one ethnic group may have limited applicability to other ethnic groups. He said that after 9 months or so, our research will pretty much be shut down. Personally, I can not see how Bush, an anti-choicer, can be in favor this since he believes that life begins at conception. But then there's not much loot to be made by allowing women access to safe abortions. But there's tons of potential booty here. And I've never known a Republican to let their "ethics" get in the way of that.
Slashdot Mirror
The funny part of this is that the memo indicates that business are opting for commodity (Intel) servers. What Microsoft does not seem to understand is that the reasons that businesses are migrating in this direction is the same reason they're migrating to Linux.
Lastly, I recall back in the early '90's seeing TCO industry projections for Mac's as being 1/2 that of a PC. You can see the result. They don't call it "sticker-shock" for nothing.
Looks like he's eating well.....
...in my opinion. Whatever benefit capitalism has for society...it is and "emergent" one. Certainly corporations, under capitalism, only have an obligation to benefit shareholders. Customers benefit only as long as there is competition for their business. When that competition is gone, there is nothing left but shareholder profit. Look at XP...that says it all.
Did I not read on this site that M$ paid no federal taxes last year? Maybe Bill could help out Uncle Sam a little and pay his fair share..or under a Republican administration is his fair share zero!!!!
Congrats to the poster in pointing out something really critical. In order to design well, and communicate that design well, we need to emphasize abstractions. In the old days, this would have been known as the "conceptual design phase" of a project. I've come in on lots of projects in the middle and invariably, when I ask for the system design documents., somebody hands me a generated dump (pun intended) of all the objects/entities - whatever in the system (often, the proud owner points to a massive paper mosaic pinned to their wall.) I always reply...no let me see the conceptual design documents. Usually there ain't none. And I mean that in every way...there aint no conceptual design documents and there aint no conceptual design.
...or lots of beer.
You gonna date this twinkie-suckin bit-pusher? I don't think so. What this guy needs to realize is that this is as good as it gets. Send him a "Slashdot" pocket protector and lets move on. Face it kid, WE are your only friends.
As for calling you...well you'll have to send me the number of the YMCA you're staying at first...
Yea, yea... I read all those books while completing my Masters degree in computer science. Normalization trades off performance because of the join processing that must be done. Nothing wrong with that trade but it's a fact. I suggest you expand your reading to include how database management system actually work to produce results.
When you normalize, you force the database server to recombine an object from it's spread-out parts at runtime. This is why database companies have spent the last 20 years working on transaction optimization...so they could match the transaction performance of databases based upon a navigational data model.
Relational databases where originally conceived to help with decision-support processing - like data warehouses where queries are the majority of processing. Navigational data models (like object models) are optimal for on-line transactions because the records usually have pointers that take you directly to the data you want to access - rather than have the DBMS server perform join-processing.
Please, don't get me wrong, I got Normalization-fever just like the poster. You just don't do it to speed things up... Normalization generally has the opposite effect....except perhaps for updates on highly denormalized systems.
What a staggering analogy! Lets continue. If I do not like my current headlights, I can go to the auto parts store and pick from a wide variety of different headlights and easily put them in myself. And by the way, one thing I would never do is go back to the dealership and have them install new headlights - cause as George Kostanza says,
The assumption here is that the cost of changing is low. Microsoft's challenge will be to make the cost of switching too high. This will be easy from the perspective that they will (unlawfully I believe) leverage their OS to tie you into their services. From the other side, this will be difficult since most companies should know by now that if their service is sufficiently lucrative, Microsoft will copy and brand it - putting the originator and all other competitors out of business.
The poster also discusses relationships. Remember that it will not just be users that they have to pamper but they will also have to cut deals with other companies to establish a web presence that exploits Microsoft's proprietary (or I should say propriaterily implemented) features.
Right on... I do not mind entering my credit card information. I prefer to do that almost always. My address is a bit of a pain but not so much that I want to be taxed by Microsoft to avoid it. My household buys maybe one thing a week over the web. Finally, my login. Would it not have been a lot simpler to associate login management with my client-side browser software rather than have Microsoft keep it for me for a fee? Really, this has to be a simple programming problem. Like I come up on a web page that wants me to login (maybe the browser detects the input tag with the password attribute) and a window pops up with all of my user names and passwords. Then its cut and past or whatever. Really. These guys but the creep in feature-creep. Jeeze.
I suggest you get certified on at least 5 to 10 new design tools every year. If you can get your company to pay for your training - even better.
And there is everything else in between.
The BEST software developers are those who LOVE to write programs. They enjoy watching it come together, tracking down bugs, and especially watching someone use the final product to do something that they could never do before. They are there for the end-game on a project - when temper's flair and time is short. Also, they are lifetime learners...they study new programming languages and techniques long before the majority of their peers - and don't wait for their company to offically send them to training before cracking something open and finding out what makes it tick. The best developers are counted on to suggest a better way to implement something and never miss a chance to pick-up and improve their own approach when shown a better way.
The best architects that I've known are those who are dragged into the position after being top developers. A top developer will be acquiring the skills of an architect because this is how the concepts of design patterns, analysis patterns, data modelling, object modelling etc. are presented. Top software developers learn architectural principles because they realize at some point that allowing somone else to screw this up affects their work. Acquiring these skills allows them to review proposed "architecures" for flaws and "buildability."
My last remarks on this are to look in the paper and see how many jobs there are for software developers vs. architects. To be hired as an architect (as opposed to being "promoted" into the position internally) will require you to have significant education (MS in computer science - which you should have anyway) as well as significant software development experience and architecture experience.
And one last thing, on the interview, don't tell them that you got out of programming because you couldn't hack it anymore. Even if it's true, you should probably keep that to yourself.
Slashdot Retard? Now that's hitting below the belt!
I wish I could write like this guy. He expresses my feelings exactly.
These security problems are the result of a flawed business strategy that tries to exploit an installed base of software without creating a layered architcture to work from. The motivation for not building a layered architecture is not a technical one it is a selfish one. What is truly astoundingly ironic is that they have critically harmed themselves by doing this. Bad karma Bill. You harm no one but yourself.
There's gold in them thar bugs...and somebody's gotta stomp 'em...
Microsoft software intentionally violates the well known software engineering principle of avoiding module coupling in system design. This is done so that you can not replace one layer of functionality with another. The only exception I can think of right now is ODBC which was required because Oracle had the advantage as client-server emerged and they needed to play nice to play at all. To say that comparing Microsoft's GUI work to KDE GUI work is unfair seems to punish KDE because they are working on a system developed with sound software engineering principles. As most software developers will tell you, if you build the foundation of your system correctly in the beginning, everything else you do down the road will be much easier and robust. I think the comparison is fair and intructional.
I would like to rise up in support of this opinion as well. Look at how Micropatch has tried to associate the various "open" movements with communist sympathizing hippies. The CLEAR motive here is to form an image in the "collective minds" (sorry!) of IT management that open software is not "corporate" software. I'm not suggesting that we stoop to the level of dishonesty, but after all, it is a Microsoft Virus.
You know I think M$ really missed the market here. Perhaps they should start giving their software away for free and sell the patches. They could change their name from Microsoft to Micropatch.
Yea, I hear that SP1 id due out next Friday...
I was listening to NPR in the car about 1/2 hour ago. The scientist being interviewed said this winds up being a ban on future research. This is because many of the 60 lines are in Europe. He said we should not expect to get access to them. Of the 5 to 10 lines that are available here he said that Mr. Bush got it wrong when he said that these lines could be extended on and on. He said that after a finite number of extensions, they are worthless. Lastly, he said that data gathered on stem-cells from one ethnic group may have limited applicability to other ethnic groups. He said that after 9 months or so, our research will pretty much be shut down. Personally, I can not see how Bush, an anti-choicer, can be in favor this since he believes that life begins at conception. But then there's not much loot to be made by allowing women access to safe abortions. But there's tons of potential booty here. And I've never known a Republican to let their "ethics" get in the way of that.