Then again, software "coders" are almost never the ones that are responsible for the success or failure of the project -either-. Architects, analysts, team leaders, researchers, etc, are. Actually "building" the thing is completly trivial once these people succeed on their side of things.
Almost like building a house. Once the plans are drawn, the ressources allocated, etc, if the project is managed correctly, its almost impossible for the coders to screw something up.
Or use outdated or forked versions of the last version of the software under GPL2, which in these cases, as of today, is probably 99.9% alike to the GPL3 version, hmm? Of course, that won't last though...
I agree, though admitedly, often, even when in a team made entirely of people who breath by cross-browser sites, certain requirements and considerations end up with the team making IE-only web apps (for internal use though! IE-only internet facing web apps are simply not economically viable anymore except in extremely rare scenarios).
I'm actually in one of these situations right now, where while it only takes a few extra minutes (when the requirements are simple anyway) to make a web app cross-browser, its a few extra minutes we don't have, nor need. Rare, awkward, and it makes me winces, but it was the right choice to make in this case. So its not always from incompetence:)
Yup, when the DS came out, Nintendo was very specific that they did not intend to replace the Gameboy with it, and that they would be in parallel, with the DS being designed for more "original" games, and possibly target an older audience, while keeping the gameboy the traditional handheld.
However, quickly (especially after the DS light, but even at launch or almost), normal games that would normally have found themselves on gameboy ended up on the DS, and the DS ate up all the gameboy's market, on top of its own incredible one. Thus Nintendo changed their mind about keeping the gameboy around (they talked about it quite a while ago actually). In other words, the first company to "defeat" the GB, has been Nintendo. Kind of amusing.
You -have- to implement incredibly advanced concepts. Those concepts are just not CS. Software architecture, development methods, design patterns, the development decision trees (just to name the simpler ones that most should have heard of), and more, could easily fill an undergraduate curriculum.
The thing is, most programmers, especially those that come from the CS worlds, don't even KNOW they need this, because all they do is follow the orders of someone with 10+ years of experience who studied all this on his/her own to get where they are, when it could easily have all been taught in school, and that all the developers (minus the ones implementing the math-y, more scientific stuff) should know, but don't.
The way I see it, is like a doctor vs a biologist. A biologist will learn some (a lot) of human biology, a doctor obviously learns a ton of very abstract, scientific stuff, but at the end, they are 2 different paths. The same should be true in the computer fields. There should be a software development degree, with quite a bit of math and theory, but not the entire focus on it, and CS as is, and depending in what interests you, you pick the appropriate one.
Because for every algorythm that a non-science developer can't figure out, there's a CS major that got the algorythm right, and implemented it poorly. The best example I can think of, of this, is the Postgres code base. You can tell by the code that the people who wrote it had all the science background they needed for it, but had no clue whatsofreagin ever about -basic- software development and architecture concepts.
Software development, and computer science, are 2 fields with enough stuff to fill 3 Ph.Ds over. But they are 2 very different things, and most schools only offer one of the two, which is a big, big issue.
In the same way someone with an MIS or IT degree are limited in the software development field, so are CS. None of these majors really teach you what you need to -develop software-.
CS majors are a bit ahead because they can write complex algorythms and do a lot of low level things, but thats just a fraction of what software development is. Very, very few colleges will offer the appropriate software development and engineering classes required to do real software development and architecture, but those, along with the more passionate people with the appropriate experience, have the skills. Everyone else is just part of the peanut gallery.
If you end up working for Intel, Nvidia, or game companies (stuff that uses CS quite heavily), you'll have the upper hand simply because of all the CS (not software development) stuff involved, but as soon as you end up in the business world, CS skills and MIS/IT are more or less useless: your elective classes and experience will be all whats left thats useful, aside for the odd thing in an entire project that calls for these skills.
If the new version is digital, then yes. If the new version is bound to some hardware, let say they decided to remake Zelda 64 on 10 gb SD cards, well, no:)
Could point, but then if thats true, whats the issue with Disney? They're scared that people will start releasing Mickey Mouse stuff, if i understand correctly...if its as simple as you say, then it shouldn't be a problem, but it is... It is very very possible I'm not understanding things right, however.
I don't know, but I feel that as long as you use/enforce it semi-regularly, you should be able to keep it, but copyrighted material shouldn't be allowed to die in obscurity.
To make a parallel of the gaming world (which is relevent, copyright is copyright), it would be freakishly weird to see Sony and Microsoft be allowed to make Mario and Zelda games already, but at the same time, obscure games from 25 years ago that no one hears of anymore should be freely available as ROMs, its too late to just come back and say "oups, I want to sell that game again on the Wii's VC!".
So as long as you use it, its yours, and if you don't for X (short) amount of times, its everyone's. I'm sure there are flaws with this, but it would be a start, and make everyone happy, to some extent.
Yup, i'm exactly in that situation right now. Some hot shot wanna be developer, who thinks he knows everything and has a bit too much influence, is making changes in the project scope constantly. So now we went from a scope that meant "Holy crap, this is gonna be some rough deadlines" to "err...I'm going to retire before this project is halfway through..."
And there is a difference between coding before the specs are set in stone, and building over a prototype. Its 2 totally different, 100% unrelated things.
I had to quit a job because of that. My boss kept asking me to make prototypes to prove concepts, which I did...then ask me to build entire ERP and CRM systems around the prototypes, which obviously was impossible within the constraints given, and definately impossible to keep maintainable, and mostly secure (they were web based systems exposed to the net, and the prototypes were not security aware...). I quit rather than take the responsability behind that...
I'm a developer, sysadmin, software architect, system architect, business and functional analyst, in both unix and windows environments, and senior in a few high demand development environment: no one gives a flying duck.
If its a software architect job, they don't give a flying duck about the system architecture. If its a developer job, they don't care about my Java experience if its a C# job. Even worse? They don't care about my C# experience if its a VB.NET job, and vice versa (wtf, there's like 6 keywords of difference and a slightly altered event model).
So what I do? 1 resume per job I apply to. That works. Well. Just have to creatively word things out: "X years experience in.NET environments including Y years in C#". That catches the fish, while stating things as they are do not.
Dell's promotions and stuff is err..."dynamic", to say the least. You can find the same system at like 5 different price depending where you look. So its no surprise that stuff like that would happen.
They're not losing their customers. They're only losing subscribers to their software assurance program, which probably wasn't much to begin with, since people complained strongly about it from day 1. The only worthwhile software subscription MS has is MSDN.
Software Assurance sucks, cuz it puts MS in a serious conflict of interest. So all it means is that they'll go back to a more reasonable model. Which I can't wait for.
I don't know about these, but the more advanced ORMs out there will take multiple operations in a batch, parse them, optimize them, then make a single SQL query (or at least, as few as possible), and send that in one shot. Or at least, have the option to do so, so its rarely a problem.
The one thing that tends to cause issues, are very specific things like query planner hints, but some ORMs even support those. But for the 3 queries in a 5000 tables with replication system you have that require such things, again, most good ORMs will also support mapping stored procedures to objects, so you can have the best of both worlds, no issues:)
Save for support, software development is pretty close to fix cost... that you sell it to 4 billion people, or you sell it to 5000, its the same... So technically speaking, if Microsoft had the choice between selling Windows to every single human being on earth at 3$, or to sell it to a fraction of the western world at 400$, they'd probably pick the former =)
terrabytes of data here. Works pretty good. Where's the problem exactly? In most scenarios, it will be doing the exact same query anyway (not a billion ways of doing CRUD for example, and most data mining operations should be done on an OLAP system anyway, so....)
Half a day of work is usually 3.5 to 4 hours for most places, so that comes pretty close to your "couple of hours". And of course you can tune MSSQL, it has quite a few options to tweak and optimise...it just tends to configure them correctly on its own.
4GL tools tend to have this issue, but look at more recent ones...which while they DO have issues, solve most of the problems you're talking of (going to use MS examples, sorry!):
Error handling: SSIS and WWF handle this quite nicely, with error paths and exception blocks, respectively.
Duplication: Reusable blocks are in most of the recent tools.
Knowing what it does: If its made in Java,.NET, etc, you can probably decompile it and get the code, often with the comments! (its what I did once having to tweak something very specific in SSIS
Limited functionalities: Most tools have extensibility features now, that allow for even native code.
Source control: no problem if it uses XML.
Debugging: thats usually fine in the recent tools, if a little bit different. Good ones allow for breakpoints, variable inspection, etc.
Maybe in the field you're in, the tools really suck, but they're not -ALL- bad, thats for sure! What i like is how you can use the program itself as its own documentation, which saves a lot of time, and a lot of issues syncing documentation with last minute changes...
Correct. And thats where the confusion comes: There is an incredibly high demand for software -developers-, as opposed to computer scientist, people who are intimately familiar with the software development flow, design patterns, software architecture, and a bunch of other things most CS curriculum don't teach. Thats where articles like this one comes from.
What it fails at, is that there's also a fairly high demand for actual computer scientist. So what does it mean? That we leave CS just like it is, and that there should be -2- different programs in university, because sticking a CS major in a pure software development job is dumb. Just as dumb as sticking a software developer or architect to develop some new compression algorythm or program a pace maker.
Of course, for every programmer who doesn't know about O(n), there's someone straight out of CS school who has no clue about theta, omega, or just actually what Big O is good for, which is quite amusing:)
Slashdot Mirror
Then again, software "coders" are almost never the ones that are responsible for the success or failure of the project -either-. Architects, analysts, team leaders, researchers, etc, are. Actually "building" the thing is completly trivial once these people succeed on their side of things.
Almost like building a house. Once the plans are drawn, the ressources allocated, etc, if the project is managed correctly, its almost impossible for the coders to screw something up.
An thus restrict it to the 0.1% "elite" part of the population!
Or use outdated or forked versions of the last version of the software under GPL2, which in these cases, as of today, is probably 99.9% alike to the GPL3 version, hmm? Of course, that won't last though...
I agree, though admitedly, often, even when in a team made entirely of people who breath by cross-browser sites, certain requirements and considerations end up with the team making IE-only web apps (for internal use though! IE-only internet facing web apps are simply not economically viable anymore except in extremely rare scenarios).
:)
I'm actually in one of these situations right now, where while it only takes a few extra minutes (when the requirements are simple anyway) to make a web app cross-browser, its a few extra minutes we don't have, nor need. Rare, awkward, and it makes me winces, but it was the right choice to make in this case. So its not always from incompetence
Yup, when the DS came out, Nintendo was very specific that they did not intend to replace the Gameboy with it, and that they would be in parallel, with the DS being designed for more "original" games, and possibly target an older audience, while keeping the gameboy the traditional handheld.
However, quickly (especially after the DS light, but even at launch or almost), normal games that would normally have found themselves on gameboy ended up on the DS, and the DS ate up all the gameboy's market, on top of its own incredible one. Thus Nintendo changed their mind about keeping the gameboy around (they talked about it quite a while ago actually). In other words, the first company to "defeat" the GB, has been Nintendo. Kind of amusing.
The only IE-only sites I've seen in years were internal corporate intranet applications, and even those are starting to be rare...
You misunderstood my point.
You -have- to implement incredibly advanced concepts. Those concepts are just not CS. Software architecture, development methods, design patterns, the development decision trees (just to name the simpler ones that most should have heard of), and more, could easily fill an undergraduate curriculum.
The thing is, most programmers, especially those that come from the CS worlds, don't even KNOW they need this, because all they do is follow the orders of someone with 10+ years of experience who studied all this on his/her own to get where they are, when it could easily have all been taught in school, and that all the developers (minus the ones implementing the math-y, more scientific stuff) should know, but don't.
The way I see it, is like a doctor vs a biologist. A biologist will learn some (a lot) of human biology, a doctor obviously learns a ton of very abstract, scientific stuff, but at the end, they are 2 different paths. The same should be true in the computer fields. There should be a software development degree, with quite a bit of math and theory, but not the entire focus on it, and CS as is, and depending in what interests you, you pick the appropriate one.
Because for every algorythm that a non-science developer can't figure out, there's a CS major that got the algorythm right, and implemented it poorly. The best example I can think of, of this, is the Postgres code base. You can tell by the code that the people who wrote it had all the science background they needed for it, but had no clue whatsofreagin ever about -basic- software development and architecture concepts.
Software development, and computer science, are 2 fields with enough stuff to fill 3 Ph.Ds over. But they are 2 very different things, and most schools only offer one of the two, which is a big, big issue.
In the same way someone with an MIS or IT degree are limited in the software development field, so are CS. None of these majors really teach you what you need to -develop software-.
CS majors are a bit ahead because they can write complex algorythms and do a lot of low level things, but thats just a fraction of what software development is. Very, very few colleges will offer the appropriate software development and engineering classes required to do real software development and architecture, but those, along with the more passionate people with the appropriate experience, have the skills. Everyone else is just part of the peanut gallery.
If you end up working for Intel, Nvidia, or game companies (stuff that uses CS quite heavily), you'll have the upper hand simply because of all the CS (not software development) stuff involved, but as soon as you end up in the business world, CS skills and MIS/IT are more or less useless: your elective classes and experience will be all whats left thats useful, aside for the odd thing in an entire project that calls for these skills.
If the new version is digital, then yes. If the new version is bound to some hardware, let say they decided to remake Zelda 64 on 10 gb SD cards, well, no :)
Could point, but then if thats true, whats the issue with Disney? They're scared that people will start releasing Mickey Mouse stuff, if i understand correctly...if its as simple as you say, then it shouldn't be a problem, but it is... It is very very possible I'm not understanding things right, however.
I don't know, but I feel that as long as you use/enforce it semi-regularly, you should be able to keep it, but copyrighted material shouldn't be allowed to die in obscurity.
To make a parallel of the gaming world (which is relevent, copyright is copyright), it would be freakishly weird to see Sony and Microsoft be allowed to make Mario and Zelda games already, but at the same time, obscure games from 25 years ago that no one hears of anymore should be freely available as ROMs, its too late to just come back and say "oups, I want to sell that game again on the Wii's VC!".
So as long as you use it, its yours, and if you don't for X (short) amount of times, its everyone's. I'm sure there are flaws with this, but it would be a start, and make everyone happy, to some extent.
Yup, i'm exactly in that situation right now. Some hot shot wanna be developer, who thinks he knows everything and has a bit too much influence, is making changes in the project scope constantly. So now we went from a scope that meant "Holy crap, this is gonna be some rough deadlines" to "err...I'm going to retire before this project is halfway through..."
I am not american :)
And there is a difference between coding before the specs are set in stone, and building over a prototype. Its 2 totally different, 100% unrelated things.
I had to quit a job because of that. My boss kept asking me to make prototypes to prove concepts, which I did...then ask me to build entire ERP and CRM systems around the prototypes, which obviously was impossible within the constraints given, and definately impossible to keep maintainable, and mostly secure (they were web based systems exposed to the net, and the prototypes were not security aware...). I quit rather than take the responsability behind that...
Hey thats cool, if not even PCI Express is that fast, it means Microsoft could realise one of their wet dreams, and sell Direct X "as a service"!!!!.
I'm a developer, sysadmin, software architect, system architect, business and functional analyst, in both unix and windows environments, and senior in a few high demand development environment: no one gives a flying duck. If its a software architect job, they don't give a flying duck about the system architecture. If its a developer job, they don't care about my Java experience if its a C# job. Even worse? They don't care about my C# experience if its a VB.NET job, and vice versa (wtf, there's like 6 keywords of difference and a slightly altered event model). So what I do? 1 resume per job I apply to. That works. Well. Just have to creatively word things out: "X years experience in .NET environments including Y years in C#". That catches the fish, while stating things as they are do not.
Dell's promotions and stuff is err..."dynamic", to say the least. You can find the same system at like 5 different price depending where you look. So its no surprise that stuff like that would happen.
They're not losing their customers. They're only losing subscribers to their software assurance program, which probably wasn't much to begin with, since people complained strongly about it from day 1. The only worthwhile software subscription MS has is MSDN.
Software Assurance sucks, cuz it puts MS in a serious conflict of interest. So all it means is that they'll go back to a more reasonable model. Which I can't wait for.
I don't know about these, but the more advanced ORMs out there will take multiple operations in a batch, parse them, optimize them, then make a single SQL query (or at least, as few as possible), and send that in one shot. Or at least, have the option to do so, so its rarely a problem. The one thing that tends to cause issues, are very specific things like query planner hints, but some ORMs even support those. But for the 3 queries in a 5000 tables with replication system you have that require such things, again, most good ORMs will also support mapping stored procedures to objects, so you can have the best of both worlds, no issues :)
Save for support, software development is pretty close to fix cost... that you sell it to 4 billion people, or you sell it to 5000, its the same... So technically speaking, if Microsoft had the choice between selling Windows to every single human being on earth at 3$, or to sell it to a fraction of the western world at 400$, they'd probably pick the former =)
terrabytes of data here. Works pretty good. Where's the problem exactly? In most scenarios, it will be doing the exact same query anyway (not a billion ways of doing CRUD for example, and most data mining operations should be done on an OLAP system anyway, so....)
Half a day of work is usually 3.5 to 4 hours for most places, so that comes pretty close to your "couple of hours". And of course you can tune MSSQL, it has quite a few options to tweak and optimise ...it just tends to configure them correctly on its own.
The oracle thing made me laugh out loud =P
4GL tools tend to have this issue, but look at more recent ones...which while they DO have issues, solve most of the problems you're talking of (going to use MS examples, sorry!): Error handling: SSIS and WWF handle this quite nicely, with error paths and exception blocks, respectively. Duplication: Reusable blocks are in most of the recent tools. Knowing what it does: If its made in Java, .NET, etc, you can probably decompile it and get the code, often with the comments! (its what I did once having to tweak something very specific in SSIS
Limited functionalities: Most tools have extensibility features now, that allow for even native code.
Source control: no problem if it uses XML.
Debugging: thats usually fine in the recent tools, if a little bit different. Good ones allow for breakpoints, variable inspection, etc.
Maybe in the field you're in, the tools really suck, but they're not -ALL- bad, thats for sure! What i like is how you can use the program itself as its own documentation, which saves a lot of time, and a lot of issues syncing documentation with last minute changes...
Correct. And thats where the confusion comes: There is an incredibly high demand for software -developers-, as opposed to computer scientist, people who are intimately familiar with the software development flow, design patterns, software architecture, and a bunch of other things most CS curriculum don't teach. Thats where articles like this one comes from.
:)
What it fails at, is that there's also a fairly high demand for actual computer scientist. So what does it mean? That we leave CS just like it is, and that there should be -2- different programs in university, because sticking a CS major in a pure software development job is dumb. Just as dumb as sticking a software developer or architect to develop some new compression algorythm or program a pace maker.
Of course, for every programmer who doesn't know about O(n), there's someone straight out of CS school who has no clue about theta, omega, or just actually what Big O is good for, which is quite amusing