I seem to remember highly trumpeted press releases about 3 years ago when M$ and SGI joinly announced a coming together of Direct3D and OpenGL into a new standard. This wasn't to happen immediately but was being worked on. If I remember correctly (and my mind is hazy on the subject) the crux of the deal was some standard for how the resultant code hit the graphics card.
The idea was that those graphic card manufactureres who were thinging that they might develop OpenGL compatible cards were led in a particular direction which in fact has left DirectX pretty much in the driving seat.
I've not heard anything since. So perhaps this approach worked to stifle any competition for a while (although NVIDIA still seem to offer OpenGL stanards for their cards).
With this idea of sealed boxes the scariest thing for me is that maybe they'll sell one box for video editing and one for gaming but not one that does both well. So instead of paying for one card I have to buy a whole system to get the same functionality. The beauty of PCs is how customizable they are, it would be a great loss to loose that.
I don't know why its scary - the point about this "sealed" box is that it is cheap. Think more along the lines that you have a several ball point pens (maybe each with a different colour inks) than a quill pen and lots of bottles of ink.
I like the general tone of this post, but some of the facts are glaringly wrong...
Object orientation is necessary and natural.
No, it's not (necessary). I like OO, but it is not the silver bullet that kills all programming problems. Other, radically different styles (functional programming springs to mind) have shown big increases in effectiveness over OO languages, by any useful metric you care to choose, at least in certain areas. OO is good, but hardly perfect and clearly not necessary based on the number of non-OO projects in the world.
I too like the tone of the top level post, and I think the quoted bit above starts to make a valid point.However I think it misses one fundemental...
Programming in an object oriented way is fundementally different from programming in an functional way. In the first "processes" are smeared across a number of actors (objects) which cluster the parts in a number of separate functions around the data they are working on. In the second "processes" are implemented (as functions) in one place.
Just using an object oriented language does not guarentee that you use an object oriented approach. In fact a programmer has to almost make a complete philosophy shift to move to an OO way having gained experience of functional programming. For programmers who try and do this as a day job (ie work as programmers for a living) this shift often takes about 6 months.
This means, I think that the effectiveness that you refer to above is as a result of this bias towards a functional style and the difficulty many OO projects have in getting it right.
However an OO project that does get it right makes massive and rapid progress once the basics are in place.
In my view - we are seeing KDE make rapid advances at the moment precisely because (at least with KDE 2 - I have no experience of KDE 1)they have got some fundementals right. The challenge they will face as they attract more supporting developers will be that the layers above which probably won't have as rigourous architectural control over them will suffer from people with a different approach start to work on the applications.
On the other hand, I don't think a new and better C++ will solve the problem and be the saviour that Anonymous Brave Guy thinks.
I think that small companies are where Linux is going to have to make inroads into the market.
Remember the days when Xerox copies where the only ones ever used. I even think that Xerox may have been forcing a charge per copy on large organisations (certainly renting rather than owning a photocopier was the norm)
Then the Japanese started to attack small businesses with cheap simple photocopies which you bought. Almost wiped Xerox out.
Just out of curiosity, why? Hasn't Emacs always been open source? Or were you paying to get a compiler license to build Emacs?
I don't know - remember in those days we (as a company) were not connected to the internet and I had certainly not heard of open source. As a commercial company we were not in the university stream where software was shared - the only way to get software was to buy it from someone. I think we bought it along with our Unix System V licence from a UK company called The Instruction Set
It certainly wasn't a compiler licence but the actual editor. A collegue and I have been talking about it and we think we used an editor called "Author" before that... but is all a bit hazy.
Why not use the money (and the willing effort of thousands?) to create a media blitz on how yet another small voluntary group has been squashed by big business.
Do not spare putting your point of view that you had prior art into this argument, regardless of whether or not it will stand up in court. Just believe you are in the right.
Do say you cannot afford lawyers but try and get some media company that wants to give these guys a bloody nose the opportunity to stump up the fee.
Wait for about 10 years. Compare the state of free software to the state of "shared source" or whatever they call it this week.
A lot can happen in "about" 10 years. It was a bit longer than that, but in 1985 IBM (who now embrace Linux) not Microsoft was the dominant party and I had to pay to buy an Emacs licence to run on my Unix operating system (having just switched over to System V from my Microsoft Xenix operating system).
Don't get me wrong, I like GPL. I don't see a problem with it "infecting" source, other than that it will turn some appalingly rich people into merely very rich people. But I do think that it is fair to describe it as viral, or Borg-like, or whatever. The mechanism is the same.
The problem with that approach is that M$'s PR department makes mincemeat out of it. So whilst you can be logically correct in your argument you still lose out because you get the media misinterpreting the words and the man in the street drawing the wrong conclusions.
I think we should start a propaganda war back - which in the obtuse way of a M$ PR department you do slightly off the side and start to talk about their licencing conditions as deliberately restricting the quality of software and the GPL as an innoculation against that cancer.
And I'll add - I used to be in brilliant coder - but in 1975:-). {Actually I was never very fast - just accurate, each person added a pint of beer into our project drinking pot for every bug found. I was great at finding bugs though:-)}.
I then spent 20 years acting in this curious interface between technology, business needs and management.
But over the last 6 years I have been in Marketing at work, and I am missing the technical input.
I am trying to build up to doing some open source coding - but one disadvantage is that the last time I coded seriously it was in Assembler (PDP 11 - correcting Dave Cutler's bugs in RSX-11M:-) ).
But I would welcome the opportunity to find other ways to help.
So you are not alone in this - and I am sure I am also not the only one who would join you in those thoughts.
If copy protection devices are going to come, the open source movement MUST fight for the mechanisms to allow someone to copy UNLOCK their data (by that I mean that once such an UNLOCK is applied to data a subsequent copy of that data connot be copy locked).
But there are interesting litte differences that make it a whole lot easier to do what you want.
For instance, packets due to be forwarded also hit the input chain. Now they don't. As a result, its a lot easier to write firewall rules that are different for forwarded packets to ones reaching the firewall box.
This is a typical response my a market oriented company to something that worries them. Attack it at what appears to be its weak points.
For me, however, he makes points which we should attack back, and which are his weak points.
"Open source is an intellectual-property destroyer," Allchin said. "I can't imagine something that could be worse than this for the software business and the intellectual-property business."
Surely the response to this, and one that the Open Source Movement should be shouting from the rooftops is that this is not in customers best interests. Microsoft has already been shown to stiffle things in the interest of the software business at the expense of customers. Open source actually is a far better way to get code that is stable and bug free because it emulates the scientific method and exposes it to peer review.
I feel that this argument is not used nearly enough. There are often complaints out there that large corporations are worried to switch to Linux because of the belief that it is not backed by a large corporation that guarentees them support. We should make the point that actually getting support for problems is a lot faster because one only has to raise the issue at the right place and almost immediately there are people looking at the problem and coming up with fixes.
I think there is a difference here. Microsoft calls their product "Microsoft Word" (two words) not just "Word". If someone had a product called "Open MSWord" then that would be different. We already have "Microsoft Office", "KOffice" and "Star Office".
The other problem, is that he is getting support e-mails from users of OpenSSH - demonstrating clearly that there is a confusion in users minds
But surely it Ximian who have put out this press release NOT HP. Its Ximian who have much to gain from this announcement, because it announces to the world that they are company that large companies do business with.
It seems to me that the article fails to understand two key points.
Firstly, although Linus heads up to programme to move the kernel forward, there is a vast array of people out there on the net supporting the activity. You always have to have somebody heading that programme, and whoever does do it is going to get up some peoples noses sometimes. I don't think a commercial organisation will find it any easier to control the kernel, or ensure it moves along any faster - indeed they would come under enormous political pressures from all sides during the transition phase from the way it is now. Typically the way that an organisation reacts so such pressures is to form committees. Hardly conducive to moving the kernel forward faster. In other words, to change from the way things happen now is a massive risk which should used to argue against those who see the grass as greener on the other side of an imaginary fence
Secondly, the reality is that Linux, to be useable is also a vast array of supporting applications. These two have to continue to move forward (perhaps even more so) if Linux is to see off the competition.
Whilst all this looks wonderful, it already assumes you see an image. What is actually transmitted is (presumably) a bit stream. How do you decode this into being an image?
Do you think that it is because your mind is better tuned or something, or that it is simply a matter of matching the paradigm to the person?
I would not characterise my approach as better just different. What is clear, is that when looking at a system most programmers I encounter start to think about the functions that it should carry out. An initial breakdown then starts to split the system into functional silo's. This is of course wrong for an OO approach.
I, on the other hand, had always (and I started working on these things in the late '60s before OO had been characterised) had always seen things in terms on intercommunicating actors, with functionality shared between these according to an "architecture". I can vivedly remember having arguments with people about how we should structure systems - but my problem was that whilst I always had an instinctive feel about the way to do things that was always different from the rest of my collegues, I couldn't persuade them that I was right. When OO came along, and I realised that the basic concept of an entitiy, with functionality clustered around it was the paradigm I had some formality to hand what had been instinct.
That was the crux of my point. OO might be the theoretical correct way to do things, but actually there are two ways to thing about the architecture of a solution (functional decomposition or entity decomposition) and the thought processes right at the start need to follow the second approach.
I think that an approach to OO can be taught - the anecdotal evidence that is discussed in my organisation is that it takes the average person about 6 months to make the mind switch. This is the real reason why short term costs may be higher - but there is an expected longer term benefit (I say expected because I have no evidence only my own prejudiced view that this is so).
Object Orientedness is as much a mindset and philosophy as it is language constructs. (OO languages are designed to facilitate the philosophy, not guarantee it.) Many studies say in practical terms that most OOP projects fail to produce benefits over non-OOP methods because for political, technical, or training reasons, OOP is "not being used properly
My failure to realise this in the mid '90s was why I fell off the greasy pole. I was a technically oriented manager who pushed one of my project teams to build the next product in an object oriented way. I failed to understand that whilst an OO mindset was obvious to me, to the programming team they couldn't get a handle on where to start. Net result is a project that went nowhere for 6 months and didn't get moving until I abandoned my insistence on an OO approach. This killed me with the costs of the project overspend which brought the rest of my profit centre to its knees. I am no longer a manager with these sort of responsibilities.
Even today, I still feel architectually the OO is right and works the way I happen to think - but I have to be honest and realise most programmers don't seem to have the same mindset and that kills the benefits that can be had.
If the purpose of this technology is to be developed to protect copyright, does the new standard allow the ability to protect CopyLeft. That is, can it be used so that some data can be given a lock which prevents this data from being copy protected in the future by some other (unknown) person, and if not why not
So once the game companies starts to see that even ordinary gamers/users installs linux with ease and has no problems in learning it then they will finally develop games for the platform and sell them through retail. IMHO linux still has a very long way to go before that happens.
I think that most of you have it wrong when you think that end users install windows. Most end users buy a computer with it already installed and configured. What they do is buy extras (hardware and software) that comes with self install (you install the CD into the drive - it automatically comes up and ask you to click on the install button - you select all the defaults and you are up and running).
When people buy hardware (PC or gaming) with Linux pre-installed and only then will software developers start developing games (in any serious way) for it.
I don't think that has to be as far off as the previous post implied, but I think it is more likely because a new hardware manufacturer enters the market with some down market home computing hardware that is not a PC.
It is true that sometimes adhering to standards is free, performance-wise; that going along with the "flow" and adopting other's formats and methods into your own only helps the performance and usability of your platform. However, in gaming I think that a system that adopts standards too strictly will be too restrictive for the gaming environment.
I tend to disagree - provided things are architected reasonably well then these sorts of standards do add a degree of overhead - I guess we are talking 10% sort of numbers. With CPU speed doubling every 18 months (according to Moore's law) this sort of overhead is fluff.
Whats more, don't forget that if there is a standard that everyone is using in the open source area, then there is potentially a large number of minds that can look at these inefficiencies and optimise most of them away.
Don't forget that windows imposes standards (DirectX) on gamers too, but we only have M$ to optimise this.
Following a thread of a few days ago about the new ATA spec which allows people with copywriteable content to prevent it being copied, shouldn't the open source community lobby for the manufacturers of these disks to add to the spec a facility to allow data to be GPL'd (ie locked so that someone else can't subsequently lock it).
Slashdot Mirror
... is a must for all software engineers.
The idea was that those graphic card manufactureres who were thinging that they might develop OpenGL compatible cards were led in a particular direction which in fact has left DirectX pretty much in the driving seat.
I've not heard anything since. So perhaps this approach worked to stifle any competition for a while (although NVIDIA still seem to offer OpenGL stanards for their cards).
I don't know why its scary - the point about this "sealed" box is that it is cheap. Think more along the lines that you have a several ball point pens (maybe each with a different colour inks) than a quill pen and lots of bottles of ink.
No, it's not (necessary). I like OO, but it is not the silver bullet that kills all programming problems. Other, radically different styles (functional programming springs to mind) have shown big increases in effectiveness over OO languages, by any useful metric you care to choose, at least in certain areas. OO is good, but hardly perfect and clearly not necessary based on the number of non-OO projects in the world.
I too like the tone of the top level post, and I think the quoted bit above starts to make a valid point.However I think it misses one fundemental ...
Programming in an object oriented way is fundementally different from programming in an functional way. In the first "processes" are smeared across a number of actors (objects) which cluster the parts in a number of separate functions around the data they are working on. In the second "processes" are implemented (as functions) in one place.
Just using an object oriented language does not guarentee that you use an object oriented approach. In fact a programmer has to almost make a complete philosophy shift to move to an OO way having gained experience of functional programming. For programmers who try and do this as a day job (ie work as programmers for a living) this shift often takes about 6 months.
This means, I think that the effectiveness that you refer to above is as a result of this bias towards a functional style and the difficulty many OO projects have in getting it right.
However an OO project that does get it right makes massive and rapid progress once the basics are in place.
In my view - we are seeing KDE make rapid advances at the moment precisely because (at least with KDE 2 - I have no experience of KDE 1)they have got some fundementals right. The challenge they will face as they attract more supporting developers will be that the layers above which probably won't have as rigourous architectural control over them will suffer from people with a different approach start to work on the applications.
On the other hand, I don't think a new and better C++ will solve the problem and be the saviour that Anonymous Brave Guy thinks.
Then the Japanese started to attack small businesses with cheap simple photocopies which you bought. Almost wiped Xerox out.
Each to his own - I nearly always use dselect and hardly ever use apt-get.
Why?
I don't know - remember in those days we (as a company) were not connected to the internet and I had certainly not heard of open source. As a commercial company we were not in the university stream where software was shared - the only way to get software was to buy it from someone. I think we bought it along with our Unix System V licence from a UK company called The Instruction Set
It certainly wasn't a compiler licence but the actual editor. A collegue and I have been talking about it and we think we used an editor called "Author" before that. .. but is all a bit hazy.
Why not use the money (and the willing effort of thousands?) to create a media blitz on how yet another small voluntary group has been squashed by big business.
Do not spare putting your point of view that you had prior art into this argument, regardless of whether or not it will stand up in court. Just believe you are in the right.
Do say you cannot afford lawyers but try and get some media company that wants to give these guys a bloody nose the opportunity to stump up the fee.
A lot can happen in "about" 10 years. It was a bit longer than that, but in 1985 IBM (who now embrace Linux) not Microsoft was the dominant party and I had to pay to buy an Emacs licence to run on my Unix operating system (having just switched over to System V from my Microsoft Xenix operating system).
Turn and turn about
The problem with that approach is that M$'s PR department makes mincemeat out of it. So whilst you can be logically correct in your argument you still lose out because you get the media misinterpreting the words and the man in the street drawing the wrong conclusions.
I think we should start a propaganda war back - which in the obtuse way of a M$ PR department you do slightly off the side and start to talk about their licencing conditions as deliberately restricting the quality of software and the GPL as an innoculation against that cancer.
And I'll add - I used to be in brilliant coder - but in 1975:-). {Actually I was never very fast - just accurate, each person added a pint of beer into our project drinking pot for every bug found. I was great at finding bugs though:-)}.
I then spent 20 years acting in this curious interface between technology, business needs and management.
But over the last 6 years I have been in Marketing at work, and I am missing the technical input.
I am trying to build up to doing some open source coding - but one disadvantage is that the last time I coded seriously it was in Assembler (PDP 11 - correcting Dave Cutler's bugs in RSX-11M:-) ).
But I would welcome the opportunity to find other ways to help.
So you are not alone in this - and I am sure I am also not the only one who would join you in those thoughts.
We need to fight back with proposals of our own!
For instance, packets due to be forwarded also hit the input chain. Now they don't. As a result, its a lot easier to write firewall rules that are different for forwarded packets to ones reaching the firewall box.
There are other differences too.
For me, however, he makes points which we should attack back, and which are his weak points.
Surely the response to this, and one that the Open Source Movement should be shouting from the rooftops is that this is not in customers best interests. Microsoft has already been shown to stiffle things in the interest of the software business at the expense of customers. Open source actually is a far better way to get code that is stable and bug free because it emulates the scientific method and exposes it to peer review.
I feel that this argument is not used nearly enough. There are often complaints out there that large corporations are worried to switch to Linux because of the belief that it is not backed by a large corporation that guarentees them support. We should make the point that actually getting support for problems is a lot faster because one only has to raise the issue at the right place and almost immediately there are people looking at the problem and coming up with fixes.
The other problem, is that he is getting support e-mails from users of OpenSSH - demonstrating clearly that there is a confusion in users minds
My first reaction to this sort of thing is to ask why not contact the trading standards office. Is this just a UK facility?
But surely it Ximian who have put out this press release NOT HP. Its Ximian who have much to gain from this announcement, because it announces to the world that they are company that large companies do business with.
Firstly, although Linus heads up to programme to move the kernel forward, there is a vast array of people out there on the net supporting the activity. You always have to have somebody heading that programme, and whoever does do it is going to get up some peoples noses sometimes. I don't think a commercial organisation will find it any easier to control the kernel, or ensure it moves along any faster - indeed they would come under enormous political pressures from all sides during the transition phase from the way it is now. Typically the way that an organisation reacts so such pressures is to form committees. Hardly conducive to moving the kernel forward faster. In other words, to change from the way things happen now is a massive risk which should used to argue against those who see the grass as greener on the other side of an imaginary fence
Secondly, the reality is that Linux, to be useable is also a vast array of supporting applications. These two have to continue to move forward (perhaps even more so) if Linux is to see off the competition.
Whilst all this looks wonderful, it already assumes you see an image. What is actually transmitted is (presumably) a bit stream. How do you decode this into being an image?
I would not characterise my approach as better just different. What is clear, is that when looking at a system most programmers I encounter start to think about the functions that it should carry out. An initial breakdown then starts to split the system into functional silo's. This is of course wrong for an OO approach.
I, on the other hand, had always (and I started working on these things in the late '60s before OO had been characterised) had always seen things in terms on intercommunicating actors, with functionality shared between these according to an "architecture". I can vivedly remember having arguments with people about how we should structure systems - but my problem was that whilst I always had an instinctive feel about the way to do things that was always different from the rest of my collegues, I couldn't persuade them that I was right. When OO came along, and I realised that the basic concept of an entitiy, with functionality clustered around it was the paradigm I had some formality to hand what had been instinct.
That was the crux of my point. OO might be the theoretical correct way to do things, but actually there are two ways to thing about the architecture of a solution (functional decomposition or entity decomposition) and the thought processes right at the start need to follow the second approach.
I think that an approach to OO can be taught - the anecdotal evidence that is discussed in my organisation is that it takes the average person about 6 months to make the mind switch. This is the real reason why short term costs may be higher - but there is an expected longer term benefit (I say expected because I have no evidence only my own prejudiced view that this is so).
My failure to realise this in the mid '90s was why I fell off the greasy pole. I was a technically oriented manager who pushed one of my project teams to build the next product in an object oriented way. I failed to understand that whilst an OO mindset was obvious to me, to the programming team they couldn't get a handle on where to start. Net result is a project that went nowhere for 6 months and didn't get moving until I abandoned my insistence on an OO approach. This killed me with the costs of the project overspend which brought the rest of my profit centre to its knees. I am no longer a manager with these sort of responsibilities.
Even today, I still feel architectually the OO is right and works the way I happen to think - but I have to be honest and realise most programmers don't seem to have the same mindset and that kills the benefits that can be had.
I think that most of you have it wrong when you think that end users install windows. Most end users buy a computer with it already installed and configured. What they do is buy extras (hardware and software) that comes with self install (you install the CD into the drive - it automatically comes up and ask you to click on the install button - you select all the defaults and you are up and running).
When people buy hardware (PC or gaming) with Linux pre-installed and only then will software developers start developing games (in any serious way) for it.
I don't think that has to be as far off as the previous post implied, but I think it is more likely because a new hardware manufacturer enters the market with some down market home computing hardware that is not a PC.
I tend to disagree - provided things are architected reasonably well then these sorts of standards do add a degree of overhead - I guess we are talking 10% sort of numbers. With CPU speed doubling every 18 months (according to Moore's law) this sort of overhead is fluff.
Whats more, don't forget that if there is a standard that everyone is using in the open source area, then there is potentially a large number of minds that can look at these inefficiencies and optimise most of them away.
Don't forget that windows imposes standards (DirectX) on gamers too, but we only have M$ to optimise this.
Following a thread of a few days ago about the new ATA spec which allows people with copywriteable content to prevent it being copied, shouldn't the open source community lobby for the manufacturers of these disks to add to the spec a facility to allow data to be GPL'd (ie locked so that someone else can't subsequently lock it).