The real reason Linux isn't popular has more to do with marketing that technical issues. People have always pointed to one feature or another where Linux is weak and say that it won't be viable until the feature is there. The simple fact is that Linux is now ready for the desktop technically. It is the marketing which Linux and more generally open source needs to perfect.
To address the parent:
1. Fonts are not something I even notice a difference in. I can't imagine anyone making a decision on this basis.
2. Linux is now just as easy to use as Windows for the average user. Many devices will be supported without installation of special drivers, and in many respects this experience is easier than windows. For example, my GPS device plugs straight in and works. To use it under Windows I have to keep installing a driver. Not just once but every time I use it. I don't know why. I don't know how to fix it on Windows.
3. Graphics issues - Desktops like Suse and Ubuntu are well integrated with consistent styles. While there is a broader range of layouts than with Windows, this is not a barrier to adoption.
4. Lack or help. I don't know of any software which has effective help; be that Windows or Linux. Linux has man pages of course, but thats too technical. I agree that documentation could be better, but popular applications are generally easy to use without detailed help. The lack of local help is not a big factor, and is mitigated by good online resources such as FAQ's and mailing lists.
5. This last one is odd. You want a "bundle of software that fits my needs". Linux may have been inspired by a philosophy, but there is no suggestion that users must share it. The fact is that under Linux you have access to a huge number of applications out of the box. Under Windows you will need to purchase software piece at a time. I would rather just be able to download a program automatically.
None of these reasons are real reason why Linux is not popular on the desktop. One real reason is gaming support - one of the primary reasons many of my associates say they still have Windows partitions. If only I could play CS on Linux....
Say you go shopping, and in the local store there is a stand of tomatoes and a sign saying "Free Tomatoes". Inanimate objects cannot be free in the "not detained" sence. Therefore the only definition of free that makes sence is without cost. If you see a movie and a man is released from prison, and he says "thank god I'm free", it doesn't mean others can purchase him without cost, it means he is no longer detained.
The point I'm making is that the distiction between the definitions is made on the basis of whether the object concerned is a person capable of free will, or an inanimate object. This is the common case I believe. I have never seen any confusion in people using the term free, as the context defines meaning.
Now however we have "free software", which is inanimate, in that it does not have free will. It has no intent, no capacity to decide for itself. Perhaps someday software might achieve this, but right now we are talking about operating systems and word processors. If you walk into a software store and there is a sign saying "free software" the average joe would expect it to be without cost. They are not expecting it to be "free as in freedom" because its not; inanimate objects cannot be "free as in freedom".
The problem with Stallman's term is that it isn't the software which is free, but the developer. The GPL gives freedom not to software, but to people, and personanally that is far more compelling. So perhaps what we should be saying is "Open Source Software == Free Developers". Stallman is indeed a hero, but because he valued the importance of freedom of expression for developers, not the freedom of the software, which of course cannot exercise any freedom.
When I looked at developing in Java I wanted to know what tool sets I should use. With Microsoft it is easy - you buy Visual Studio, develop you apps using MS-SQL, deploy to Windows servers. I was working for a company in 1998 and implemented a non-Microsoft email system called "NT Mail". When I left they ripped the NT Mail solution out and put in Exchange just because it was from Microsoft. It is easy to simply decide "Microsoft is the way".
When you move to Open Source you have a large number of decisions to make. For those of us who are in the OSS community we are comfortable with the idea of mixing many projects together to suit your needs. For example, I use velocity instead of JSP's. If I were in the Microsoft world it would be ASP or nothing. It makes the choices easier, but removing choice decreases diversity.
However, the problem exists for developers new to open source. They ask the question - what should I be using? And they get 100 answers that are all different. Every Linux user has their own favorite distro - and it changes from year to year. We have different development environments, such as Netbeans, Eclipse, JEdit and so on. We have different languages. We have different databases. And each has its proponents who will tell you why you should use their approach.
What we do not have is a "Visual Studio" package that includes everything you need to develop software in one easy to use install. I mean include the language, development environment, database, build tools, and put it into a package that is well documented.
I first got a Fujifilm S5500, which has a nice optical zoom, reasonable resolution, good feature set including various manual options, and is flexible enough for most lighting conditions. Good photography is about controlling many exposure parameters, such as focal depth. For any light condition there is a combination of ISO, exposure speed and F stop. For small camera phones the solution is to have a small lens with infinite focus. The limitations are that it can't zoom in; except digitally, and you can't get a shot that brings out the subject. Is that bad? Well, if you just want a snap a camera phone is great.
However, even the Fujifilm, which is a great camera, has its limits. The big one for me is start up time and shutter lag. When it comes to being able to quickly react to unexpected events that last only a few seconds there is simply no substitute for a DSLR. The ability to turn on your camera and have it ready to shoot before you get the eyepeice to your eye is vital. For example, before I had my Nikon D70 I went on a trip watching dolphins. There was a fantastic display of dolphins jumping clear out of the water, but I didn't get a single shot because unless you have the shutter half way pressed on most dogitals it takes a second or so to actually fire.
However, I have now been the photographer at two weedings, and the D70 has been wonderful; you can see the shot and press the shutter in a fraction of a second. At an airshow recently I captured a boeing jet taking off at a high angle of attack; impossible with cameras with shutter lag. The quality of the images is slightly better, and the lenses provide a little more flexibility, and there are more complex lighting options, such as speedlights that are better than the built in ones on smaller digitals, but the biggest feature for me on a DSLR is the instant shutter.
The review in this context limits the test to things a camera phone can do - that is take a photo of something reasonably close and not moving. Thats not to say I think camera phones are bad or poor quality images, just that they don't have the flexibility of a DSLR.
The Novell deal will have no effect on open source developers at all. The 'license' is just not related to open source. It might provide limited coverage to Novell customers, by which I mean limited to Microsoft, but it will do nothing to protect Linux users from claims made by other companies such as the claim by FireStar against Red Hat. In any case, Novell, Red Hat and other companies have stood behind Linux with indemnity for their customers, and so Novell hasn't bought anything of benefit to their customers; only reduced their own risk profile.
Lets stop beating up on Novell; the move was pragmatic - it got them some cash so they could continue to survive. The interesting thing will happen if and when Microsoft attacks another Linux company, because then Novell will be aware of a patent, and then they won't be able to distribute according to the GPL.
Personally I think that the community has vastly over reacted to the move by Novell, and that if we want to be taken more seriously we should perhaps look at ourselves. We were baited by this one, and we took that bait hook, line and sinker.
New Zealand rules. Sure the Telecom companies have us over a barrel, the Govrenment wants us to pay several times for the same service, the health system is falling apart, public transport and roading is a joke, but on the flip side we have a Government who isn't in the pocket of big business, have the best natural scenery, a pretty relaxed way of life. I can recommend New Zealand; except that Americans arn't really that popular right now. For god sake stay and vote the Republicans out so the US can start to heal the damage the Republicans have done.
I have a nine year old son who really wants a mobile phone. I'm not willing to give him one simply because he has no real need for one. His friends all go to the same school, and if he wants to call one he just has to pick up the phone. He also has free access to the net - no filtering - but the computer is in the front room. Mobile phones also are being used as a medium to bully other kids; sending mean txt messages and generally harrasing classmates out of school as well as in school.
As somebody with a mild addiction to technology myself I know that just because you can (have a phone) doesn't mean you should. Mobile phones are unnessasary for a nine year old. I would prefer not to have one at all; except sometimes its good to be able to txt people when coordinating meetings on the run. Also, at work we have one phone between six - as our roles do not involve much external communication, and most of that communication is by email.
So why does it matter that MS gets Vista out before Christmas? I'm betting that one of the big problems Microsoft faces is that its Software Assurance program was meant to deliver value, in that it ensured the right to upgrade. If I were an IT manager who had signed up for Software Assurance and paid out tens or even hundreds of thousands of dollars, and got exactly sqat I'd probably be a little pissed. I'd probably be reluctant to sign up for another three years because even if MS to deliver Vista the older machines can't even run it, so you can't upgrade. You might also be a little worried that your entire business could be cut off at the knees if Microswoft ever decided to use Windows Genuine Advantage to disable PC's.
I'm guessing here, but I suspect this is a reaction to a recent Judge ruling that Blogging is a form of Journalism that attracts the same protection as tradisional media enjoy with regards to not divulging their sources. If just by blogging you can become a "journalist" people might start blogging to prevent the Police from catching the evil terrists.
The question I supose is - should Journalists be registered? Will we end up with a society where a select few have the right to freedom of expression. If someone you know finds out something terrible about a company, and you publish that information on condition that they remain anonymous, then why should you be forced to divulge the source? Sounds like the road to state controlled media to me.
Okay, try to understand whats being said. If you have two companies, company A who spends half its resources on developing new GPL software and half on supporting users, and company B spends all of its resources on support, company B can use all of the new stuff from company A while not commiting the programming resource. Red Hat relies on its good reputation as a good corporate citizen and loyal free software community member to maintain its position as one of the lead Linux distro providers. Red Hat is a Type A company. The only thing protecting Red Hat from a type B is the perception that Type B's are leeches, and are not supported by the community.
However, by dropping Red Hat Distro for the Desktop they have weakened their position somewhat. The average joe cares less about Red Hat now because generally they use Suse, Debian, Ubuntu, and even Fedora. They are moving towards being just another Linux Distro - and when that happens competing on service quality won't work. They will need to drop the developers and focus on providing excellent support if they lose the support and protection of the community.
Which brings be to a uncomfortable truth. Eventually the users will have to start to pay the developers again. I don't mean reverting to closed source; I mean that the myth of service revenue supporting development will die the death it deserves, and corporations will learn that there is benefit in supporting open source shared infrastructure projects. We can't afford to have open source remain something done only after hours. We need developers working on it full time, and they need to be paid. Service revenue can't be a long term solution, we need to find a way to pay for development.
Re:I got the CVS cop-out from the cscope maintaine
on
The CVS Cop-Out
·
· Score: 2, Interesting
This is not uncommon. I know of a Apache project that has had books written about it, which hasn't had a release since 2003. It is still very much in active use, and there has been a major bug in it since 2003 and a fix for about as long in CVS, but no release. Fair enough if one of the dev team doesn't feel like doing something, but for active projects with several develoeprs on a major project, its important to release.
Perhaps not once a bugfix; but some major OSS projects which are still in very active use have failed to release a new version despite people waiting on major bug fixes for over three years. Its all very well to get cute about users demanding a fix the instant they find something, but the reality is some projects just don't seem to be releasing at all.
This has two effects. First it means that developers who might want to help are discouraged because they can't see improvements released. This feeds back so that less actual development occurs. One of the core principles of OSS is to release early and release often. Ignoring this principle results in stagnation and alienation of users.
Re:The diplomatic response
on
The CVS Cop-Out
·
· Score: 2, Insightful
Without naming names I know of a high profile project that has had serious known issues for over three years without a release, although that it has been known for almost all of that time, and was fixed in the CVS tree years ago. The problem is very real; by not releasing often, or even at all, and not fixing obvious bugs quickly, it makes the project look abaondoned. The problem in this case is that the project is used as a major component of our applications, and there is no reliable forward planning on getting a real fix. We can't deploy code from CVS to clients.
This "you can fix it yourself" attitide doesn't fly because; 1. Often you don't have commit access, and even if you do fix it there is no confidence in getting it into a release anytime soon. 2. I don't have time to learn the code of every open source project I use. I do maintain my own open source projects, and for these I am happy to fix bugs and create releases, but to expect a user to make changes, even if they are technically able, if often a poor use of time.
There is an attitude that because we are not paying that users should just be happy with what is supplied. Personally I would like to see important projects properly funded, and am willing to help with that, rather than having them stagnate because developers are busy on their day job.
Just a random thought while I read your comment. As you say, Java never feels 'hacked together'. Lets look at dates. Originally there was the Date object, but as there is more than one calendar and date system in the world most of the Date functionality in Java was dropped in favour of GregorianCalendar. Suddenly date manipulation became a horrid mismash of creating the objects you needed to access the date manipulation methods required.
The same applies to Swing. In Delphi you could start a project, drop a few components on a form and hook them up with a few lines of code. With Java (NetBeans/JBuilder) just getting to the point of dropping something on a form required an in depth understanding of the Swing graphical layouts. It is perhaps another example of a very flexible system, but like the Date changes, it shows two underlying issue; being more generic (ie catering for a wider class of situations) was valued over being easier to use for the majority.
Sun engineers seem to have come up with systems which cater for complex situations, while developers trying to do real world apps are left writing huge swags of code to do reasonably simple things.
Another two examples. EJB. When I started a project about four years ago I had to make a choice; whether to use EJB as a persistence layer; which was viewed at the time as the 'correct' way, to manually code JDBC into data persistence objects, or to design my own persistence engine. Well, I decided to code my own persistence engine, and the design decisions meant it turned out pretty much like Hibernate looks today. Why does Hibernate exist? Because EJB was over engineered, too complex to use for the vast majority of projects.
Second example is Mail. I tried to use the Mail API that came with Java to send emails. Once again it seemed that it had been over engineered to such an extent that it was impossible to use. I resorted to a small open source class - a single class - that could do everything very simply. Perhaps the class wouldn't do everything the JavaMail API could, but it did everything I needed, and perhaps everything most other developers would need as well.
This is a common thread; over engineering everything to be generic while making life more difficult for the majority of coders. Luckily projects such as Hibernate, Struts, Tomcat and Velocity make development much easier for me day to day.
Sure, its possible to write maintainable code in any language if you have a skilled developer who has the disipline to code that way. However, the difference here is whether the Language itself makes it hard or easy to write nice code. With Java the language itself enforces a certain way of doing things; primarily encapsulation of everything in a Object. It also has very definite convensions which you will be derided for if broken.
For small systems developed by one person Perl might be okay, but when you have whole teams of people with various skill levels (you can't always have the best) Java has what it takes to encourage good coding practise. Perl on the other hand encourages the "many ways to skin a cat" approach. Being interpreted brevity is valued over clarity, and that hurts when it comes to trying to maintain stuff after the fact.
Now, I am a Java developer, but I also like Python - its not quite as strict as Java, but doesn't end up with the same hideous spagetti code messes that I've seen in Perl and PHP. Is Java perfect? Hell no. If I had a choice to make just on Language it would be Python. Sadly there are still not the frameworks available for Python to the same level as Java; although projects like DJango are getting good.
You can't prove a negitive. If there is a statement that child porn is increasingly using younger children, and you are using this as a basis for a new law, you *better* come prepared with some evidence to back up your position. Even then you better make sure the new bill is also targetting the crime at issue.
The problem here is that keeping these records is only incedentally useful, and may not even be admissible in a court of law, since you can't prove that they user of the account is also the computer operator. There may also be more than one user of a computer, and it may have been hijacked.
The point Mr Quinn is making isn't that the Linux crowd in general should wear suits. The average programmer isn't going to be doing sales. I personally think the issue is a lack of open source advocates who are business friendly at all. When a organisation wants to move to open source they want to know the company doing it is professional. Having someone turn up in a T Shirt and jandels doesn't do much for their confidence.
Appearance is important, we cannot possibly invest the effort to get to know 'the real person' for everyone we meet. In order to function we need our sterotypes. Its a bit like justice - justice not only must be done, it must be seen to be done. Similarly, quality in service must be seen. Thats not to say everyone needs to be in a suit; but certainly if you are making a open source presentation to a large organisation you should have appropriate attire; or not bother at all.
Not bothering is fine of course - nobody is forcing you to advocate and sell OSS solutions, but if you are in that game, and you do care about getting that contract, then perhaps how you dress will impact your chances.
The cost of switching for most people and organisations will be pretty low. Some organisations which do depend on complex macros and templates may have difficulty, but in my experience organisations implementing OpenOffice have faced very few implementation issues. Having other applications be able to modify documents has been the bane of my life prior to OO when it comes to document management. OpenOffice made document generation easy without needing to have server side Word instances.
Often Office is used as a glorified viewer - where people have Office just to view content created by others. In the last six months the only PPT files I've received were slideshows of joke images; seems that power point is being used more for recreation than business.
I agree. The comparision of source code to slaves is terrible. We should be fighting for freedoms of people, not source code. Open Source brings freedom to developers, allows them to build their own culture not owned by corporations. We have a strong moral sense that people should be free to share if they choose. What I strongly disagree with is Stallmans misguided and unethical attitude to having the right to use the work of others even if its against their will.
Okay, so you get a great new job defending the fatherland - sorry - homeland from evil terrorists. You generally wander around in your small hick town for a couple of years, but no turbin wearing insane arabs appear to be around, so you have to do something to justify your salary. What could be better than protecting the public from indecency; surely your local conservative community would back a hard line on "protecting the children"?
Oh god, not another big up front design fan. You don't need in depth design documents, you simply need developers who have been properly briefed about what the project is, and have proper oversight. I run a development shop, and I've got guys who can pretty much manage themselves once given a task. They have great lattitude in how they get the job done, but they work on what they are given, not on their pet project.
This guy didn't really put the vision across, and didn't have a user in the process. Where were te users? In open source the developer is the user, and thus one person can be everything. But in this project the users are school teachers. Where were they? Were there user stories provided by users? Was there any consultation with users? This is really they key, not design documents that go stale.
Developers will always want to use the tool they think is best, however if every developer in a organisation used what they thought was best you would have huge interoperability and intergration issues. At some point a manager type has to make a decision on how they project will be written, and usually its one language - perhaps two. A application may require several technologies, and some of those technologies will be more advanced in another language. However, overall it is better to take a hit on the fringe than fracture the architecture. For example, Ant and JUnit provide a way to autorun unit tests on build. Having Python, perl and whatever else in the mix just makes it harder to test - especially the interfaces between those languages.
This is somewhat true. If a company doesn't keep up with the times it will rot. However there is more than just a boolean decision here. You can choose to develop in a language, and then later move to another language, training all the developers equally. The problem with no standardization is that you have developers who become indispensable because they know weird toolsets required for specific projects. It might feel nice for developers to have the freedom to write in anything they wish, but in a larger organisation it doesn't lead to flexibility at all.
A classic example is a company I used to work for who used MagicLanguageX. MagicLanguageX (X) was retired by the company who made it, but the victim company had developed a entire system with it, on which the business depended. The developers of the system are now the only ones who know the language. They are paid a heap of money, but there is zero demand for them outside. Company bought itself into a huge problem because of a poor choice of language.
There are a few things I could say about standardization:
First of all it means that any developer can work on any project. If you have a developer leave, die or go on holiday their projects are not left in limbo while some other developer gets up to speed. In large shops it might not matter so much, but with a small shop of four or five developers it's a different story.
Maintaining a level of professional proficiency in any language means spending time on a regular basis developing in it. Languages and utilities change all the time, and to keep up takes time. I can't imagine a single developer being proficient in more than three languages. For example, I used to do Delphi, but havn't really done anything in it in a couple of years, so I'm not really in touch with it anymore.
From a business perspective its good to concentrate on being good with one technology rather than being mediocre in several. Learning new languages takes time, and so having standardized on one language means not needing to train new employees in several languages, and also keeps the employment pool wide.
It means you can be clear about which projects you are aiming for, who your clients are, and allows you to concentrate on what you are good at. Obviously you don't want to chose a language which paints you into a corner - it must be flexible, generic, well supported by employee availability, and accepted by clients (and yes, some clients do care).
However, that doesn't mean you stop evaluating the options. Right now we are moving from Java technology towards Python. Our bread and butter is still Java, but Python is giving us a faster more effective way to develop web applications without sacrificing our favorite language features. We even mix languages in projects.
However, unregulated use of languages is not permitted because it would mean having no clear strategy for the future support and maintenance of the project. Moving to python for me means moving all our developers to that technology, and making sure we don't lose the company advantages we built up by using Java.
Slashdot Mirror
The real reason Linux isn't popular has more to do with marketing that technical issues. People have always pointed to one feature or another where Linux is weak and say that it won't be viable until the feature is there. The simple fact is that Linux is now ready for the desktop technically. It is the marketing which Linux and more generally open source needs to perfect.
To address the parent:
1. Fonts are not something I even notice a difference in. I can't imagine anyone making a decision on this basis.
2. Linux is now just as easy to use as Windows for the average user. Many devices will be supported without installation of special drivers, and in many respects this experience is easier than windows. For example, my GPS device plugs straight in and works. To use it under Windows I have to keep installing a driver. Not just once but every time I use it. I don't know why. I don't know how to fix it on Windows.
3. Graphics issues - Desktops like Suse and Ubuntu are well integrated with consistent styles. While there is a broader range of layouts than with Windows, this is not a barrier to adoption.
4. Lack or help. I don't know of any software which has effective help; be that Windows or Linux. Linux has man pages of course, but thats too technical. I agree that documentation could be better, but popular applications are generally easy to use without detailed help. The lack of local help is not a big factor, and is mitigated by good online resources such as FAQ's and mailing lists.
5. This last one is odd. You want a "bundle of software that fits my needs". Linux may have been inspired by a philosophy, but there is no suggestion that users must share it. The fact is that under Linux you have access to a huge number of applications out of the box. Under Windows you will need to purchase software piece at a time. I would rather just be able to download a program automatically.
None of these reasons are real reason why Linux is not popular on the desktop. One real reason is gaming support - one of the primary reasons many of my associates say they still have Windows partitions. If only I could play CS on Linux....
Say you go shopping, and in the local store there is a stand of tomatoes and a sign saying "Free Tomatoes". Inanimate objects cannot be free in the "not detained" sence. Therefore the only definition of free that makes sence is without cost. If you see a movie and a man is released from prison, and he says "thank god I'm free", it doesn't mean others can purchase him without cost, it means he is no longer detained.
The point I'm making is that the distiction between the definitions is made on the basis of whether the object concerned is a person capable of free will, or an inanimate object. This is the common case I believe. I have never seen any confusion in people using the term free, as the context defines meaning.
Now however we have "free software", which is inanimate, in that it does not have free will. It has no intent, no capacity to decide for itself. Perhaps someday software might achieve this, but right now we are talking about operating systems and word processors. If you walk into a software store and there is a sign saying "free software" the average joe would expect it to be without cost. They are not expecting it to be "free as in freedom" because its not; inanimate objects cannot be "free as in freedom".
The problem with Stallman's term is that it isn't the software which is free, but the developer. The GPL gives freedom not to software, but to people, and personanally that is far more compelling. So perhaps what we should be saying is "Open Source Software == Free Developers". Stallman is indeed a hero, but because he valued the importance of freedom of expression for developers, not the freedom of the software, which of course cannot exercise any freedom.
When I looked at developing in Java I wanted to know what tool sets I should use. With Microsoft it is easy - you buy Visual Studio, develop you apps using MS-SQL, deploy to Windows servers. I was working for a company in 1998 and implemented a non-Microsoft email system called "NT Mail". When I left they ripped the NT Mail solution out and put in Exchange just because it was from Microsoft. It is easy to simply decide "Microsoft is the way".
When you move to Open Source you have a large number of decisions to make. For those of us who are in the OSS community we are comfortable with the idea of mixing many projects together to suit your needs. For example, I use velocity instead of JSP's. If I were in the Microsoft world it would be ASP or nothing. It makes the choices easier, but removing choice decreases diversity.
However, the problem exists for developers new to open source. They ask the question - what should I be using? And they get 100 answers that are all different. Every Linux user has their own favorite distro - and it changes from year to year. We have different development environments, such as Netbeans, Eclipse, JEdit and so on. We have different languages. We have different databases. And each has its proponents who will tell you why you should use their approach.
What we do not have is a "Visual Studio" package that includes everything you need to develop software in one easy to use install. I mean include the language, development environment, database, build tools, and put it into a package that is well documented.
I first got a Fujifilm S5500, which has a nice optical zoom, reasonable resolution, good feature set including various manual options, and is flexible enough for most lighting conditions. Good photography is about controlling many exposure parameters, such as focal depth. For any light condition there is a combination of ISO, exposure speed and F stop. For small camera phones the solution is to have a small lens with infinite focus. The limitations are that it can't zoom in; except digitally, and you can't get a shot that brings out the subject. Is that bad? Well, if you just want a snap a camera phone is great.
However, even the Fujifilm, which is a great camera, has its limits. The big one for me is start up time and shutter lag. When it comes to being able to quickly react to unexpected events that last only a few seconds there is simply no substitute for a DSLR. The ability to turn on your camera and have it ready to shoot before you get the eyepeice to your eye is vital. For example, before I had my Nikon D70 I went on a trip watching dolphins. There was a fantastic display of dolphins jumping clear out of the water, but I didn't get a single shot because unless you have the shutter half way pressed on most dogitals it takes a second or so to actually fire.
However, I have now been the photographer at two weedings, and the D70 has been wonderful; you can see the shot and press the shutter in a fraction of a second. At an airshow recently I captured a boeing jet taking off at a high angle of attack; impossible with cameras with shutter lag. The quality of the images is slightly better, and the lenses provide a little more flexibility, and there are more complex lighting options, such as speedlights that are better than the built in ones on smaller digitals, but the biggest feature for me on a DSLR is the instant shutter.
The review in this context limits the test to things a camera phone can do - that is take a photo of something reasonably close and not moving. Thats not to say I think camera phones are bad or poor quality images, just that they don't have the flexibility of a DSLR.
The Novell deal will have no effect on open source developers at all. The 'license' is just not related to open source. It might provide limited coverage to Novell customers, by which I mean limited to Microsoft, but it will do nothing to protect Linux users from claims made by other companies such as the claim by FireStar against Red Hat. In any case, Novell, Red Hat and other companies have stood behind Linux with indemnity for their customers, and so Novell hasn't bought anything of benefit to their customers; only reduced their own risk profile.
Lets stop beating up on Novell; the move was pragmatic - it got them some cash so they could continue to survive. The interesting thing will happen if and when Microsoft attacks another Linux company, because then Novell will be aware of a patent, and then they won't be able to distribute according to the GPL.
Personally I think that the community has vastly over reacted to the move by Novell, and that if we want to be taken more seriously we should perhaps look at ourselves. We were baited by this one, and we took that bait hook, line and sinker.
New Zealand rules. Sure the Telecom companies have us over a barrel, the Govrenment wants us to pay several times for the same service, the health system is falling apart, public transport and roading is a joke, but on the flip side we have a Government who isn't in the pocket of big business, have the best natural scenery, a pretty relaxed way of life. I can recommend New Zealand; except that Americans arn't really that popular right now. For god sake stay and vote the Republicans out so the US can start to heal the damage the Republicans have done.
I have a nine year old son who really wants a mobile phone. I'm not willing to give him one simply because he has no real need for one. His friends all go to the same school, and if he wants to call one he just has to pick up the phone. He also has free access to the net - no filtering - but the computer is in the front room. Mobile phones also are being used as a medium to bully other kids; sending mean txt messages and generally harrasing classmates out of school as well as in school.
As somebody with a mild addiction to technology myself I know that just because you can (have a phone) doesn't mean you should. Mobile phones are unnessasary for a nine year old. I would prefer not to have one at all; except sometimes its good to be able to txt people when coordinating meetings on the run. Also, at work we have one phone between six - as our roles do not involve much external communication, and most of that communication is by email.
So why does it matter that MS gets Vista out before Christmas? I'm betting that one of the big problems Microsoft faces is that its Software Assurance program was meant to deliver value, in that it ensured the right to upgrade. If I were an IT manager who had signed up for Software Assurance and paid out tens or even hundreds of thousands of dollars, and got exactly sqat I'd probably be a little pissed. I'd probably be reluctant to sign up for another three years because even if MS to deliver Vista the older machines can't even run it, so you can't upgrade. You might also be a little worried that your entire business could be cut off at the knees if Microswoft ever decided to use Windows Genuine Advantage to disable PC's.
I'm guessing here, but I suspect this is a reaction to a recent Judge ruling that Blogging is a form of Journalism that attracts the same protection as tradisional media enjoy with regards to not divulging their sources. If just by blogging you can become a "journalist" people might start blogging to prevent the Police from catching the evil terrists.
The question I supose is - should Journalists be registered? Will we end up with a society where a select few have the right to freedom of expression. If someone you know finds out something terrible about a company, and you publish that information on condition that they remain anonymous, then why should you be forced to divulge the source? Sounds like the road to state controlled media to me.
Okay, try to understand whats being said. If you have two companies, company A who spends half its resources on developing new GPL software and half on supporting users, and company B spends all of its resources on support, company B can use all of the new stuff from company A while not commiting the programming resource. Red Hat relies on its good reputation as a good corporate citizen and loyal free software community member to maintain its position as one of the lead Linux distro providers. Red Hat is a Type A company. The only thing protecting Red Hat from a type B is the perception that Type B's are leeches, and are not supported by the community.
However, by dropping Red Hat Distro for the Desktop they have weakened their position somewhat. The average joe cares less about Red Hat now because generally they use Suse, Debian, Ubuntu, and even Fedora. They are moving towards being just another Linux Distro - and when that happens competing on service quality won't work. They will need to drop the developers and focus on providing excellent support if they lose the support and protection of the community.
Which brings be to a uncomfortable truth. Eventually the users will have to start to pay the developers again. I don't mean reverting to closed source; I mean that the myth of service revenue supporting development will die the death it deserves, and corporations will learn that there is benefit in supporting open source shared infrastructure projects. We can't afford to have open source remain something done only after hours. We need developers working on it full time, and they need to be paid. Service revenue can't be a long term solution, we need to find a way to pay for development.
This is not uncommon. I know of a Apache project that has had books written about it, which hasn't had a release since 2003. It is still very much in active use, and there has been a major bug in it since 2003 and a fix for about as long in CVS, but no release. Fair enough if one of the dev team doesn't feel like doing something, but for active projects with several develoeprs on a major project, its important to release.
Perhaps not once a bugfix; but some major OSS projects which are still in very active use have failed to release a new version despite people waiting on major bug fixes for over three years. Its all very well to get cute about users demanding a fix the instant they find something, but the reality is some projects just don't seem to be releasing at all.
This has two effects. First it means that developers who might want to help are discouraged because they can't see improvements released. This feeds back so that less actual development occurs. One of the core principles of OSS is to release early and release often. Ignoring this principle results in stagnation and alienation of users.
Without naming names I know of a high profile project that has had serious known issues for over three years without a release, although that it has been known for almost all of that time, and was fixed in the CVS tree years ago. The problem is very real; by not releasing often, or even at all, and not fixing obvious bugs quickly, it makes the project look abaondoned. The problem in this case is that the project is used as a major component of our applications, and there is no reliable forward planning on getting a real fix. We can't deploy code from CVS to clients.
This "you can fix it yourself" attitide doesn't fly because; 1. Often you don't have commit access, and even if you do fix it there is no confidence in getting it into a release anytime soon. 2. I don't have time to learn the code of every open source project I use. I do maintain my own open source projects, and for these I am happy to fix bugs and create releases, but to expect a user to make changes, even if they are technically able, if often a poor use of time.
There is an attitude that because we are not paying that users should just be happy with what is supplied. Personally I would like to see important projects properly funded, and am willing to help with that, rather than having them stagnate because developers are busy on their day job.
Just a random thought while I read your comment. As you say, Java never feels 'hacked together'. Lets look at dates. Originally there was the Date object, but as there is more than one calendar and date system in the world most of the Date functionality in Java was dropped in favour of GregorianCalendar. Suddenly date manipulation became a horrid mismash of creating the objects you needed to access the date manipulation methods required.
The same applies to Swing. In Delphi you could start a project, drop a few components on a form and hook them up with a few lines of code. With Java (NetBeans/JBuilder) just getting to the point of dropping something on a form required an in depth understanding of the Swing graphical layouts. It is perhaps another example of a very flexible system, but like the Date changes, it shows two underlying issue; being more generic (ie catering for a wider class of situations) was valued over being easier to use for the majority.
Sun engineers seem to have come up with systems which cater for complex situations, while developers trying to do real world apps are left writing huge swags of code to do reasonably simple things.
Another two examples. EJB. When I started a project about four years ago I had to make a choice; whether to use EJB as a persistence layer; which was viewed at the time as the 'correct' way, to manually code JDBC into data persistence objects, or to design my own persistence engine. Well, I decided to code my own persistence engine, and the design decisions meant it turned out pretty much like Hibernate looks today. Why does Hibernate exist? Because EJB was over engineered, too complex to use for the vast majority of projects.
Second example is Mail. I tried to use the Mail API that came with Java to send emails. Once again it seemed that it had been over engineered to such an extent that it was impossible to use. I resorted to a small open source class - a single class - that could do everything very simply. Perhaps the class wouldn't do everything the JavaMail API could, but it did everything I needed, and perhaps everything most other developers would need as well.
This is a common thread; over engineering everything to be generic while making life more difficult for the majority of coders. Luckily projects such as Hibernate, Struts, Tomcat and Velocity make development much easier for me day to day.
Sure, its possible to write maintainable code in any language if you have a skilled developer who has the disipline to code that way. However, the difference here is whether the Language itself makes it hard or easy to write nice code. With Java the language itself enforces a certain way of doing things; primarily encapsulation of everything in a Object. It also has very definite convensions which you will be derided for if broken.
For small systems developed by one person Perl might be okay, but when you have whole teams of people with various skill levels (you can't always have the best) Java has what it takes to encourage good coding practise. Perl on the other hand encourages the "many ways to skin a cat" approach. Being interpreted brevity is valued over clarity, and that hurts when it comes to trying to maintain stuff after the fact.
Now, I am a Java developer, but I also like Python - its not quite as strict as Java, but doesn't end up with the same hideous spagetti code messes that I've seen in Perl and PHP. Is Java perfect? Hell no. If I had a choice to make just on Language it would be Python. Sadly there are still not the frameworks available for Python to the same level as Java; although projects like DJango are getting good.
You can't prove a negitive. If there is a statement that child porn is increasingly using younger children, and you are using this as a basis for a new law, you *better* come prepared with some evidence to back up your position. Even then you better make sure the new bill is also targetting the crime at issue.
The problem here is that keeping these records is only incedentally useful, and may not even be admissible in a court of law, since you can't prove that they user of the account is also the computer operator. There may also be more than one user of a computer, and it may have been hijacked.
The point Mr Quinn is making isn't that the Linux crowd in general should wear suits. The average programmer isn't going to be doing sales. I personally think the issue is a lack of open source advocates who are business friendly at all. When a organisation wants to move to open source they want to know the company doing it is professional. Having someone turn up in a T Shirt and jandels doesn't do much for their confidence.
Appearance is important, we cannot possibly invest the effort to get to know 'the real person' for everyone we meet. In order to function we need our sterotypes. Its a bit like justice - justice not only must be done, it must be seen to be done. Similarly, quality in service must be seen. Thats not to say everyone needs to be in a suit; but certainly if you are making a open source presentation to a large organisation you should have appropriate attire; or not bother at all.
Not bothering is fine of course - nobody is forcing you to advocate and sell OSS solutions, but if you are in that game, and you do care about getting that contract, then perhaps how you dress will impact your chances.
The cost of switching for most people and organisations will be pretty low. Some organisations which do depend on complex macros and templates may have difficulty, but in my experience organisations implementing OpenOffice have faced very few implementation issues. Having other applications be able to modify documents has been the bane of my life prior to OO when it comes to document management. OpenOffice made document generation easy without needing to have server side Word instances.
Often Office is used as a glorified viewer - where people have Office just to view content created by others. In the last six months the only PPT files I've received were slideshows of joke images; seems that power point is being used more for recreation than business.
I second that. The headline is flat out wrong.
I agree. The comparision of source code to slaves is terrible. We should be fighting for freedoms of people, not source code. Open Source brings freedom to developers, allows them to build their own culture not owned by corporations. We have a strong moral sense that people should be free to share if they choose. What I strongly disagree with is Stallmans misguided and unethical attitude to having the right to use the work of others even if its against their will.
Okay, so you get a great new job defending the fatherland - sorry - homeland from evil terrorists. You generally wander around in your small hick town for a couple of years, but no turbin wearing insane arabs appear to be around, so you have to do something to justify your salary. What could be better than protecting the public from indecency; surely your local conservative community would back a hard line on "protecting the children"?
Oh god, not another big up front design fan. You don't need in depth design documents, you simply need developers who have been properly briefed about what the project is, and have proper oversight. I run a development shop, and I've got guys who can pretty much manage themselves once given a task. They have great lattitude in how they get the job done, but they work on what they are given, not on their pet project.
This guy didn't really put the vision across, and didn't have a user in the process. Where were te users? In open source the developer is the user, and thus one person can be everything. But in this project the users are school teachers. Where were they? Were there user stories provided by users? Was there any consultation with users? This is really they key, not design documents that go stale.
Developers will always want to use the tool they think is best, however if every developer in a organisation used what they thought was best you would have huge interoperability and intergration issues. At some point a manager type has to make a decision on how they project will be written, and usually its one language - perhaps two. A application may require several technologies, and some of those technologies will be more advanced in another language. However, overall it is better to take a hit on the fringe than fracture the architecture. For example, Ant and JUnit provide a way to autorun unit tests on build. Having Python, perl and whatever else in the mix just makes it harder to test - especially the interfaces between those languages.
This is somewhat true. If a company doesn't keep up with the times it will rot. However there is more than just a boolean decision here. You can choose to develop in a language, and then later move to another language, training all the developers equally. The problem with no standardization is that you have developers who become indispensable because they know weird toolsets required for specific projects. It might feel nice for developers to have the freedom to write in anything they wish, but in a larger organisation it doesn't lead to flexibility at all.
A classic example is a company I used to work for who used MagicLanguageX. MagicLanguageX (X) was retired by the company who made it, but the victim company had developed a entire system with it, on which the business depended. The developers of the system are now the only ones who know the language. They are paid a heap of money, but there is zero demand for them outside. Company bought itself into a huge problem because of a poor choice of language.
There are a few things I could say about standardization:
First of all it means that any developer can work on any project. If you have a developer leave, die or go on holiday their projects are not left in limbo while some other developer gets up to speed. In large shops it might not matter so much, but with a small shop of four or five developers it's a different story.
Maintaining a level of professional proficiency in any language means spending time on a regular basis developing in it. Languages and utilities change all the time, and to keep up takes time. I can't imagine a single developer being proficient in more than three languages. For example, I used to do Delphi, but havn't really done anything in it in a couple of years, so I'm not really in touch with it anymore.
From a business perspective its good to concentrate on being good with one technology rather than being mediocre in several. Learning new languages takes time, and so having standardized on one language means not needing to train new employees in several languages, and also keeps the employment pool wide.
It means you can be clear about which projects you are aiming for, who your clients are, and allows you to concentrate on what you are good at. Obviously you don't want to chose a language which paints you into a corner - it must be flexible, generic, well supported by employee availability, and accepted by clients (and yes, some clients do care).
However, that doesn't mean you stop evaluating the options. Right now we are moving from Java technology towards Python. Our bread and butter is still Java, but Python is giving us a faster more effective way to develop web applications without sacrificing our favorite language features. We even mix languages in projects.
However, unregulated use of languages is not permitted because it would mean having no clear strategy for the future support and maintenance of the project. Moving to python for me means moving all our developers to that technology, and making sure we don't lose the company advantages we built up by using Java.