You have to take it in context. Windows 95 may be useless now, but back at the time it offered features other systems didn't, if not in its own code, then through the third-party ecosystem it created.
For a business relying on critical software, being able to legally and practically hire a contractor to fix a problem in open source software is a huge advantage over having to track down a developer legally and technically able to fix a problem.
Even if 99% of people can't fix the problem, having that 1% is enough to save a business. If it's 99.9999% of people who can't fix it, leaving a mere handfull of developers who can (for legal or technical reasons), you're pretty much sunk and have to take the disaster recovery or migration cost head-on.
Open source is a guarantee that things can be fixed legally and practically. You may not need it, but if you do, it can save your business. A lot of companies learn that the hard way, and that's why open source and open standards are growing and growing.
That's just the thing... If you don't care, why did you bother being hostile in response to hostilities? Now you've just riled up more people who've made more hostile responses. You must get something out of it, otherwise why do it?
If you don't care, don't do it. If you do care, do it better. I hope I'm not coming off as hostile. I'm just not sure I understand how this works for you.
You know, if you manage to offend an entire tree of responders, you may be doing something wrong. It's kind of like those smartasses everyone's met in school/university by now, who may even be right, but aren't making any friends showing off. I'm just saying, take a close look at your priorities.
Right, because with an installer that occupies an entire DVD, Visual Studio is *so* much leaner than Eclipse' 100-200MB + JRE.
You can fit Eclipse with JDT, CDT, PyDev, RDT, Subclipse, WST, DTP, etc. and the JDK (which includes source and documentation for the entire API), Python, Ruby, and heaps more, on one CD, with room left over. I know because I've done it. 7zip is your friend.
It's good you bring up plastic surgery, because that used to be the exclusive domain of the rich, and now it's almost a commodity in some countries. It's clear genetic engineering will be too, the question is when.
Regular mating is pretty lossy too. For all we know, advancements in cloning experimentation could lead to improvements in conception safety.
It may even be the case that we will be able to submit genetic samples from two partners, regardless of gender, and have a technological process create the new genome and begin its gestation. We may even be able to select parameters in the process, such as selecting gender and which traits to inherit.
I prefer to see this as the next step up for medicine, not as "playing God". We've come a long way from caves and flint axes, and it's only the fundamentalists and conservatives who insist what we have is where it should end.
Microsoft Office could pass a standard ODF-ACID test but still allow extensions. These won't be part of the standard, but if the document itself still qualifies for the standard, it will only work in Office anyway. Microsoft has entire divisions dedicated to this kind of evil.
Fair enough. I like H2 for pure Java projects, because although SQLite is my favorite in general, its Java bindings suck. H2 is pure Java (no bindings, no native components, so you can embed it anywhere) and performs very well. You can even run a client-server split if you find you need that, but if it comes to that I'd rather just use PostgreSQL and the JDBC driver for that (which, IIRC, is also pure Java).
I can think of a case where Ctrl+C could be a leak of information exploit... if you rely on your program to zero its memory before shutting down, Ctrl+C could kill it before it zeroes its memory and the sensitive information will be left over, subject to any of the usual memory exploits. If this is on a swap file you're even worse off.
All of that pre-installed crapware subsidises the cost of the machine. The more there is, the cheaper the machine gets.
Once you've bought it, you're welcome to install Linux on a clean slate, but unless you bought a separate XP CD you're probably going to get the crapware back every time you reinstall or recover.
I build my own machines too, and for laptops I don't even let them boot until after I've wiped the default install. I don't even want the initial "phone home" most crapware does on those installs. The less they gain from their practices, the better. Honest market research is one thing, but installing spies and advertisers on our machines is plain evil.
This kind of fear and paranoia is exactly what Microsoft wants, and insituting this kind of witchunt is going to do much more damage than any malicious contributor possibly could.
Linux has always had a "bad" security record. It is not difficult for a motivated and skilled attacker to, at the very least, cause severe denial of service against a Linux host. Black hats have many more exploits than developers and auditors know about.
If Microsoft really wants to damage the Linux ecosystem, and legal means (patent threats, copyright threats, market confusion, etc) run out, the next logical resort will be to hire blackhats. They may already have some. Why would they tell anyone?
The recent Debian/Ubuntu vulnerability, which technically left me trivially vulnerable for two years, made me seriously consider switching back to FreeBSD. Fortunately my OpenVPN keys were generated specifically on BSD so they were not vulnerable. It's not like BSD hasn't had exploits either, but Ubuntu in particular has been an embarrasment from its first release, where the system administrator password was recorded in a world-readable log file.
Well, we do have that. GNOME is the beginner mode and KDE is the expert mode. XFce is somewhere in the middle.
Personally I prefer GNOME's design lately, even though KDE's technology is vastly better even in 3.5. Being based on Qt gives KDE a serious advantage, but it's a shame that doesn't always result in elegant designs.
Put down the crack pipe..NET is based on the same bytecode + JIT design as Java, with years less maturity. Microsoft only created.NET in the first place because they were legally barred from "embrace, extend, extinguish"ing Java itself..NET has its merits over Java (and vice versa), but speed is almost identical.
Back at the time, this meant Linux could serve static files over HTTP with a tiny tiny fraction of the resources required for Windows or even BSD to do the same. So for that little kernel option and configuration, you could save 90% of your static file server resources.
Now Linux and user-space software alike have been optimized to the point that khttpd's performance is barely better and sometimes worse, so it is no longer practical.
I used to be a userland purist, but let's face it, back when all we had were 133Mhz Pentiums, fast is good. Microkernels never took off for the same reasons - elegant and maintainable are great but sometimes you just need fast. Otherwise we'd have kernels written in Python.
Steve Jobs is Buddhist. He could believe he is a Buddha himself, and the funny thing is, even if he's wrong there's nothing wrong with that. It's not like there's an ISO Standard Buddha, and nobody would take his sutras seriously anyway.
That's true, Python list comprehensions do not support ordering. However, since Python is partly a functional language anyway, it doesn't matter:
a = sorted([comprehension of b])
Or specifying your own comparator:
a = sorted([comprehension of b], cmp = comparator or lambda)
It's even cleaner in Ruby which has nicer closures, but there it's OO instead of functional.
You already know that, but I don't suppose every Slashdot reader does.
And while you have a point that Python list comprehensions in particular do not support being translated to SQL, Ruby's OO version certainly do. The.filter,.map, etc. can all generate code behind the scenes, which is only run when the result is requested as an iretable. ORMs do this, though I haven't used any Ruby ORMs personally.
Except that it affects all keys generated on Debian. Even if your client is not running Debian, any number of your servers may be, whether they're in your administrative domain or not. Keys may have been generated on Debian machines and then used on Windows servers. The companies you rely on to store your sensitive details may have already been compromised. How the hell can you say you're not affected? Everyone's potentially affected. That's kind of the problem with PKI in general, which is meant to be "solved" with strong entropy.
When you already have a dynamic language, type inference, anonymous types and extension methods are implicit. I'm not arguing every language should be dynamic, but the additions to C# are just solving problems specific to static languages. That's fine - static languages are still much faster than dynamic ones. Just saying, as far as languages go, C# is the one playing catchup, and with about a 20 year gap for many features.
Slashdot Mirror
You have to take it in context. Windows 95 may be useless now, but back at the time it offered features other systems didn't, if not in its own code, then through the third-party ecosystem it created.
For a business relying on critical software, being able to legally and practically hire a contractor to fix a problem in open source software is a huge advantage over having to track down a developer legally and technically able to fix a problem.
Even if 99% of people can't fix the problem, having that 1% is enough to save a business. If it's 99.9999% of people who can't fix it, leaving a mere handfull of developers who can (for legal or technical reasons), you're pretty much sunk and have to take the disaster recovery or migration cost head-on.
Open source is a guarantee that things can be fixed legally and practically. You may not need it, but if you do, it can save your business. A lot of companies learn that the hard way, and that's why open source and open standards are growing and growing.
In that case, it's kind of like this: http://xkcd.com/386/
It's funny because it's true.
That's just the thing... If you don't care, why did you bother being hostile in response to hostilities? Now you've just riled up more people who've made more hostile responses. You must get something out of it, otherwise why do it?
If you don't care, don't do it. If you do care, do it better. I hope I'm not coming off as hostile. I'm just not sure I understand how this works for you.
You know, if you manage to offend an entire tree of responders, you may be doing something wrong. It's kind of like those smartasses everyone's met in school/university by now, who may even be right, but aren't making any friends showing off. I'm just saying, take a close look at your priorities.
Right, because with an installer that occupies an entire DVD, Visual Studio is *so* much leaner than Eclipse' 100-200MB + JRE.
You can fit Eclipse with JDT, CDT, PyDev, RDT, Subclipse, WST, DTP, etc. and the JDK (which includes source and documentation for the entire API), Python, Ruby, and heaps more, on one CD, with room left over. I know because I've done it. 7zip is your friend.
It's good you bring up plastic surgery, because that used to be the exclusive domain of the rich, and now it's almost a commodity in some countries. It's clear genetic engineering will be too, the question is when.
Regular mating is pretty lossy too. For all we know, advancements in cloning experimentation could lead to improvements in conception safety.
It may even be the case that we will be able to submit genetic samples from two partners, regardless of gender, and have a technological process create the new genome and begin its gestation. We may even be able to select parameters in the process, such as selecting gender and which traits to inherit.
I prefer to see this as the next step up for medicine, not as "playing God". We've come a long way from caves and flint axes, and it's only the fundamentalists and conservatives who insist what we have is where it should end.
Microsoft Office could pass a standard ODF-ACID test but still allow extensions. These won't be part of the standard, but if the document itself still qualifies for the standard, it will only work in Office anyway. Microsoft has entire divisions dedicated to this kind of evil.
Fair enough. I like H2 for pure Java projects, because although SQLite is my favorite in general, its Java bindings suck. H2 is pure Java (no bindings, no native components, so you can embed it anywhere) and performs very well. You can even run a client-server split if you find you need that, but if it comes to that I'd rather just use PostgreSQL and the JDBC driver for that (which, IIRC, is also pure Java).
I can think of a case where Ctrl+C could be a leak of information exploit... if you rely on your program to zero its memory before shutting down, Ctrl+C could kill it before it zeroes its memory and the sensitive information will be left over, subject to any of the usual memory exploits. If this is on a swap file you're even worse off.
All of that pre-installed crapware subsidises the cost of the machine. The more there is, the cheaper the machine gets.
Once you've bought it, you're welcome to install Linux on a clean slate, but unless you bought a separate XP CD you're probably going to get the crapware back every time you reinstall or recover.
I build my own machines too, and for laptops I don't even let them boot until after I've wiped the default install. I don't even want the initial "phone home" most crapware does on those installs. The less they gain from their practices, the better. Honest market research is one thing, but installing spies and advertisers on our machines is plain evil.
Just out of curiosity, why do you prefer HSQL over its successor, H2?
Because Wild Tangent is spyware.
A window manager is only a tiny part of a desktop environment.
This kind of fear and paranoia is exactly what Microsoft wants, and insituting this kind of witchunt is going to do much more damage than any malicious contributor possibly could.
Linux has always had a "bad" security record. It is not difficult for a motivated and skilled attacker to, at the very least, cause severe denial of service against a Linux host. Black hats have many more exploits than developers and auditors know about.
If Microsoft really wants to damage the Linux ecosystem, and legal means (patent threats, copyright threats, market confusion, etc) run out, the next logical resort will be to hire blackhats. They may already have some. Why would they tell anyone?
The recent Debian/Ubuntu vulnerability, which technically left me trivially vulnerable for two years, made me seriously consider switching back to FreeBSD. Fortunately my OpenVPN keys were generated specifically on BSD so they were not vulnerable. It's not like BSD hasn't had exploits either, but Ubuntu in particular has been an embarrasment from its first release, where the system administrator password was recorded in a world-readable log file.
Well, we do have that. GNOME is the beginner mode and KDE is the expert mode. XFce is somewhere in the middle.
Personally I prefer GNOME's design lately, even though KDE's technology is vastly better even in 3.5. Being based on Qt gives KDE a serious advantage, but it's a shame that doesn't always result in elegant designs.
Put down the crack pipe. .NET is based on the same bytecode + JIT design as Java, with years less maturity. Microsoft only created .NET in the first place because they were legally barred from "embrace, extend, extinguish"ing Java itself. .NET has its merits over Java (and vice versa), but speed is almost identical.
Back at the time, this meant Linux could serve static files over HTTP with a tiny tiny fraction of the resources required for Windows or even BSD to do the same. So for that little kernel option and configuration, you could save 90% of your static file server resources.
Now Linux and user-space software alike have been optimized to the point that khttpd's performance is barely better and sometimes worse, so it is no longer practical.
I used to be a userland purist, but let's face it, back when all we had were 133Mhz Pentiums, fast is good. Microkernels never took off for the same reasons - elegant and maintainable are great but sometimes you just need fast. Otherwise we'd have kernels written in Python.
And thus far, aliens are vaporware. When we finally find them, they'll probably have far fewer features and far more DRM than rumored.
Steve Jobs is Buddhist. He could believe he is a Buddha himself, and the funny thing is, even if he's wrong there's nothing wrong with that. It's not like there's an ISO Standard Buddha, and nobody would take his sutras seriously anyway.
That's true, Python list comprehensions do not support ordering. However, since Python is partly a functional language anyway, it doesn't matter:
.filter, .map, etc. can all generate code behind the scenes, which is only run when the result is requested as an iretable. ORMs do this, though I haven't used any Ruby ORMs personally.
a = sorted([comprehension of b])
Or specifying your own comparator:
a = sorted([comprehension of b], cmp = comparator or lambda)
It's even cleaner in Ruby which has nicer closures, but there it's OO instead of functional.
You already know that, but I don't suppose every Slashdot reader does.
And while you have a point that Python list comprehensions in particular do not support being translated to SQL, Ruby's OO version certainly do. The
Except that it affects all keys generated on Debian. Even if your client is not running Debian, any number of your servers may be, whether they're in your administrative domain or not. Keys may have been generated on Debian machines and then used on Windows servers. The companies you rely on to store your sensitive details may have already been compromised. How the hell can you say you're not affected? Everyone's potentially affected. That's kind of the problem with PKI in general, which is meant to be "solved" with strong entropy.
It's a pretty thin layer compared to an entire extra base library + bytecode runtime + OS abstraction.
PyGTK layers:
Your code (python)
PyGTK code (python)
Python runtime (C)
PyGTK->GTK binding (C)
GTK+libc code (C)
kernel (C)
IronPython + Gtk# layers:
Your code (python)
IronPython code (python)
IronPython runtime (CLI)
Gtk# code (CLI)
Mono base (CLI)
Mono runtime (C)
Gtk# -> GTK binding (C)
GTK+libc code (C)
kernel (C)
That's a fun one to deploy, let me tell you.
When you already have a dynamic language, type inference, anonymous types and extension methods are implicit. I'm not arguing every language should be dynamic, but the additions to C# are just solving problems specific to static languages. That's fine - static languages are still much faster than dynamic ones. Just saying, as far as languages go, C# is the one playing catchup, and with about a 20 year gap for many features.