I had heard somewhere that the bytecode format would remain the same, so forward compatibility of compiled Java would work because the new language features would only be visible to the compiler.
I thought that the compiled Java would remain compatible with the bytecode format used by previous versions. However, this seems not be the case and I get this message:
java.lang.UnsupportedClassVersionError: HelloWorld (Unsupported major.minor version 49.0)
Whilst code that uses the new language features must obviously be compiled with the v1.5 JSDK, this means that it must also be run on the v1.5 JRE.
This may inhibit the use of Java 5 by projects that want their programs to run on a v1.4 JRE.
About six months ago I wanted to try out Gentoo, so I installed it on an old PC I had lying around: a 200Mhz AMD processor, 96MB ram, no harddrive. I used nfs for the root fs, and the compiling needed swap-over-nbd (network block device).
I think this could be a good thing; Schwartz' jealousy of IBM's reputation could bring out some competitiveness.
The Open Source community will benefit greatly from two companies vying to outdo each other in a quest for our affections by seeing who can contribute the most.
My next stop on the Linux train will be Gentoo, if it comes to that.
IMHO you should try this anyway. I'd been using Mandrake for years when I tried Gentoo earlier this year, and it isn't hard to use for people that know what they're doing on Mandrake.
I have seen benefits that I didn't appreciate before I tried it, such as the design of the Portage system which means that each Gentoo install contains all of the build information for the whole distribution, and is simple to extend yourself.
Actually, it would be good to see a distribution like Mandrake switch to using Portage because it makes installation of updates (for which there aren't binary packages) so smooth. Portage is a better design than RPM, which Mandrake currently use for historic reasons. They could maintain their own repository and distribute using binary packages on CDs as they do now.
FCC regulations prohibit open source firmware in 802.11
Is this really true? This comment seems to suggest otherwise:
Every one of these 802.11b and 802.11a wireless networking cards that I've pried the case from includes a hardware bandpass filter. This is true even for the manufacturers who refused (at first) to repease drivers or specs.
Though CSS is the ideal choice for logical placement of objects, tables and fixed widths to position material still reign.
Using CSS for layout can be a pain. It is, for instance, difficult to construct a page that has a banner across the bottom of the browser window, which will instead appear below the rest of the content if that doesn't fit in the window. It can probably be done with min-height, but IE doesn't have that.
Most CSS layouts I've seen horribly break if the font size is increased by 300%.
The old./configure prefix=, make && make install is too archaic and relies on a good understanding of directory structure
Gentoo's approach to this is brilliant. It would be great if other distributions were made using (and supplied with) Portage, and then they could still be distributed as binary packages if necessary. Upgrading or installing a custom version of a package then becomes very easy.
I've recently done this with a small webserver to keep it running. Some sort of deposit had appeared on top of the electrolytic capacitors on the motherboard, and the machine became unbearably unstable.
I took out one of it's mirrored drives and connected it in a different (larger) machine and then booted it using Usermode Linux. I found it was best to be running 2.6.7 on both the host and the uml and it is bridged onto the host's network, so it appears exactly as before.
The article is biased because it, seemingly deliberately, omits crucial parts of the discussion. For instance: 2. "Open Source software allows you to get under the hood and fix problems"
The author says that the idea that OSS allows you to tinker with the source code is a "myth". He is totally missing the point; The freedom to fix the software is important, not because every user will be able to do so, but because they will all ultimately benefit from this access being available to the programmers that will submit patches.
I'm familiar with the use of relative units, but unfortunately they don't solve the problem. Because of a lack of guidance from the W3C, different browsers measure some of these sizes differently. To see the problem, try to find a page with a reasonably complex CSS layout that scales well as the font size is increased above 300% (CTRL-+ in Moz).
Does anyone know if this revision will specify more precisely how it should be displayed by a web browser?
The problem I've found with fully using XHTML+CSS for web pages is that it is not possible to layout pages that will scale accurately as the font size is increased. So much for accessibility!
I wonder why it was decided that it would be useful for text that doesn't fit, to run-over other text and elements on the same page. It would be better if we could tell the browser that when elements expand the other parts of the page must move out of the way.
I don't see where in your quote that the original author suggested that GPL code couldn't be used in commercial products.
In the second paragraph, he is contrasting the PHP license to the GPL and he says: "They are just happy that it's a PHP license and they can... ship it with their commercial products"
His entire comment shows his huge personal anti-GPL bias which is exhibited in the labelling and falsehoods therein. I think you need to question people's motives more carefully.
I didn't claim that the code has "rights". The grandparent post (by dmaxwell) is exactly correct. The benefit of concentrating on keeping the code Free is for its users.
Andi Gutmans seems to have considered his comments for very little time, or is intentionally choosing to label a diverse group of people with FUD like this:
"The GNU community, in my opinion, is a very fanatic community and I don't think it represents the real serious open source users. It definitely doesn't represent the PHP user base," Gutmans said.
"Most of PHP's user base are people that are using PHP to make a living and they wouldn't care less. "They are just happy that it's a PHP license and they can do whatever they want with it and can ship it with their commercial products," he said.
It seems that almost everyone that has an alterior motive for disliking the GPL chooses to hide behind this incorrect criticism. GPL'ed code *can* be used in commercial products, and the weird thing about hearing this crap from Gutmans is that PHP is already Open Source.
the FSF wants software to be free - but part of being free is having the right to develop non-free software.
No it isn't. The freedom that the FSF want to safeguard is that of the code, not the developer. You can't maintain the freedom of the code by using it in non-free software.
Maven's 'repository', under Linux (and with the default properties), is under ~/.maven/repository/
You're right - my memory had become slight hazy about this. It was the per-user package repositories that I was remembering as a potential problem.
What it needs now is source compilation... [I know that the Java bytecode won't vary much with recompilation, but it is using the "preferred source format" that I like]
I've only used Maven in the way that it has been part of projects that I've downloaded. Something that gives me reservations is the way that it manages library dependencies by downloading them into the project tree. On a Unix system this feels inconsistent because I prefer to manage installed software (libraries) on a system-wide basis.
It would be nice if there were a consensus on Java library dependency resolution/installation. I used to use the exellent jpackage.org on Mandrake, but now I appreciate the source compilation aspect of Gentoo (although many of Gentoo's Java packages are old). A combination of the two would be ideal.
And whilst I'm forming a wishlist, I'd also like Java applications to have greater integration with the system, not on the desktop, but in the way that multiple invocations might have less of a memory penalty (like a Java subsystem).
If only CSS1&2 were actually implemented in a standard way across the different browsers. Whilst applying styles does simplify writing webpages, using them for layout produces irritatingly haphazard results on different browsers, and not a single one scales in a sensible way as the font-size is increased (i.e. any CSS-layout page with the size increased by 5x will usually display horribly).
On that subject, I'd like to make a wireless router for my parents' network. I don't want an off the shelf one, so that it can contain a harddrive and I can use it remotely.
I could use a case large enough to fit a PCI wireless card, or I've noticed that some of VIA's boards have a Cardbus slot, although I can't help thinking that it isn't a good idea to have the 802.11b/g card inside the metal case.
I'm aware of the nice Soekris boards, but they are much more expensive that VIA's. It's just a shame VIA don't offer a miniPCI slot.
I'd like to know what other people who've done this used.
Similarly, author's disgust at some people using tabs to display separate pages seems ridiculous...
Indeed. In Konqueror you can have a good combination of both approaches. If you have it set to "open links in a new tab" then a middle-click will open a directory in a new tab, and files can be moved between tabs by mousing over the tab for a second whilst dragging which switches to that tab's view.
This is why it's great to have one high-powered machine that has some X thin clients attached. That way I almost always get the speed of the server, and I only have to maintain the desktop software for one machine (on the server).
Slashdot Mirror
(forward compatibility is somewhat impossible :p)
I had heard somewhere that the bytecode format would remain the same, so forward compatibility of compiled Java would work because the new language features would only be visible to the compiler.
- Brian.
Whilst code that uses the new language features must obviously be compiled with the v1.5 JSDK, this means that it must also be run on the v1.5 JRE.
This may inhibit the use of Java 5 by projects that want their programs to run on a v1.4 JRE.
- Brian.
About six months ago I wanted to try out Gentoo, so I installed it on an old PC I had lying around: a 200Mhz AMD processor, 96MB ram, no harddrive.
I used nfs for the root fs, and the compiling needed swap-over-nbd (network block device).
It took a week to compile KDE.
- Brian.
I think this could be a good thing; Schwartz' jealousy of IBM's reputation could bring out some competitiveness.
The Open Source community will benefit greatly from two companies vying to outdo each other in a quest for our affections by seeing who can contribute the most.
- Brian.
My next stop on the Linux train will be Gentoo, if it comes to that.
IMHO you should try this anyway. I'd been using Mandrake for years when I tried Gentoo earlier this year, and it isn't hard to use for people that know what they're doing on Mandrake.
I have seen benefits that I didn't appreciate before I tried it, such as the design of the Portage system which means that each Gentoo install contains all of the build information for the whole distribution, and is simple to extend yourself.
Actually, it would be good to see a distribution like Mandrake switch to using Portage because it makes installation of updates (for which there aren't binary packages) so smooth.
Portage is a better design than RPM, which Mandrake currently use for historic reasons. They could maintain their own repository and distribute using binary packages on CDs as they do now.
- Brian.
Is this really true? This comment seems to suggest otherwise:
- Brian
Though CSS is the ideal choice for logical placement of objects, tables and fixed widths to position material still reign.
Using CSS for layout can be a pain. It is, for instance, difficult to construct a page that has a banner across the bottom of the browser window, which will instead appear below the rest of the content if that doesn't fit in the window.
It can probably be done with min-height, but IE doesn't have that.
Most CSS layouts I've seen horribly break if the font size is increased by 300%.
- Brian.
The old ./configure prefix=, make && make install is too archaic and relies on a good understanding of directory structure
Gentoo's approach to this is brilliant. It would be great if other distributions were made using (and supplied with) Portage, and then they could still be distributed as binary packages if necessary.
Upgrading or installing a custom version of a package then becomes very easy.
- Brian.
I've recently done this with a small webserver to keep it running. Some sort of deposit had appeared on top of the electrolytic capacitors on the motherboard, and the machine became unbearably unstable.
I took out one of it's mirrored drives and connected it in a different (larger) machine and then booted it using Usermode Linux.
I found it was best to be running 2.6.7 on both the host and the uml and it is bridged onto the host's network, so it appears exactly as before.
- Brian.
The article is biased because it, seemingly deliberately, omits crucial parts of the discussion. For instance:
2. "Open Source software allows you to get under the hood and fix problems"
The author says that the idea that OSS allows you to tinker with the source code is a "myth". He is totally missing the point; The freedom to fix the software is important, not because every user will be able to do so, but because they will all ultimately benefit from this access being available to the programmers that will submit patches.
- Brian.
I'm familiar with the use of relative units, but unfortunately they don't solve the problem.
Because of a lack of guidance from the W3C, different browsers measure some of these sizes differently.
To see the problem, try to find a page with a reasonably complex CSS layout that scales well as the font size is increased above 300% (CTRL-+ in Moz).
- Brian.
Does anyone know if this revision will specify more precisely how it should be displayed by a web browser?
The problem I've found with fully using XHTML+CSS for web pages is that it is not possible to layout pages that will scale accurately as the font size is increased. So much for accessibility!
I wonder why it was decided that it would be useful for text that doesn't fit, to run-over other text and elements on the same page.
It would be better if we could tell the browser that when elements expand the other parts of the page must move out of the way.
- Brian.
I don't see where in your quote that the original author suggested that GPL code couldn't be used in commercial products.
... ship it with their commercial products"
In the second paragraph, he is contrasting the PHP license to the GPL and he says:
"They are just happy that it's a PHP license and they can
His entire comment shows his huge personal anti-GPL bias which is exhibited in the labelling and falsehoods therein. I think you need to question people's motives more carefully.
- Brian.
I didn't claim that the code has "rights". The grandparent post (by dmaxwell) is exactly correct. The benefit of concentrating on keeping the code Free is for its users.
- Brian.
Andi Gutmans seems to have considered his comments for very little time, or is intentionally choosing to label a diverse group of people with FUD like this:
"The GNU community, in my opinion, is a very fanatic community and I don't think it represents the real serious open source users. It definitely doesn't represent the PHP user base," Gutmans said.
"Most of PHP's user base are people that are using PHP to make a living and they wouldn't care less. "They are just happy that it's a PHP license and they can do whatever they want with it and can ship it with their commercial products," he said.
It seems that almost everyone that has an alterior motive for disliking the GPL chooses to hide behind this incorrect criticism. GPL'ed code *can* be used in commercial products, and the weird thing about hearing this crap from Gutmans is that PHP is already Open Source.
- Brian.
the FSF wants software to be free - but part of being free is having the right to develop non-free software.
No it isn't. The freedom that the FSF want to safeguard is that of the code, not the developer. You can't maintain the freedom of the code by using it in non-free software.
- Brian.
Maven's 'repository', under Linux (and with the default properties), is under ~/.maven/repository/
...
You're right - my memory had become slight hazy about this. It was the per-user package repositories that I was remembering as a potential problem.
What it needs now is source compilation
[I know that the Java bytecode won't vary much with recompilation, but it is using the "preferred source format" that I like]
- Brian
I've only used Maven in the way that it has been part of projects that I've downloaded.
Something that gives me reservations is the way that it manages library dependencies by downloading them into the project tree. On a Unix system this feels inconsistent because I prefer to manage installed software (libraries) on a system-wide basis.
It would be nice if there were a consensus on Java library dependency resolution/installation. I used to use the exellent jpackage.org on Mandrake, but now I appreciate the source compilation aspect of Gentoo (although many of Gentoo's Java packages are old). A combination of the two would be ideal.
And whilst I'm forming a wishlist, I'd also like Java applications to have greater integration with the system, not on the desktop, but in the way that multiple invocations might have less of a memory penalty (like a Java subsystem).
- Brian.
It seemed obvious that something was wrong with this story the first time it was discussed on /. because there were no links to Dell.
u estar/
The Register confirms that this move has little to do with Dell:
http://www.theregister.co.uk/2004/07/07/dell_vs_q
- Brian.
> the CSS3 standard
If only CSS1&2 were actually implemented in a standard way across the different browsers.
Whilst applying styles does simplify writing webpages, using them for layout produces irritatingly haphazard results on different browsers, and not a single one scales in a sensible way as the font-size is increased (i.e. any CSS-layout page with the size increased by 5x will usually display horribly).
- Brian.
I've seen the problem on Win98 and 2000, where the graphics card drivers are dodgy.
You can see the same effect by running IE using Wine.
- Brian.
Unfortunately, what all of these PNG-on-IE sites fail to mention is that they only work if the user has a working DirectX setup.
Worse still, without DirectX the images don't display at all, which is worse than not using the hack at all.
- Brian.
On that subject, I'd like to make a wireless router for my parents' network. I don't want an off the shelf one, so that it can contain a harddrive and I can use it remotely.
I could use a case large enough to fit a PCI wireless card, or I've noticed that some of VIA's boards have a Cardbus slot, although I can't help thinking that it isn't a good idea to have the 802.11b/g card inside the metal case.
I'm aware of the nice Soekris boards, but they are much more expensive that VIA's. It's just a shame VIA don't offer a miniPCI slot.
I'd like to know what other people who've done this used.
- Brian.
Similarly, author's disgust at some people using tabs to display separate pages seems ridiculous ...
Indeed. In Konqueror you can have a good combination of both approaches. If you have it set to "open links in a new tab" then a middle-click will open a directory in a new tab, and files can be moved between tabs by mousing over the tab for a second whilst dragging which switches to that tab's view.
- Brian.
This is why it's great to have one high-powered machine that has some X thin clients attached.
That way I almost always get the speed of the server, and I only have to maintain the desktop software for one machine (on the server).
- Brian