I would have liked to have seen any areas where BSD and Linux beat Solaris, on Sun hardware.
Yes, but then you would be testing 2 different OSs on 2 completely different architectures (x86 and SPARC). Unless of course you mean we should test Linux/SPARC against Solaris/SPARC - but that would probably give the advantage to Sun, considering that Linux/SPARC is probably far less optimized than Linux/x86 (the inverse is true for Solaris).
Ummm... but GTK+ and Qt are built on top of X. You would still need an emulation layer for Xlib in any new windowing system.
I think what he means is that, if a new windowing system was developed, we could just port GTK+ and QT to the new windowing system (and keep the GTK / QT API the same) - that way most existing applications could continue to work, without the overhead of some kind of X emulation layer.
Besides, who's to say that an add-on package has to make remote-display slower?
I'm not an X hacker, but I'd guess that if you wrote a new window system that was not network transparent (which would make it simpler to design and implement, I suppose), it would involve MUCH more than an 'add-on package' to add network transparency (which I believe is done in the X protocol currently).
Rather than webmail (which kinda sucks - I'd much prefer mutt for email than Netscape, or even Mozilla), what about someone providing off-shore secure POP or IMAP4-SSL? Most good mail clients (including fetchmail and mutt) support IMAP + SSL.
Is Hushmail off-shore?
Also, I'm Canadian. How are Canada's laws on personal privacy (excuse my ignorance, Canada is pretty boring;D). If Canada is reasonably lax, I might look into setting up something myself...
I don't know if anyone has mentioned this (or if the GIMP people already plan to implement it), but I find a minor feature of Photoshop to be very useful. When you copy a section of an image to the clipboard, and go to 'new image', the default size of the new image is at exactly the same resolution/pixel size as the contents of the clipboard.
I bet it wouldn't take long at all to implement, but it makes a lot of tasks faster. I would love it if there was a way to do this in the GIMP already, but AFAIK there isn't.
So what you are saying (and apparently the moderaters agree with you)
When you're a Slashdot moderator, you don't moderate up posts that you agree with! You moderate up posts that you think the Slashdot community would find interesting to read.
Sealand could be a very important step in providing a place for groups with controversial material to host a site without fear of the government taking it down. From the HavenCo site:
HavenCo is donating free colocation space to Non-Governmental Organizations of our choosing. In general, the types of organizations that we will want to provide hosting for are those that promote
free speech
promote human rights
give a voice to minority and oppressed groups that otherwise may not be heard
Among other groups, Freenet have expressed interest in getting a large Freenet node hosted there (of course, that's very preliminary, and was just posted to freenet-chat a couple hours ago). Between Freenet and HavenCo, this is an exciting time for free speech and the Internet.
Linux also ignited the "open source" movement of shared software that has been partly adopted by companies from Netscape to Oracle
Netscape, sure, but what Open Source software has Oracle released? (I could just be uninformed, so feel free to enlighten me).
It's amazing how that article was able to allegedly cover the beginnings of the GNU/Linux movement, and *never* once mentioned the words "FSF", "Richard M. Stallman", or "GNU". The article makes it seem like Linus wrote the kernel, and all the rest of the GNU software kind of just 'appeared' out of thin air! I'm no FSF fanatic, but I think that RMS et al have a point. It's articles like this that make me think we should try to emphasize 'GNU/Linux' over just 'Linux'. And 'Free Software', rather than just 'Open Source'.
Linux also ignited the "open source" movement of shared software that has been partly adopted by companies from Netscape to Oracle
Netscape, sure, but what Open Source software has Oracle released? (I could just be uninformed, so feel free to enlighten me).
It's amazing how that article was able to allegedly cover the beginnings of the GNU/Linux movement, and *never* once mentioned the words "FSF", "Richard M. Stallman", or "GNU". The article makes it seem like Linus wrote the kernel, and all the rest of the GNU software kind of just 'appeared' out of thin air! I'm no FSF fanatic, but I think that RMS et al have a point. It's articles like this that make me think we should try to emphasize 'GNU/Linux' over just 'Linux'. And 'Free Software', rather than just 'Open Source'.
The other thing to keep in mind is that consoles like the PS2 use the hardware as the graphics API. Some people think this is crazy, but it's not a big deal. You take the 2% of your code that deals with actually drawing things and rewrite it to use PS2-specific features.
Your comments are interesting, but GreenMarine commented that it took Epic 6 months to port UT to the PS2. Assuming they had a fairly large team of coders working on it, it seems like 6 months is an awful long time to spend rewriting a 'mere' 2% of the total codebase.
I don't know anything about graphics coding, so maybe I'm wrong. What's the real verdict?
Did anyone else notice the section of the 'Technology' page that describes 'Gnome Basic'? Helix plan to copy Microsoft, by producing a clone of VB (syntax compatible), and embedding it in Gnumeric documents and other GNOME office stuff.
People! Haven't we just spent the last couple days making fun of M$ weenies who received hundreds of 'ILOVEYOU' emails? Did we really need, or want, embedded scripting in data files? What good will GB do to enhance Gnumeric or the rest of the GNOME office? Does it really justify the security risk of embedding logic in data files? I guess that's my main objection. I write most of my documents in plain text and / or TeX. A data file contains... data. That's it. It's not going to pop up any dialog boxes, query any databases, or do anything else that's completely unnecessary. It just contains my data, plus some minor formatting information (in the case of TeX / troff).
Of all the people I know who use Word (probably several hundred), not a single one has ever used VB macros in a document. And many of them have suffered through VB macro viruses. Can anyone think of anything that justifies the inclusion of a scripting language in Gnumeric data files? I know that the GB implementation will probably be more secure, but it seems like Helix want to copy VB 'feature' for feature, bug for bug.
Remember when the Rage 128 was announced? It was supposed to crush the TNT1 and Voodoo2 SLI, and (if it had shipped on time), it probably would have. But by the intended shipping date (Christmas 99, IIRC), they only had early alpha silicon which ran far hotter than the specs had claimed. The final Rage 128 shipped more than 3 months late, at a slower speed than was originally announced. Several weeks later, the V3 / TNT2 came out, blowing the Rage 128 out of the water.
My memory is a bit fuzzy, so if someone could correct me, I'd appreciate it.
AFAIK, pentium optimizations have been reject by the Debian developers for several reasons.
- they rarely enhance performance at all - pgcc has problems compiling quite a lot of software - any good optimizations that pgcc developes are eventually folded back into gcc/egcs - Debian runs on many different platforms, and spending time producing 2 x86 distrox probably isn't a very good way to use Debian's resources
Fairly recently, nVidia worked extensively with Microsoft on the X-Box. Since then, nVidia's support for their Linux drivers has evaporated. XF86 4.0 has been released, and there is no sign of the high performance drivers that were promised. And now this - is it at least possible that nVidia could be influenced by Microsoft and interested in keeing gaming essential Windows only?
Just a kind of 'off the wall' rumour I heard. Do you think there is any truth to it?
He means the Shockwave (Flash?) web movies that come bundled with many Macromedia products, including Dreamweaver and Flash.
There is text, and when you click the 'start' button, the cursur on the screenshot does something, and you see what happens - i.e. if it is a documentation page on writing 'helloworld.c', it would show the user opening the IDE, starting a project / loading a file / whatever, typing in the code, going to the 'compile' IDE option, and executing the executable (for arguments sake, of course).
It's a pretty good idea for certain applications. For 99% of traditional UNIX stuff, it's dumb (err, how is that useful if I want to configure Apache?). But for stuff like GUI design tools (the GIMP, video editors, etc), it could be pretty neat (for newbies at least).
Macromedia bundles the movies with the product (as a direct part of the documentation), so you don't have to download them.
I took the Brainbench "Linux Administrator" test, and it's *really* easy. Especially if you refer to manpages + online docs, which I didn't do.
IIRC, I got ~3.24 - either that or 3.74, out of 5. Certainly not amazing - most/. could pass it easily, I would expect. It's pretty decent though - they mail you an official looking certification (I don't think it even mentions you took the exam online), and it's something to put on your resume that ignorant HR people won't know to distrust.
It's a cool concept (free online certifications) - I would love for Redhat, LPI, or a similar authority to offer some kind of online exam. Of course, there are numerous problem - how would the exam be 'verified' - you could easily get the answers from someone who took the exam earlier. If you force the applicants to take the exam at the same time, you destroy the core advantages of the online exam contest. IIRC, the Brainbench exam has a database of questions, and only asks ~60 for each exam. This could be a good approach, but would of course involve writing more, quality questions. The questions would be divided into topics to ensure each test taker would get an approximately equal amount of questions on each topic. What are the other security problems with this concept? I'm probably missing something obvious.
I don't know anything about the RH certification, but has anyone done the LPI courses? I looked at the topics on their website, and they looked pretty in-depth.
Also, how much would it cost to take the LPI exam? Is the RH exam a better idea?
The article mentions that GNU/Linux isn't innovative (although it doesn't say MS is either). I've heard this argument before, usually as a criticism of Linux.
Would you agree with this? IMHO, Linux is very innovative - how else could you account for its success? GNOME, KDE, Enlightenment, Apache + JServ, etc. Even the Open Source concept (while not directly accreditable to Linux), is different from traditional closed source idealogy.
Yeah - make that x86 + alpha. I personally don't see any more BSD mergers coming. BSDI never had much a 'niche' IMHO - I still don't understand the appeal of "corporate accountibility". Most people have figured out that using an open source product, in general, provides better accountibility. FreeBSD definately has its niche, as does OpenBSD. NetBSD? Sure, why not (NetBSD could be cool for the Playstation 2!)
I have used FreeBSD in the past, and really enjoyed it. However, I've never used BSDi. What do you people think of it? Is it any good?
What advantages does the BSDi codebase have over the existing FreeBSD code? What's likely to be incorporated into FreeBSD? Other than more coders, is anything else going to change?
Here's a quote from the article:
"Once BSDI releases the code to FreeBSD, it wall fall under a very liberal license. Basically, if the code is incorporated into an existing open source project, it will fall under the licensing terms of that project. That means that any open source project can incorporate BSDI's code..."
If the code can be re-licensed for use by ANY open source project, would that include Linux? For those kernel hackers, is there anything in BSDI that you would like to see included in the Linux kernel, perhaps for 3.0?
Well, I don't know either, but I think this quote from the article indicates that SRP *does* offer session encryption, not just secure password transmission.
3.SRP exchanges a session key in the process of authentication. This key can be used to encrypt the user's login session and protect it from both snooping and malicious active attack.
ssh does a similar thing - a session key is changed and can be used for encrypting the user's login session, etc.
FYI, I've heard that the XFree86 3.3.6 debs for stable (and others in the series) are quite unstable and buggy (and this is from the official Debian X maintainer, Branden Robinson). I've never used them - but I figured out might want to know.
Also, why not use unstable? It can be a bit annoying sometimes (when something breaks, and fscks with dpkg), but you get access to up to the second software, the latest packages and dpkg enhancements, and a chance to help out Debian (by reporting any bugs you find in unstable). I've used it for quite a while now, without any major problems.
Slashdot Mirror
Yes, but then you would be testing 2 different OSs on 2 completely different architectures (x86 and SPARC). Unless of course you mean we should test Linux/SPARC against Solaris/SPARC - but that would probably give the advantage to Sun, considering that Linux/SPARC is probably far less optimized than Linux/x86 (the inverse is true for Solaris).
I think what he means is that, if a new windowing system was developed, we could just port GTK+ and QT to the new windowing system (and keep the GTK / QT API the same) - that way most existing applications could continue to work, without the overhead of some kind of X emulation layer.
I'm not an X hacker, but I'd guess that if you wrote a new window system that was not network transparent (which would make it simpler to design and implement, I suppose), it would involve MUCH more than an 'add-on package' to add network transparency (which I believe is done in the X protocol currently).
But correct me if I'm wrong...
Is Hushmail off-shore?
Also, I'm Canadian. How are Canada's laws on personal privacy (excuse my ignorance, Canada is pretty boring ;D). If Canada is reasonably lax, I might look into setting up something myself...
Oh! My bad! The last time I looked for this was in 1.0.x branch. 1.1.23 has it.
Thanks for the heads up.
I bet it wouldn't take long at all to implement, but it makes a lot of tasks faster. I would love it if there was a way to do this in the GIMP already, but AFAIK there isn't.
When you're a Slashdot moderator, you don't moderate up posts that you agree with! You moderate up posts that you think the Slashdot community would find interesting to read.
HavenCo is donating free colocation space to Non-Governmental Organizations of our choosing. In general, the types of organizations that we will want to provide hosting for are those that promote
Among other groups, Freenet have expressed interest in getting a large Freenet node hosted there (of course, that's very preliminary, and was just posted to freenet-chat a couple hours ago). Between Freenet and HavenCo, this is an exciting time for free speech and the Internet.
Netscape, sure, but what Open Source software has Oracle released? (I could just be uninformed, so feel free to enlighten me).
It's amazing how that article was able to allegedly cover the beginnings of the GNU/Linux movement, and *never* once mentioned the words "FSF", "Richard M. Stallman", or "GNU". The article makes it seem like Linus wrote the kernel, and all the rest of the GNU software kind of just 'appeared' out of thin air! I'm no FSF fanatic, but I think that RMS et al have a point. It's articles like this that make me think we should try to emphasize 'GNU/Linux' over just 'Linux'. And 'Free Software', rather than just 'Open Source'.
Netscape, sure, but what Open Source software has Oracle released? (I could just be uninformed, so feel free to enlighten me).
It's amazing how that article was able to allegedly cover the beginnings of the GNU/Linux movement, and *never* once mentioned the words "FSF", "Richard M. Stallman", or "GNU". The article makes it seem like Linus wrote the kernel, and all the rest of the GNU software kind of just 'appeared' out of thin air! I'm no FSF fanatic, but I think that RMS et al have a point. It's articles like this that make me think we should try to emphasize 'GNU/Linux' over just 'Linux'. And 'Free Software', rather than just 'Open Source'.
Your comments are interesting, but GreenMarine commented that it took Epic 6 months to port UT to the PS2. Assuming they had a fairly large team of coders working on it, it seems like 6 months is an awful long time to spend rewriting a 'mere' 2% of the total codebase.
I don't know anything about graphics coding, so maybe I'm wrong. What's the real verdict?
People! Haven't we just spent the last couple days making fun of M$ weenies who received hundreds of 'ILOVEYOU' emails? Did we really need, or want, embedded scripting in data files? What good will GB do to enhance Gnumeric or the rest of the GNOME office? Does it really justify the security risk of embedding logic in data files? I guess that's my main objection. I write most of my documents in plain text and / or TeX. A data file contains... data. That's it. It's not going to pop up any dialog boxes, query any databases, or do anything else that's completely unnecessary. It just contains my data, plus some minor formatting information (in the case of TeX / troff).
Of all the people I know who use Word (probably several hundred), not a single one has ever used VB macros in a document. And many of them have suffered through VB macro viruses. Can anyone think of anything that justifies the inclusion of a scripting language in Gnumeric data files? I know that the GB implementation will probably be more secure, but it seems like Helix want to copy VB 'feature' for feature, bug for bug.
My memory is a bit fuzzy, so if someone could correct me, I'd appreciate it.
For more info, check Debian.org.
For more info, check Debian.org.
- they rarely enhance performance at all - pgcc has problems compiling quite a lot of software - any good optimizations that pgcc developes are eventually folded back into gcc/egcs - Debian runs on many different platforms, and spending time producing 2 x86 distrox probably isn't a very good way to use Debian's resources
Check the debian-devel archives for more info.
Just a kind of 'off the wall' rumour I heard. Do you think there is any truth to it?
There is text, and when you click the 'start' button, the cursur on the screenshot does something, and you see what happens - i.e. if it is a documentation page on writing 'helloworld.c', it would show the user opening the IDE, starting a project / loading a file / whatever, typing in the code, going to the 'compile' IDE option, and executing the executable (for arguments sake, of course).
It's a pretty good idea for certain applications. For 99% of traditional UNIX stuff, it's dumb (err, how is that useful if I want to configure Apache?). But for stuff like GUI design tools (the GIMP, video editors, etc), it could be pretty neat (for newbies at least).
Macromedia bundles the movies with the product (as a direct part of the documentation), so you don't have to download them.
IIRC, I got ~3.24 - either that or 3.74, out of 5. Certainly not amazing - most /. could pass it easily, I would expect. It's pretty decent though - they mail you an official looking certification (I don't think it even mentions you took the exam online), and it's something to put on your resume that ignorant HR people won't know to distrust.
It's a cool concept (free online certifications) - I would love for Redhat, LPI, or a similar authority to offer some kind of online exam. Of course, there are numerous problem - how would the exam be 'verified' - you could easily get the answers from someone who took the exam earlier. If you force the applicants to take the exam at the same time, you destroy the core advantages of the online exam contest. IIRC, the Brainbench exam has a database of questions, and only asks ~60 for each exam. This could be a good approach, but would of course involve writing more, quality questions. The questions would be divided into topics to ensure each test taker would get an approximately equal amount of questions on each topic. What are the other security problems with this concept? I'm probably missing something obvious.
Also, how much would it cost to take the LPI exam? Is the RH exam a better idea?
Would you agree with this? IMHO, Linux is very innovative - how else could you account for its success? GNOME, KDE, Enlightenment, Apache + JServ, etc. Even the Open Source concept (while not directly accreditable to Linux), is different from traditional closed source idealogy.
Yeah - make that x86 + alpha. I personally don't see any more BSD mergers coming. BSDI never had much a 'niche' IMHO - I still don't understand the appeal of "corporate accountibility". Most people have figured out that using an open source product, in general, provides better accountibility. FreeBSD definately has its niche, as does OpenBSD. NetBSD? Sure, why not (NetBSD could be cool for the Playstation 2!)
What advantages does the BSDi codebase have over the existing FreeBSD code? What's likely to be incorporated into FreeBSD? Other than more coders, is anything else going to change?
Here's a quote from the article:
"Once BSDI releases the code to FreeBSD, it wall fall under a very liberal license. Basically, if the code is incorporated into an existing open source project, it will fall under the licensing terms of that project. That means that any open source project can incorporate BSDI's code..."
If the code can be re-licensed for use by ANY open source project, would that include Linux? For those kernel hackers, is there anything in BSDI that you would like to see included in the Linux kernel, perhaps for 3.0?
3.SRP exchanges a session key in the process of authentication. This key can be used to encrypt the user's login session and protect it from both snooping and malicious active attack.
ssh does a similar thing - a session key is changed and can be used for encrypting the user's login session, etc.
Also, why not use unstable? It can be a bit annoying sometimes (when something breaks, and fscks with dpkg), but you get access to up to the second software, the latest packages and dpkg enhancements, and a chance to help out Debian (by reporting any bugs you find in unstable). I've used it for quite a while now, without any major problems.