On x86, I prefer to run Debian. I'm considering getting a new Titanium G4 Powerbook, and I'd like to dual-boot OSX and a Linux distro. Has anyone tried Debian/PPC on the new Powerbook? I'd be very interested to hear your comments (positive or negative).
We fight amongst ourselves so much that we can't present a unified front against (much more organized) Closed Source efforts.
You mean like the "unified" database market -- Oracle versus DB2 versus Sybase versus Cloudscape versus MS SQL Server? Or the "unified" enterprise OS market -- Solaris versus HPUX versus AIX versus Linux/BSD versus NT? Or the "unified" web application market (too many products to list)?
Do we now judge the merits of an OS based on how it was implemented? Rather than, say, what it does and how reliably it does it?
Partially, yes. Obviously, AtheOS can't compare to Linux, BSD and even Win2k in terms of "what it does and how reliably it does it". But then again, neither could Linux in 1992. Perhaps the most important quality of any new OS is how easy it is to develop and extend and the quality of the core system design. I haven't looked at the source yet, but it seems to score very well on both those fronts. It's done in C++, which I personally find much more pleasing and useful than C. It has a very modular design (microkernel based?), allowing new components to be added without core modifications.
One impressive factor is SMP support: already! OpenBSD still doesn't support SMP and NetBSD just added it recently (not a flame of the BSD folks -- obviously, they're focusing on different goals. But I still think it's an impressive feature to have this early in its development).
Does the government have a legal monopoly on force?
Yes -- for example, if I shoot someone, that's illegal. If the government subsequently locks me up for 20 years (which is a form of physical force), that's perfectly legal.
I thought the raison d'etre of the 2nd Amendment was to prevent the government gaining a monopoly of force.
Right -- it's more complicated than that. As a citizen, you have the right to defend yourself -- and if necessary, to respond to an act of force with violence; but only the government can legally initiate the use of force. (That's Ayn Rand's view anyway...)
The government cannot take down Microsoft, but Microsoft can take down the government...
The government chose not to 'take down' Microsoft, dumb ass. Since the government has a legal monopoly of force, of course they are able to, if they choose to (whether they should be allowed to interfere in the economy is another issue entirely).
And also, Microsoft didn't 'take down' the US government -- not even intentionally. If anything, this attack will take the whitehouse website -- which is butt-ugly and serves zero official function -- down for a few hours. Big deal...
Take Netscape staying at 4.7 forever, then announcing their next version would be version 6, with version 5 never having been available.
It was. There was never a commercially available project, but Netscape released the code to the Mozilla project when it was started. The reason the next Netscape browser was called 6 was because it shared virtually no code with the Classic codebase (i.e. 4.x and 5-pre-pre-alpha).
Okay, let's discount this community - we're not what you might consider 'normal'.
This community, and the community of warez/gamerz/hackerz. It's "free" software, either way:-)
They would probably use it to backup something
If that was the case, few people would need a CD rewriter. I'd say people typically use them to burn music CDs (my family wants me to burn CDs for them all the time, as well as MP3s), portable documents (i.e. burn this presentation, take it in to work -- more reliable and better storage than floppy), or illegal software (even computer illiterates don't like paying for MS products:-)). The number of non-technies who back up anything is quite small, in any case.
I would definately like to get my hands on quad density CDs.
Take it easy chief... I read your post again, and I don't think I misinterpreted anything. Would you be so kind as to point out what else I should have read in your original post?
3dfx management may be trying to benefit themselves at the expense of their own shareholder who are the true owners of the company, and that we believe there are ways to dispose of 3dfx's assets that would return more of it's value to us.
I wasn't talking about that. I was talking about the aspect of your post when you talked about nVidia 'screwing over 3dfx investors'. You actually didn't talk about 3dfx's management at all in your original post.
It looks like nVidia may not actually be "buying" 3dfx. Rather, it looks like they will be cherry picking the few assets worth anything, like the designs for 3dfx's next products [...] and leaving 3dfx as a hollow shell containing nothing but a near-worthless boardmaking plant and lots of debt--in other words, completely screwing over the shareholders in the worst way possible.
You sound bitter - why's that? nVidia should be concerned with one thing: profit. They should take steps to make as much money as they can off the video card industry. If they judge that the best way to do this is picking the bones of 3dfx's corpse, so be it. Really, what do you think? That nVidia should be concerned about the financial well-being of 3dfx investors? What kind of fscked up society would that be!
Its really a shame with all the problems that 3dfx has had that they couldn't pull it out of the gutter...they started out great, and made the best products back in the day
A shame? 3dfx, like any other company, are in the business of making profit -- in any way they can. If they are unable to do this, they shouldn't exist. That's integral to the concept of a capitalist economy. I've never understood people having "pity" for a failing company. If it's failing, there is almost always one primary reason: the company is incompetant! If 3dfx were a good enough company to make a profit, they would have continued; obviously, they were not.
Note: I'm not anti-3dfx at all (I still own a couple V2s) -- but I just think sometimes people need reminding we do live in a capitalism.
Just imagine if Web sites could make ready-made WebSQL scripts for their portals for users to download and use with their favorite spreadsheet or database.
You mean exactly like RSS? It's XML based and numerous web sites (including Slashdot and Freshmeat) use it. That's how Slashdot's Slashboxes work. You could set it up like so: user creates an account @ foo.com. The user creates a portfolio (say, when buying stocks, or to get news updates on those companies). The website gives you an option of getting automatically regenerated RSS summaries of your portfolio and the latest prices (as well as RSS of market indices, currency/commodities, etc). You can use any scripting language with an XML parser (i.e. basically any) to process the XML and present it to the user. And all thjs without needing to buy a $499 software product, or use 6GB of local disk space.
From what I've heard, WebQL doesn't sound that innovative (if it even works at all). But since there was no technical info in the article, it's hard to tell.
People who bitch about UIs are becoming really annoying. I really hate the Mac UI. Seriously. Whenever I use a Mac, the experience is painful. Now that said, I understand how *certain types of users* find the Mac's interface intuitive. Personally, I'm fine with vi (I actually prefer Blackbox, but you get the point). Different UIs for different types of people. My ideal UI would probably be Blackbox + powershell (multiple xterms - rock!).
The idiocy of ranking UIs is that - what users are we talking about? You, a "person who owns multiple computers"? Those aren't exactly very impressive qualifications. Basically, you're ranking the UIs on your personal reaction to them. Fine - but what makes you think this applies to anyone else?
Window's UI is hardly useable for daily work
That's absurd. Tens of millions of people use Windows every day, at work and at home. You may not like it, but it's the truth.
Sorry for the rant. I just find it aggravating to hear Mac OS people say that no OS has ever approached the UI of the original Mac OS. As I've tried to show, that idea is nonsensical.
Well, it's being developed by the same company that pays for the time of most of the Postgres core team. From what I understand, most of the code will be initially open sourced, but some of it (the really enterprise level stuff I guess) will need to be purchased. Once that has been out for a while, that is also going to be open sourced, and included with the main Postgres release.
I'm happy we're getting replication at all. The companies that really need it won't mind paying for it, and the money goes into further PgSQL development. And (AFAIK) it's all going to be open source eventually.
I was thinking more like rsyncing pg_dump files. Considering that the vast majority of the output will be the same between synchronizations, this would be much more efficient than using scp - if you need security, you can run rsync over ssh.
You're right, of course. rsyncing the actual database files while the DB serving is running is incredibly dumb.
There's a project to add replication to PostgreSQL. From what I've heard, a public beta should be out fairly soon. It's at www.erserver.com.
If you'd rather do it the 'old fashioned' way, what is the query load like? If the data is mostly static, you could send all the writes to 1 database and use all the DBs for reads. Every 2 or so hours, update the read-only databases. Something like rsync would be very good for that. Until erServer is released, this may be your best choice.
There's been some discussion about this on the PostgreSQL mailing lists - check the pgsql-general archives @ postgresql.org
Seems like it would be very easy for someone to crack into the local client S/W and figure out to breach security on a Peer's machine by sending scripts/etc.
No more easy than hacking into a web server and using the same crack to break into hundreds of other sites running the same software. How is this any different from traditional client/server stuff? Besides, a badly written piece of software will tend to have security holes. It's up to the author to code carefully, and it's up to us as consumers to only use proven secure software (obligitary OpenBSD plug here).
Infected clients could update S/W from a site other than the one intended by the end-user (and thus infect more computers, etc).
Auto-upgrading is dumb, but who says that 'infected clients' would automatically download and install updates from untrusted remote sites? A few simple checks (e.g. PGP signatures, MD5 checksums from a trusted source) would basically prevent this.
still issues of passing cracks disguised as data remain
You mentioned this before, but I think it would be unlikely for a client to have such a security vulnerability. The separation between executable code and network data should be clear. The same idea applies to email, and with the notable exception of Outlook, such flaws are very rare (of course, they occur all the time, but that is thanks to the stupidity of HTML mail w/ scripting, and sloppy coding on MS's part).
it really needs high speed always connected computers.
Yeah, such as the millions of people now getting broadband internet at home (through cable, DSL, or whatever). Not to mention sympathetic corporations (or sympathetic sysadmins:P). Freenet is unlike Napster in that every client does not automatically become a server (AFAIK). Inserting data is completely different from becoming a node and serving data.
Anyone else have any better ideas?
I'm still not clear what you think the problem with Freenet is. It may not be the all-dancing, all-singing Messiah of information storage, but it's still a damn useful piece of software.
Rather than just dumping the results of the database query to disk, why not dump the complete, generated HTML into a static file? This doesn't even require modifying a lot of code - just write a simple daemon to send an http request ('lynx -dump' or whatever) to the correct page every X seconds/minutes, and save the results to a file. In the code, just divert requests to known static pages (either with a redirect, or just send the file on disk). You can get the web code to ignore the static file if the timestamp is too old.
Don't know if PHP and Perl/DBI have similar features.
Perl does, if you're using mod_perl. It's called Apache::DBI (caveat: it allows persistent connections, not true 'pools'. In other words, each Apache child established one, and only one, connection to the DB. It usually works just as well as a true connection pool).
You can increase the limit to 32K by recompiling Postgres. You can use the lztext type, which compresses data before inserting/retrieving it.
You can also wait for the 7.1 release, which will support rows of unlimited length. The changes are in the current CVS sources - I use them for development and they're perfectly stable.
Postgres doesn't support records larger than 8K (OK, it does if you modify the source, but I'm told there is a performance hit for doing this)
As you said, you can up the limit to 32K with a recompile. I haven't heard anything about a performance hit though. Also, the 'lztext' data type compresses all data before storing it, meaning you can get far more than 8K in. Finally, you can use Large Objects to store data of unlimited size (see the Postgres Programmers Guide for more info). BTW, this limit is fixed in the current CVS sources, and the upcoming 7.1 release (beta soon).
It doesn't seem like this will help. He said that he experiences the same connection problems when trying to access MySQL locally. It doesn't seem like a network congestion issue.
Also, the author seems under the impression that he has more than enough hardware to serve the content. It just seems like the software (MySQL, possibly the queries he's running) is causing problems.
Yeah, but Slash has tons more hardware (IIRC, the database server is a quad xeon with 2 gigs of RAM). The point is that he has more than enough hardware power to serve the content - but MySQL seems to be falling over itself under that kind of load.
I did overhear something about Stockwell Day saying he was going to legalize marijuana. I'm not a drug user (save caffeine), but I'll vote for him without hesitation if he really means it.
I haven't heard anything about this - but I'm not really informed... The two things I know about Day's policies are that he would have censored Eminem, and he supports religion in the classroom. I'd like to see marijuana legalized, but most of the rest of his policies seem rather authoritarian.
Slashdot Mirror
On x86, I prefer to run Debian. I'm considering getting a new Titanium G4 Powerbook, and I'd like to dual-boot OSX and a Linux distro. Has anyone tried Debian/PPC on the new Powerbook? I'd be very interested to hear your comments (positive or negative).
One impressive factor is SMP support: already! OpenBSD still doesn't support SMP and NetBSD just added it recently (not a flame of the BSD folks -- obviously, they're focusing on different goals. But I still think it's an impressive feature to have this early in its development).
Yes -- for example, if I shoot someone, that's illegal. If the government subsequently locks me up for 20 years (which is a form of physical force), that's perfectly legal.
I thought the raison d'etre of the 2nd Amendment was to prevent the government gaining a monopoly of force.
Right -- it's more complicated than that. As a citizen, you have the right to defend yourself -- and if necessary, to respond to an act of force with violence; but only the government can legally initiate the use of force. (That's Ayn Rand's view anyway...)
The government chose not to 'take down' Microsoft, dumb ass. Since the government has a legal monopoly of force, of course they are able to, if they choose to (whether they should be allowed to interfere in the economy is another issue entirely).
And also, Microsoft didn't 'take down' the US government -- not even intentionally. If anything, this attack will take the whitehouse website -- which is butt-ugly and serves zero official function -- down for a few hours. Big deal...
It was. There was never a commercially available project, but Netscape released the code to the Mozilla project when it was started. The reason the next Netscape browser was called 6 was because it shared virtually no code with the Classic codebase (i.e. 4.x and 5-pre-pre-alpha).
This community, and the community of warez/gamerz/hackerz. It's "free" software, either way :-)
They would probably use it to backup something
If that was the case, few people would need a CD rewriter. I'd say people typically use them to burn music CDs (my family wants me to burn CDs for them all the time, as well as MP3s), portable documents (i.e. burn this presentation, take it in to work -- more reliable and better storage than floppy), or illegal software (even computer illiterates don't like paying for MS products :-)). The number of non-technies who back up anything is quite small, in any case.
I would definately like to get my hands on quad density CDs.
Take it easy chief... I read your post again, and I don't think I misinterpreted anything. Would you be so kind as to point out what else I should have read in your original post?
3dfx management may be trying to benefit themselves at the expense of their own shareholder who are the true owners of the company, and that we believe there are ways to dispose of 3dfx's assets that would return more of it's value to us.
I wasn't talking about that. I was talking about the aspect of your post when you talked about nVidia 'screwing over 3dfx investors'. You actually didn't talk about 3dfx's management at all in your original post.
You sound bitter - why's that? nVidia should be concerned with one thing: profit. They should take steps to make as much money as they can off the video card industry. If they judge that the best way to do this is picking the bones of 3dfx's corpse, so be it. Really, what do you think? That nVidia should be concerned about the financial well-being of 3dfx investors? What kind of fscked up society would that be!
A shame? 3dfx, like any other company, are in the business of making profit -- in any way they can. If they are unable to do this, they shouldn't exist. That's integral to the concept of a capitalist economy. I've never understood people having "pity" for a failing company. If it's failing, there is almost always one primary reason: the company is incompetant! If 3dfx were a good enough company to make a profit, they would have continued; obviously, they were not.
Note: I'm not anti-3dfx at all (I still own a couple V2s) -- but I just think sometimes people need reminding we do live in a capitalism.
That makes no sense - if that was the true, the faster you learn, the less you will get done. Something like this would be more accurate:
[(time available * speed of learning) / ('barrier to entry': skills required to be useful)] * ability to work with others
You mean exactly like RSS? It's XML based and numerous web sites (including Slashdot and Freshmeat) use it. That's how Slashdot's Slashboxes work. You could set it up like so: user creates an account @ foo.com. The user creates a portfolio (say, when buying stocks, or to get news updates on those companies). The website gives you an option of getting automatically regenerated RSS summaries of your portfolio and the latest prices (as well as RSS of market indices, currency/commodities, etc). You can use any scripting language with an XML parser (i.e. basically any) to process the XML and present it to the user. And all thjs without needing to buy a $499 software product, or use 6GB of local disk space.
From what I've heard, WebQL doesn't sound that innovative (if it even works at all). But since there was no technical info in the article, it's hard to tell.
The idiocy of ranking UIs is that - what users are we talking about? You, a "person who owns multiple computers"? Those aren't exactly very impressive qualifications. Basically, you're ranking the UIs on your personal reaction to them. Fine - but what makes you think this applies to anyone else?
Window's UI is hardly useable for daily work
That's absurd. Tens of millions of people use Windows every day, at work and at home. You may not like it, but it's the truth.
Sorry for the rant. I just find it aggravating to hear Mac OS people say that no OS has ever approached the UI of the original Mac OS. As I've tried to show, that idea is nonsensical.
I'm happy we're getting replication at all. The companies that really need it won't mind paying for it, and the money goes into further PgSQL development. And (AFAIK) it's all going to be open source eventually.
You're right, of course. rsyncing the actual database files while the DB serving is running is incredibly dumb.
If you'd rather do it the 'old fashioned' way, what is the query load like? If the data is mostly static, you could send all the writes to 1 database and use all the DBs for reads. Every 2 or so hours, update the read-only databases. Something like rsync would be very good for that. Until erServer is released, this may be your best choice.
There's been some discussion about this on the PostgreSQL mailing lists - check the pgsql-general archives @ postgresql.org
No more easy than hacking into a web server and using the same crack to break into hundreds of other sites running the same software. How is this any different from traditional client/server stuff? Besides, a badly written piece of software will tend to have security holes. It's up to the author to code carefully, and it's up to us as consumers to only use proven secure software (obligitary OpenBSD plug here).
Auto-upgrading is dumb, but who says that 'infected clients' would automatically download and install updates from untrusted remote sites? A few simple checks (e.g. PGP signatures, MD5 checksums from a trusted source) would basically prevent this.
You mentioned this before, but I think it would be unlikely for a client to have such a security vulnerability. The separation between executable code and network data should be clear. The same idea applies to email, and with the notable exception of Outlook, such flaws are very rare (of course, they occur all the time, but that is thanks to the stupidity of HTML mail w/ scripting, and sloppy coding on MS's part).
Yeah, such as the millions of people now getting broadband internet at home (through cable, DSL, or whatever). Not to mention sympathetic corporations (or sympathetic sysadmins :P). Freenet is unlike Napster in that every client does not automatically become a server (AFAIK). Inserting data is completely different from becoming a node and serving data.
I'm still not clear what you think the problem with Freenet is. It may not be the all-dancing, all-singing Messiah of information storage, but it's still a damn useful piece of software.
Rather than just dumping the results of the database query to disk, why not dump the complete, generated HTML into a static file? This doesn't even require modifying a lot of code - just write a simple daemon to send an http request ('lynx -dump' or whatever) to the correct page every X seconds/minutes, and save the results to a file. In the code, just divert requests to known static pages (either with a redirect, or just send the file on disk). You can get the web code to ignore the static file if the timestamp is too old.
Perl does, if you're using mod_perl. It's called Apache::DBI (caveat: it allows persistent connections, not true 'pools'. In other words, each Apache child established one, and only one, connection to the DB. It usually works just as well as a true connection pool).
You can also wait for the 7.1 release, which will support rows of unlimited length. The changes are in the current CVS sources - I use them for development and they're perfectly stable.
As you said, you can up the limit to 32K with a recompile. I haven't heard anything about a performance hit though. Also, the 'lztext' data type compresses all data before storing it, meaning you can get far more than 8K in. Finally, you can use Large Objects to store data of unlimited size (see the Postgres Programmers Guide for more info). BTW, this limit is fixed in the current CVS sources, and the upcoming 7.1 release (beta soon).
It doesn't seem like this will help. He said that he experiences the same connection problems when trying to access MySQL locally. It doesn't seem like a network congestion issue.
Also, the author seems under the impression that he has more than enough hardware to serve the content. It just seems like the software (MySQL, possibly the queries he's running) is causing problems.
Yeah, but Slash has tons more hardware (IIRC, the database server is a quad xeon with 2 gigs of RAM). The point is that he has more than enough hardware power to serve the content - but MySQL seems to be falling over itself under that kind of load.
I haven't heard anything about this - but I'm not really informed... The two things I know about Day's policies are that he would have censored Eminem, and he supports religion in the classroom. I'd like to see marijuana legalized, but most of the rest of his policies seem rather authoritarian.