Re:Dapper is good, but it's not there yet.
on
Ubuntu 6.06 Reviewed
·
· Score: 3, Insightful
My co-worker wasted a couple hours trying to install a second graphics card on Ubuntu, before inserting the card into a Windows machine and realizing that the card simply didn't work. IMO, that illustrates a significant reason to stay away from desktop Linux.
Yes. That reason: your coworker should not be doing his own installs.
I'm honestly not an elitist - I tease my coworkers about using Windows and they tease me about Linux, but it's all in fun. However, Linux tried for two hours to tell your pal that the graphics card didn't work, and he wasn't willing to believe it until Windows said the exact same thing. That doesn't indicate a flaw with either OS, but rather a serious problem with your coworker.
Did he also get frustrated when Linux wasn't able to find his SCSI drives, and even more so when Windows confirmed that he only had IDE installed?
So if that happened, then reality would eventually come to resemble Hollywood depictions -- where every agency or company is running their own custom OS...
Yep, because there are no operating systems distributed for free anymore.
I dropped the thing on my foot once and now my big toe makes this funny clicking sound whenever I flex it.
Sharpening stone? I don't know what they make ax heads out of where you live, but Sears doesn't sell the depleted uranium wedges your local store must carry.
I have two favorite keyboards: the IBM Model M I'm using here at work, and the Happy Hacking Lite I have at home. I love the feel of the Model M, but don't like the location or spacing of the function keys (LALT-F5 and higher to switch to higher-numbered desktops) gets a little painful and awkward by the end of the day. The Happy Hacking board is wonderful, but I do occasional miss the extra keys (ever play Tux Racer when you have to chord arrow keys?).
Given that my boss says he'll buy me a nice keyboard within reason - everyone else seems to have those Microsoft Unnatural keyboards that I hate - what would you recommend as a happy medium between the two?
Charging for email won't put any small organizations in the poorhouse, just cause them to rethink how they operate.
As in, "I think it's time we quit." Pick a charity - any charity. Now, imagine them asking would-be donors to give them money in advance so that they can send solicitations to them. For example, I like the EFF and enjoy getting stuff from them, but I'm not going to pay them for the privilege of receiving their donation requests and neither will anyone else. See the problem yet?
It's easy to invent schemes that will stick it to the spammers. It's incredibly difficult, though, to come up with ideas that won't cause enormous collateral damage.
This would not only kill OSS, but the whole software industry would go bankrupt in no time.
No way. There are far more of us who develop custom in-house software than people who write stuff that gets sold. You might severely hurt the software-as-a-product industry, but wouldn't touch the software-as-office-automation economy.
My point being that Jack Thompson doesn't speak for all conservative Christians -- and neither do you. Both of you claim to do so on some level, however... or you just said you did, anyway.
You're probably right. I seemed to be speaking for all conservative Christians when I'm obviously not. Still, Jackass Thompson belongs to many additional demographics that people aren't collectively blaming. Some Christians agree with him and some don't. Some men agree with him and some don't. Some 50-somethings agree with him and some don't.
It's not fair or accurate to say that his idiotic beliefs stem from his religious and political alignments when many others with those same affiliations completely disagree with him. That's really all I meant.
You act as though pointing out that no system will every be completely perfect in this regard was equivalent to saying that no system could ever be substantially better than the existing system in this area so as to warrant a change.
No. My position is that SMTP, which came out nearly 24 years ago and is extremely entrenched, is so well-suited to its environment that any would-be contenders will have to be amazingly attractive before they can replace it. There's no such animal out there now and I'm not familiar with any in development. Until such time, it's a bit silly to ask whether "the time has come to ditch email" since it clearly hasn't.
Some of the biggest problems with spam aren't just whether or not the end user is annoyed - it's how much bandwidth is taken up by junk email and how much it costs to filter it all out.
You don't have to vaccinate 100% of the population to prevent certain epidemics. Similarly, you don't have to harden 100% of mailservers before the rest are either shielded by their upstream relays or the cost of sending spam begins to outway its revenues.
I can quite easily imagine a world where spam no longer consumes many resources because no one bothers to try it anymore.
You should re-read the article from DJB. It makes a lot of sense
I did and it doesn't. I routinely need to send out 50,000 copies of a customer newsletter. Right now, SMTP allows me to start the process now and gradually spool out the copies at my network's own convenience until I'm finished. Under Dan's crackpot idea, I send a broadcast to 50,000 customers letting them know that there's a newsletter waiting for them. When they all come to work at 9AM and simultaneously attempt to download a 1MB PDF, my router cries tears of pain and my customers hate my slow-loading message.
Dan's idea sounds fine under certain very limited circumstances, but can't possibly work in the real world.
I have faith in the unlimited creativity of ethically challenged people. Beyond that, though, ask your local mathematician, cryptographer, computer scientist, or philosopher whether it's theoretically possible to design a perfect communications system that reliably delivers all wanted messages and no unwanted messages. Short answer: no.
SMTP still works exceedingly well for its purpose. Understand this: spam and viruses will propigate through any message transfer protocol that will ever be invented. We already have effective technologies for filtering that stuff out of SMTP traffic, but if admins can't be bothered to implement them for their customers, I don't know why they'd implement similar measures on other protocols.
Put another way, if you run your own mailserver and still get spam and viruses, it's because you haven't chosen to address the problem. If you use someone else's mailserver and still get spam and viruses, it's because they haven't chosen to address the problem. Nothing stands between you and a clean inbox but motivation, whether your own or your ISP's.
And no, broken hacks like DJB's "Internet Mail 2000" will never get real-world acceptance as they make it as difficult for legitimate bulk senders to broadcast as for spammers. SMTP is here to stay as the standard method for (somewhat) reliably routing messages between people on unaffiliated networks. Replacing it with a similar system with new pitfalls isn't the answer we're looking for.
Unless "the public" is helping me pay the bank loan I took out to buy the land, "the public" can kiss my butt. Alternatively, where do you park? I want to use your car, you selfish bastard.
Look at Thompson and his supporters and you'll find the same conservative Christians who have been fighting the culture war for control of expression and action in the US for decades now.
On behalf of conservative Christians throughout America: you have no idea what you're talking about. I am far more interested in regulating my own household than asking the government to do so. Jack Thompson is a nutcase who has much in common with your average Christian as he does the average man, the average 50-something, the average white person, or the average person who doesn't wear glasses.
And don't forget that liberals have been advocating censorship for decades as well. I say that not as an excuse, but as a reminder: don't think that every last person in your political demographic is as anti-censorship as you'd like to believe. Pointing at the other guys and yelling doesn't help anyone, least of all you.
Then just a short while later, really expensive DDR SDRAM with data rates in multiple gigabits per second hits the market and we've been paying for memory in a new higher price range than the old pre-RDRAM high price range ever since.
Correct - sort of. There was a temporary glut of RAM as fabs came online and started churning like made, but when a third of them went offline simultaneously (fire? earthquake? I forget) prices spiked immediately. There weren't any more parts in the pipeline to feed those empty sockets that people just learned how to fill.
RDRAM had a lot of technical problems with it.
Chief among them was that its performance sucked, and sucked hard. It was very good at streaming a huge contiguous block to the processor, but beyond horrible at switching to another block. Imagine a CPU that was excellent at applying a single operation to a large chunk of memory but awful at everything else. Voila! You've invented P4+RDRAM!
I can imagine applications where it would've rocked, like encoding video using an instruction block small enough to fit entirely in cache so that the only memory fetches were to the input data. You definitely wouldn't have wanted to run a busy multipurpose server off it, though.
Rambus developed DDR and holds the patent on it, among other things that have shown up in modern commodity RAM.
That's also partially true, and the reason that everyone in the know hates Rambus. They took part in the DDR development process, but lied to JEDEC by "forgetting" to mention that the methods they were proposing as part of that process were already patented - by them. Had they mentioned that minor fact, modern DDR would've had a different design, but one that was less convenient for Rambus's patent portfolio.
Python's loading algorithm goes something like this:
A module tries to import another module named "foo".
Find "foo.py" in the search path.
Does "foo.pyc" ("c" for "compiled") exist in the same directory? If so, is it newer than foo.py? If both are true, then load it.
Otherwise, compile "foo.py" and attempt to write the compiled bytecode to "foo.pyc".
Never in my time developing Python have I once encountered a problem with stale compiled objects. If your language of choice doesn't have this same functionality or its equivalent, then the problem is with the language and not the concept.
I can't personally vouch for its speed, because I don't run a high-volume mailserver, nor do I have the hardware to really give the MTA that much of a workout (it just becomes disk-bound on my systems).
I do, or at least one of my clients does. He runs a reasonably high-volume ecommerce site, and has many (about 50,000) opted in subscribers to his newsletter. We tried our best to get Sendmail to play nicely with that volume, but the system would inevitably slow to a crawl for long periods of time whenever he sent a batch of mail (taking the webserver on the same machine with it). By our best, I mean that we tore through the bat book, tried delayed sending, created parallel queues with their own runners - everything we could find documented or rumored on Google and Usenet.
After experimenting with Postfix on my personal servers, I convinced him to give it a shot. I installed it, ported over his Sendmail configuration, stopped one and started the other, and crossed my fingers.
It worked.
We confirmed that everything was working as expected, then he clicked the dreaded "Send now!" link. We watched as the outbound queue grew to 50,000 messages, then tailed maillog to watch them start spewing out at a record pace. Even though outbound traffic was heavy, the system never broke a sweat and the webserver kept chugging along happily.
I like Sendmail and am quite comfortable digging around in its.mc files (.cf? Therefore but by the grace of God...), but Postfix showed me what a modern MTA is capable of. I've since switched every Sendmail installation in my responsibility over to Postfix and I've never regretted it for a minute.
These devices can save $10 a month, if not more, when used instead of an old PC.
Because nothing says Return On Investment like spending $1,000 over the cost of a standard PC for small-office file and print sharing in order to save $10 per month.
Hope you plan on using the same storage appliance for 8 years; I personally don't have much use for the 18GB drives I had in 1998.
If you intend to use that exclusively as a SOHO NAS 24/7, then your garden-variety 20 Watt, $150 dedicated NAS will pay back itself in about a year by lower power consumption alone, never mind it being more environmentally friendly, taking up less space, and making less noise.
With my earlier estimate, (150-20)/1000*24*365*.12 = $137 dollars per year. Where are you finding these dirt-cheap NASes?
You seem to have an opinion on the subject, so I'll ask the same question others already have: what network filesystem would you recommend for sharing files from a FreeBSD (or possibly Linux) box to an OS X system? I'm currently using good ol' NFS to serve media files across a LAN. Is that a reasonable setup, or is there something you think should work better (faster, more compatibly)?
Some use several ARM or MIPS CPUs, which can offer suitable processing capabilities without the immense energy consumption of even a single x86 chip.
A regular Pentium 4 draws on the order of 65 watts. At $0.12/kWh, that's.065kW*24h/d*.12$/kWh*365d/y = $68 dollars per year. Even if the custom NAS's CPU runs on air, you'd have to run the Pentium system for a long time to make up the difference in cost.
Slashdot Mirror
Yes. That reason: your coworker should not be doing his own installs.
I'm honestly not an elitist - I tease my coworkers about using Windows and they tease me about Linux, but it's all in fun. However, Linux tried for two hours to tell your pal that the graphics card didn't work, and he wasn't willing to believe it until Windows said the exact same thing. That doesn't indicate a flaw with either OS, but rather a serious problem with your coworker.
Did he also get frustrated when Linux wasn't able to find his SCSI drives, and even more so when Windows confirmed that he only had IDE installed?
A few million similar votes should get the point across.
Yep, because there are no operating systems distributed for free anymore.
Sharpening stone? I don't know what they make ax heads out of where you live, but Sears doesn't sell the depleted uranium wedges your local store must carry.
Given that my boss says he'll buy me a nice keyboard within reason - everyone else seems to have those Microsoft Unnatural keyboards that I hate - what would you recommend as a happy medium between the two?
As in, "I think it's time we quit." Pick a charity - any charity. Now, imagine them asking would-be donors to give them money in advance so that they can send solicitations to them. For example, I like the EFF and enjoy getting stuff from them, but I'm not going to pay them for the privilege of receiving their donation requests and neither will anyone else. See the problem yet?
It's easy to invent schemes that will stick it to the spammers. It's incredibly difficult, though, to come up with ideas that won't cause enormous collateral damage.
No way. There are far more of us who develop custom in-house software than people who write stuff that gets sold. You might severely hurt the software-as-a-product industry, but wouldn't touch the software-as-office-automation economy.
You're probably right. I seemed to be speaking for all conservative Christians when I'm obviously not. Still, Jackass Thompson belongs to many additional demographics that people aren't collectively blaming. Some Christians agree with him and some don't. Some men agree with him and some don't. Some 50-somethings agree with him and some don't.
It's not fair or accurate to say that his idiotic beliefs stem from his religious and political alignments when many others with those same affiliations completely disagree with him. That's really all I meant.
Sure, you do. No corporation, small business, non-profit, ISP, or mailing list operator will touch it with a 20-foot pole, though.
Coincidentally, guess who tends to decide what email technologies will be used?
No. My position is that SMTP, which came out nearly 24 years ago and is extremely entrenched, is so well-suited to its environment that any would-be contenders will have to be amazingly attractive before they can replace it. There's no such animal out there now and I'm not familiar with any in development. Until such time, it's a bit silly to ask whether "the time has come to ditch email" since it clearly hasn't.
You don't have to vaccinate 100% of the population to prevent certain epidemics. Similarly, you don't have to harden 100% of mailservers before the rest are either shielded by their upstream relays or the cost of sending spam begins to outway its revenues.
I can quite easily imagine a world where spam no longer consumes many resources because no one bothers to try it anymore.
I did and it doesn't. I routinely need to send out 50,000 copies of a customer newsletter. Right now, SMTP allows me to start the process now and gradually spool out the copies at my network's own convenience until I'm finished. Under Dan's crackpot idea, I send a broadcast to 50,000 customers letting them know that there's a newsletter waiting for them. When they all come to work at 9AM and simultaneously attempt to download a 1MB PDF, my router cries tears of pain and my customers hate my slow-loading message.
Dan's idea sounds fine under certain very limited circumstances, but can't possibly work in the real world.
I have faith in the unlimited creativity of ethically challenged people. Beyond that, though, ask your local mathematician, cryptographer, computer scientist, or philosopher whether it's theoretically possible to design a perfect communications system that reliably delivers all wanted messages and no unwanted messages. Short answer: no.
Put another way, if you run your own mailserver and still get spam and viruses, it's because you haven't chosen to address the problem. If you use someone else's mailserver and still get spam and viruses, it's because they haven't chosen to address the problem. Nothing stands between you and a clean inbox but motivation, whether your own or your ISP's.
And no, broken hacks like DJB's "Internet Mail 2000" will never get real-world acceptance as they make it as difficult for legitimate bulk senders to broadcast as for spammers. SMTP is here to stay as the standard method for (somewhat) reliably routing messages between people on unaffiliated networks. Replacing it with a similar system with new pitfalls isn't the answer we're looking for.
Unless "the public" is helping me pay the bank loan I took out to buy the land, "the public" can kiss my butt. Alternatively, where do you park? I want to use your car, you selfish bastard.
On behalf of conservative Christians throughout America: you have no idea what you're talking about. I am far more interested in regulating my own household than asking the government to do so. Jack Thompson is a nutcase who has much in common with your average Christian as he does the average man, the average 50-something, the average white person, or the average person who doesn't wear glasses.
And don't forget that liberals have been advocating censorship for decades as well. I say that not as an excuse, but as a reminder: don't think that every last person in your political demographic is as anti-censorship as you'd like to believe. Pointing at the other guys and yelling doesn't help anyone, least of all you.
You got "+5: Rand"? Here, on Slashdot? You, sir, are truly worthy of my respect and admiration.
Correct - sort of. There was a temporary glut of RAM as fabs came online and started churning like made, but when a third of them went offline simultaneously (fire? earthquake? I forget) prices spiked immediately. There weren't any more parts in the pipeline to feed those empty sockets that people just learned how to fill.
RDRAM had a lot of technical problems with it.
Chief among them was that its performance sucked, and sucked hard. It was very good at streaming a huge contiguous block to the processor, but beyond horrible at switching to another block. Imagine a CPU that was excellent at applying a single operation to a large chunk of memory but awful at everything else. Voila! You've invented P4+RDRAM!
I can imagine applications where it would've rocked, like encoding video using an instruction block small enough to fit entirely in cache so that the only memory fetches were to the input data. You definitely wouldn't have wanted to run a busy multipurpose server off it, though.
Rambus developed DDR and holds the patent on it, among other things that have shown up in modern commodity RAM.
That's also partially true, and the reason that everyone in the know hates Rambus. They took part in the DDR development process, but lied to JEDEC by "forgetting" to mention that the methods they were proposing as part of that process were already patented - by them. Had they mentioned that minor fact, modern DDR would've had a different design, but one that was less convenient for Rambus's patent portfolio.
Python's loading algorithm goes something like this:
Never in my time developing Python have I once encountered a problem with stale compiled objects. If your language of choice doesn't have this same functionality or its equivalent, then the problem is with the language and not the concept.
I do, or at least one of my clients does. He runs a reasonably high-volume ecommerce site, and has many (about 50,000) opted in subscribers to his newsletter. We tried our best to get Sendmail to play nicely with that volume, but the system would inevitably slow to a crawl for long periods of time whenever he sent a batch of mail (taking the webserver on the same machine with it). By our best, I mean that we tore through the bat book, tried delayed sending, created parallel queues with their own runners - everything we could find documented or rumored on Google and Usenet.
After experimenting with Postfix on my personal servers, I convinced him to give it a shot. I installed it, ported over his Sendmail configuration, stopped one and started the other, and crossed my fingers.
It worked.
We confirmed that everything was working as expected, then he clicked the dreaded "Send now!" link. We watched as the outbound queue grew to 50,000 messages, then tailed maillog to watch them start spewing out at a record pace. Even though outbound traffic was heavy, the system never broke a sweat and the webserver kept chugging along happily.
I like Sendmail and am quite comfortable digging around in its .mc files (.cf? Therefore but by the grace of God...), but Postfix showed me what a modern MTA is capable of. I've since switched every Sendmail installation in my responsibility over to Postfix and I've never regretted it for a minute.
I can't find any specific comparison between it and NFS. What does it do better than NFS?
Because nothing says Return On Investment like spending $1,000 over the cost of a standard PC for small-office file and print sharing in order to save $10 per month.
Hope you plan on using the same storage appliance for 8 years; I personally don't have much use for the 18GB drives I had in 1998.
With my earlier estimate, (150-20)/1000*24*365*.12 = $137 dollars per year. Where are you finding these dirt-cheap NASes?
You seem to have an opinion on the subject, so I'll ask the same question others already have: what network filesystem would you recommend for sharing files from a FreeBSD (or possibly Linux) box to an OS X system? I'm currently using good ol' NFS to serve media files across a LAN. Is that a reasonable setup, or is there something you think should work better (faster, more compatibly)?
A regular Pentium 4 draws on the order of 65 watts. At $0.12/kWh, that's .065kW*24h/d*.12$/kWh*365d/y = $68 dollars per year. Even if the custom NAS's CPU runs on air, you'd have to run the Pentium system for a long time to make up the difference in cost.