We use ImageMagick to convert multi-page TIFFs that our customers send to us into our internal layout (i.e. correct size, depth, etc.). I'd like to try netpbm, but it doesn't seem to widely support multipage images:
If it weren't for the fact that it discards 86% of the input data, it'd be pretty useful. Seriously, though, I'd love to find out that I'm wrong. It seems much faster than ImageMagick for certain operations and I'd like to give it a real test.
That's interesting, and something I hadn't been aware of.
AFAICT, you can't even just ignore the GUI and configure the printers through CUPS directly.
I don't think there's anything keeping you from hacking at/etc/cups/printers.conf or similar. Also, the CUPS web interface is running on port 631 on our machine - have you tried using that instead of the GUI configurator?
just *try* to get it to deliver postscript to a remote print server.
As far as I know, I do that every day. My wife's iMac prints correctly to the HP LaserJet 1200 attached to my FreeBSD server via CUPS. Did I miss part of your point?
Erm, no. A browser misrenders a site by accepting valid input and generating out-of-spec output, such as by failing the Acid2 test. However, if the input is not deterministically parseable then all bets are off.
If you want to add 2 plus 2, but enter "2+3" into your calculator, would you get mad at it for giving an answer you didn't expect? Even if one brand of calculator often guesses that you really meant "2+2" and silently gives you that answer instead, I fail to see how you could expect another to do the same (and then claim that it "miscalculates" because it doesn't apply the same random guesswork as the first).
Really? The whole demestic spying thing isn't very liberal.
What does Bush have to do with conservatism? News flash: an increasing number of us are becoming loudly dissatisfied with the Republican leadership, which hasn't been conservative at all for quite some time.
Liberal comes from liberty; conservative comes from conserve.
No. Liberal and liberty share a common root: liber. Neither derive from the other.
One's clearly about getting more liberty, when the others about maintaining the level of liberty we had 200 years ago.
See, the funny thing is that conservatives believe that they are the group who desires more liberty. They (generally) believe that liberals are pro-society at the expense of individual liberty. Go and ask to one sometime; we're not nearly as scary as we're made out to be.
Again: neither liberals nor conservatives "own" freedom. Both of them desire it, but have different ideas about what it actually means.
No, you're wrong! ${my_issue} is much more important and urgent than ${your_issue}. In fact, Congress's lack of action on ${my_issue} is leading to the ${bad_thing} of hundreds of ${decimal_unit}s of ${country}'s citizens. Those who do not take immediate action on ${my_issue} while ignoring everything else (including ${your_issue}) are baby-eating murderers!
I'm not sure why there is such resistance here on/. (other than the fact that most/.'ers are possibly adolecent gamers) to the idea that activities you engage in for a large percentage of your time can have an impact on brain development and function.
Because the conclusion is always that the activities we/.ers grew up loving will surely turn you into a disconnected, emotionless, killing robot.
So we have to decide whether RMS is more like Joseph Smith or L. Rob Hubbard? I'll be a Free Software Mormon if you expand that polygamy thing a little bit. Oh, and also allow beer.
Thus, the bottom line here is, unfortunately, that what happened in this experiment was that one component of the total energy input, magnetic energy, which normally is not converted into heat, was converted into heat by a new mechanism.
Rats - no fusion. Instead, all we got is a previously unknown energy conversion that could possibly be useful in future creations. What's the point in getting a new energy conversion mechanism if it's not fusion?
Not just bandwidth, but if you were the head admin of their network, how thrilled would you be that somebody hung a big sign on your campus saying "please attack us"?
And yes, this challenge is sanctioned. I'm glad that the University of Wisconsin supports the genuine interests of its faculty, staff, and students, and encourages individual thought, research, discovery, and exploration. That's why it's a great place to be!
Yeah, because I like needing eight gigs of free hard drive to compile it with the options I use just so I can build in unknown shit like this. If I somehow accidentally found that on my own, I'd probably figure my machine was pwn3d and reinstall to bare metal. It wasn't funny when MS did it, and it's no more funny when OOo does it.
Grow up, folks. Stupid stunts like this hurt far more than they help. From now on, whenever people bitch about how slow OOo is, MS fanboys will have legitimate reasons to point and laugh. For that matter, I probably will too. Is it slow because it's complex and powerful, or slow because there are 300 other Easter eggs hiding out in there?
Seriously, yank this crap out and forget it never existed.
Database programmers seem to hate that approach because it shifts the center of gravity away from the database and toward the web app.
Even though I'm hardly a DBA, I hate that approach for the exact same reason. In my company, we don't have the privilege of rolling out a database that supports only one application. Instead, we have a huge legacy application written in Visual FoxPro, the web stuff that I've written in Python, and some lightweight apps (like custom image viewers) written in various.NET languages. We're moving more and more logic into the database - PostgreSQL on a beefy server - because that's the only component common to all these systems, and because that's were a lot of it belongs.
I'm assuming that you forgot to flip the xsize and the ysize in the second example. I also assume you trying to make a point that writing across memory is quicker then writing once every xsize*pixelsize bytes
Correct on both counts.
but surely than depends on the method of screen memory mapping?
Yes, but as I mentioned in another poster, images are almost always stored as an adjacent list of the values of one row, immediately followed by the list of values of the next row, etc. In other words, pixel[x,y] is most commonly at pixelarray[y*width+x], and iterating across that list in linear order will usually be the optimal algorithm.
I know what you mean in that code (tho it may contain mistakes) - it does NOT take assembly experience to see the cache benefit - granted I know 6502 and x86
But that's the thing: I don't know a single person who doesn't know assembler who understands the tricks you do. That was my point. I've gotten a lot of replies from people who understand CPU architecture who think the cache miss thing is obvious, but not one from a HLL-only programmer who saw the problem.
The two are not mutually exclusive - there's no inherent reason why you can't learn about low-level stuff when learning high-level programming - but that seems to be only a theoretical possibility. Again, I've never actually seen it in real life.
No, it's fairly obvious that one will run faster than the other. For a non-square pixel area, one will no-op for a range of pixels, or quite possibly crash before running to completion. But I'm guessing that you mistyped your example.
Bzzzt - wrong answer. Yes, that was a typo, but no, that's not why one would be faster than the other. Given that almost all images are stored with x-adjacent pixels being stored in consecutive memory locations (or nearly interleaved, with other data between them), the first example stands a much better chance of minimizing cache misses. For large images, the second example nearly guarantees an extremely high miss rate.
But that they don't usually teach that stuff in VB 101. It's not that they inherently couldn't, but that they just don't. Your typical ASM class, though, will coerce you into learning more about what your silicon is actually doing than you might want to know. That's good stuff to know, even if you don't ever plan to program in ASM again.
even though both are conceptually identical. There are many things that seem perfectly reasonable in high-level languages that turn out to be a really bad idea once you learn what's going on in the hardware. I'm sure it's possible to learn that stuff without hitting the metal a few times, but I've never, not one single time, ever met someone who's done so.
all of the "best practice" measures were taken to harden the box
No, they weren't. If all the filesystems that customers have write access to are mounted "noexec", then self-compiled binaries don't present a lot of exposure.
I'm not saying that it's not a good idea to remove GCC, just that its presence isn't an automatic compromise.
So in the end, from one legal dispute, three lawsuits emerge, and two of those suits were at best frivolous, taking advantage of the cost of a lawsuit as a tool to try to extract concessions.
Perhaps your friend and his friends should have turned that into a legal dispute, three lawsuits, and a hospitalization. Remember, never cost someone more than they'd have to pay to have you killed.
It's a pity the real poor coverage KOffice gets in the web compared to OpenOffice, being a really cool suite with great programs.
There is exactly one reason I don't use KOffice over OpenOffice.org: printed fonts look absolutely horrible. Specifically, the kerning is so awful that the results are sometimes unreadable. I'm not some hyper-nitpicky font fan, but an average user who wants his output to look nice - and KWord can't do it. There is absolutely no way I'd submit a resume or other important document printed from it.
This is apparently a common problem, and my understanding is that it's rooted in the QT libraries. In other words, it's not directly KWord's fault. Be that as it may, the end result still isn't usable for me.
Here's to hoping that the eventual QT4 version will look good and print nicely. If that happens, I'll drop OpenOffice.org like a hot potato. Until that, OOo is the only option I see for high quality output.
Well, maybe not the only option. I haven't tried AbiWord in ages, but I'm not eager to try to integrate a Gnome app into my KDE desktop. It's not that I dislike Gnome, but that its programs doesn't really blend in with all the other apps I use.
the "mark of the beast" means that you have to recant your faith to recieve the mark.
What if your faith involves not blindly ceding total control of your life to the government, e.g. by ingesting a universal tracking device? I'm not sure that recanting one's faith need be as straightforward as saying "God is dead, now tag me!"
Slashdot Mirror
If it weren't for the fact that it discards 86% of the input data, it'd be pretty useful. Seriously, though, I'd love to find out that I'm wrong. It seems much faster than ImageMagick for certain operations and I'd like to give it a real test.
AFAICT, you can't even just ignore the GUI and configure the printers through CUPS directly.
I don't think there's anything keeping you from hacking at /etc/cups/printers.conf or similar. Also, the CUPS web interface is running on port 631 on our machine - have you tried using that instead of the GUI configurator?
As far as I know, I do that every day. My wife's iMac prints correctly to the HP LaserJet 1200 attached to my FreeBSD server via CUPS. Did I miss part of your point?
You misspelled "ignore".
Erm, no. A browser misrenders a site by accepting valid input and generating out-of-spec output, such as by failing the Acid2 test. However, if the input is not deterministically parseable then all bets are off.
If you want to add 2 plus 2, but enter "2+3" into your calculator, would you get mad at it for giving an answer you didn't expect? Even if one brand of calculator often guesses that you really meant "2+2" and silently gives you that answer instead, I fail to see how you could expect another to do the same (and then claim that it "miscalculates" because it doesn't apply the same random guesswork as the first).
Thank you. It's always nice to feel appreciated. :-)
What does Bush have to do with conservatism? News flash: an increasing number of us are becoming loudly dissatisfied with the Republican leadership, which hasn't been conservative at all for quite some time.
Liberal comes from liberty; conservative comes from conserve.
No. Liberal and liberty share a common root: liber. Neither derive from the other.
One's clearly about getting more liberty, when the others about maintaining the level of liberty we had 200 years ago.
See, the funny thing is that conservatives believe that they are the group who desires more liberty. They (generally) believe that liberals are pro-society at the expense of individual liberty. Go and ask to one sometime; we're not nearly as scary as we're made out to be.
Again: neither liberals nor conservatives "own" freedom. Both of them desire it, but have different ideas about what it actually means.
Sincerely,
${someone}
I'm a conservative for the same reason.
Actually, the truth is that we're both pro-liberty, regardless of our other political leanings. Neither liberals nor conservatives "own" freedom.
Because the conclusion is always that the activities we /.ers grew up loving will surely turn you into a disconnected, emotionless, killing robot.
So we have to decide whether RMS is more like Joseph Smith or L. Rob Hubbard? I'll be a Free Software Mormon if you expand that polygamy thing a little bit. Oh, and also allow beer.
Rats - no fusion. Instead, all we got is a previously unknown energy conversion that could possibly be useful in future creations. What's the point in getting a new energy conversion mechanism if it's not fusion?
Now back to your regularly scheduled programming...
Sufficiently so, apparently:
Grow up, folks. Stupid stunts like this hurt far more than they help. From now on, whenever people bitch about how slow OOo is, MS fanboys will have legitimate reasons to point and laugh. For that matter, I probably will too. Is it slow because it's complex and powerful, or slow because there are 300 other Easter eggs hiding out in there?
Seriously, yank this crap out and forget it never existed.
Even though I'm hardly a DBA, I hate that approach for the exact same reason. In my company, we don't have the privilege of rolling out a database that supports only one application. Instead, we have a huge legacy application written in Visual FoxPro, the web stuff that I've written in Python, and some lightweight apps (like custom image viewers) written in various .NET languages. We're moving more and more logic into the database - PostgreSQL on a beefy server - because that's the only component common to all these systems, and because that's were a lot of it belongs.
Man up, nancy. The real world uses real tools for real work.
Oh, I get it! You were going for "+5: Ironic".
Correct on both counts.
but surely than depends on the method of screen memory mapping?
Yes, but as I mentioned in another poster, images are almost always stored as an adjacent list of the values of one row, immediately followed by the list of values of the next row, etc. In other words, pixel[x,y] is most commonly at pixelarray[y*width+x], and iterating across that list in linear order will usually be the optimal algorithm.
But that's the thing: I don't know a single person who doesn't know assembler who understands the tricks you do. That was my point. I've gotten a lot of replies from people who understand CPU architecture who think the cache miss thing is obvious, but not one from a HLL-only programmer who saw the problem.
The two are not mutually exclusive - there's no inherent reason why you can't learn about low-level stuff when learning high-level programming - but that seems to be only a theoretical possibility. Again, I've never actually seen it in real life.
Bzzzt - wrong answer. Yes, that was a typo, but no, that's not why one would be faster than the other. Given that almost all images are stored with x-adjacent pixels being stored in consecutive memory locations (or nearly interleaved, with other data between them), the first example stands a much better chance of minimizing cache misses. For large images, the second example nearly guarantees an extremely high miss rate.
But that they don't usually teach that stuff in VB 101. It's not that they inherently couldn't, but that they just don't. Your typical ASM class, though, will coerce you into learning more about what your silicon is actually doing than you might want to know. That's good stuff to know, even if you don't ever plan to program in ASM again.
I totally disagree. Difficult? Complicated? Sure, but not mean. People who have learned assembler are the ones who understand why
will usually run faster thaneven though both are conceptually identical. There are many things that seem perfectly reasonable in high-level languages that turn out to be a really bad idea once you learn what's going on in the hardware. I'm sure it's possible to learn that stuff without hitting the metal a few times, but I've never, not one single time, ever met someone who's done so.No, they weren't. If all the filesystems that customers have write access to are mounted "noexec", then self-compiled binaries don't present a lot of exposure.
I'm not saying that it's not a good idea to remove GCC, just that its presence isn't an automatic compromise.
Perhaps your friend and his friends should have turned that into a legal dispute, three lawsuits, and a hospitalization. Remember, never cost someone more than they'd have to pay to have you killed.
There is exactly one reason I don't use KOffice over OpenOffice.org: printed fonts look absolutely horrible. Specifically, the kerning is so awful that the results are sometimes unreadable. I'm not some hyper-nitpicky font fan, but an average user who wants his output to look nice - and KWord can't do it. There is absolutely no way I'd submit a resume or other important document printed from it.
This is apparently a common problem, and my understanding is that it's rooted in the QT libraries. In other words, it's not directly KWord's fault. Be that as it may, the end result still isn't usable for me.
Here's to hoping that the eventual QT4 version will look good and print nicely. If that happens, I'll drop OpenOffice.org like a hot potato. Until that, OOo is the only option I see for high quality output.
Well, maybe not the only option. I haven't tried AbiWord in ages, but I'm not eager to try to integrate a Gnome app into my KDE desktop. It's not that I dislike Gnome, but that its programs doesn't really blend in with all the other apps I use.
What if your faith involves not blindly ceding total control of your life to the government, e.g. by ingesting a universal tracking device? I'm not sure that recanting one's faith need be as straightforward as saying "God is dead, now tag me!"