Which has always amazed me about the name sticking so long. Within six months of "Pulp Fiction" being released, a conservative Catholic woman of my acquaintence made the statement, "you used to be able to stop at people's house and ask to use the phone if you had car trouble, but now you don't know if they've got a Gimp in the basement or what!" At that point, any lingering questions about whether they were making fun of the disabled, or trying to make people squirm with that name, were pretty much settled.
I've really wondered about all the coy denials, attempts at cute mascots, etc, over the years, (such as Wilbur, the doggish icon), and why they didn't just roll with what they'd started. They could put up a leather-and-zipper logo on the home-page, or a caricature of Mappelthorpe with the bullwhip. Or they could change it to "Gnu Raster Image Processor", or follow the lead of the Squid team and just name it "Calimari".
Of course, if I have to make a choice, I'd still rather they added adjustment layers and 16-bit support rather than changed the name, but that's one more for the bug queue.
Because if you don't torture yourself, you'll have to make an appointment with the government, and they're terribly behind schedule these days. So, in order to have the job completed in a timely manner, we're outsourcing non-essential services, and the duty of torturing yourself has fallen to you.
Oh, but you're so close to the edge... Give in, feel the evangalism, follow the One True Path. In your heart of hearts, you know that Linux is only a transient state, a ripple in the pool of computing, before the Hurd blossoms forth in all its glory. It is time. Give up Evangalism, and become a Prophet of Hurd!
I always thought the best way to secure a lab of XP machines was to take a epoxy kit, and fill in the network and USB ports. If you take that extra second to pull the CD-ROM flap down, you can secure those as well.
It tends to reduce the resale value, but it does cut down on a lot of nonsense.
A firesale Fuel could be a fun box. Probably outperformed by various commodity systems now, but so much cooler.
The NUMA architectures are nice, but doesn't IBM do those as well with the p690s? What I genuinely miss are the big vector machines. Some algorithms vectorized extremely well, and parallelized poorly. I remember a talk once by a crystallographer who said he was running 95% of his program on the CM5(?) at Pittsburgh, except for one subroutine which could only be vectorized. Every cycle through the refinement he'd pound the CM5 flat-out for an hour, then it would dump everything onto a Cray for one subroutine, pass all the data back to the CM5, and continue until done.
Outside of positronic ionization (low-performance, high cost) mass spectrometry, I thought that was the coolest abuse of technology by chemists to date.
Point taken about the SuperDrome pricing, and I'm still an Itanium fan (and beginning to understand Chingachgook at the end of "Last of the Mohicans" as a result), but the worry would be that for mid-sized problems, 2-4 proc boxes from HP, and interconnects from Myricom/Quadrics, will get you as far as most of SGI's line-up. PNNL built a massive Itanium cluster that way, and as opposed to buying SGI, you get a company (HP) which can still afford to invest in research in various technologies. Maybe SGI, sufficiently slimmed down, can keep developing their solutions and carve out a niche with Itanium-based Originish systems (if they hadn't gone Chapter 11, I was considering a mid-sized Altix at one point), but the competition isn't sitting on their hands either.
I wish them luck, but I think they spent too much time thrashing around in the wrong markets, and missed the boat. Pity, but so it goes.
Wierdly enough, if you drill hard enough you'll find their "Prism" system, which looks like the old Immersadesk, augmented by Reality Walls and Cave-type environments. OTOH, they won't tell you what it's running on, and the two Irix workstations still in the line-up are marked "EOL Dec 2006".
Therefore, they're going to compete in HPC with Itanium and Opteron systems, which seems to be a recipe for getting crushed by the Terra/Cray hybrid (under Cray's name), HP and their Itanium servers through SuperDrome systems, and IBM/Sun on the smaller Opteron boxes. Add to this that they've fired to many engineers, this has to be a delaying action before the real end: six guys running a consulting company out of a Mountain-View garage.
They really are a case of, "time to sell whatever assets are left, return the proceeds to the stockholders, and say, "it was fun"". However, since they just came out of bankruptcy, the stock is probably worth less than nothing, so time to sell whatever assets are left, order a pizza and six-pack of cheap beer with the proceeds, then turn out the lights.
You have never had to work with Physicists. Most of them I've known would have been happy if KWord (or OpenOffice, etc) had dropped the silly alphabet support, and just used MathML.
So, if you can improve MathML support, and remove support for words, you'll have a guaranteed user community.
Unfortunately, what best describes your desire is the distinctly Non-Free Aperture from Apple.
However, it's good to state those requirements and desires openly, as hopefully some amateur photographer who's a good programmer will take up the challenge. By now someone has shot his sister's wedding on a 4GB CF card in JPEG format, and is staring at 10K pictures that need tagged, identically post-processed, and sorted. This person just needs a little encouragement to start coding.
From the Original post: 'Microsoft seems to envision... but also the security that protects those computers from viruses and other online threats,'
Not to be picky, but on my Solaris boxes, I don't call up McAffee every time a security vulnerability is released, nor do I call them to protect my AIX systems from Crackers either. I expect that Sun and IBM, respectively, will secure their OS, issue patches, and provide the appropriate tools to manage security. We've been letting Microsoft get away with fobbing that duty off on third-parties for far too long. Pity if that impacts Symantec's business model, but Microsoft should have years ago either (a) fixed their OS or (b) taken the tcp/ip stack out and stuck a big, neon-orange, sticker on every box and install disk which reads, "This Products Is Terminally Insecure and If You Let It Connect to a Network, 12-Year Old Script Kiddies Will OWN Your Valuable Corporate DATA! Within 20 Minutes Or Less!"
It's hard in a case like this to know which one of them (Microsoft or Symantec) to have less sympathy for.
The GP meant the original comment about 16-bit in the context that most photographers use: 16-bits/channel for 48-bit total, versus 8-bit/channel for 24-bit total. True, you need more bits/channel to do these transitions smoothly, and PS does apparently use some sort of witchcraft (i.e. undisclosed algorithm) to interpolate values and make smoother curves.
It's honestly been years since I've heard of anyone running in 16-bit total color, though I still have a program or two around that warns you not to do it.
So, we're back where we started; I'm sure i'm not the only one who wants GIMP to take the (12-16bit/channel) output of their scanners or cameras, not throw any of it away, non-destructively edit said data, and then consistently output it to a variety of profiled devices. That's all. As I said, GIMP's interface is not one of the things that bothers me. The lack of higher than 8-bit/channel does, though hopefully the integration of GEGL will move that one out of the way as well.
Well, there's 16-bit color support, non-destructive image editing (adjustment layers), CMYK and good profiling tools for RGB (lightjets) devices. Text mode could do a better job kerning as well (the third German example on their screenshots web page illustrates the issue), and some of the tools need a little polish. Maybe the problems will go away with 16-bit color, but it tends to posterize images easily if you do harsh curves adjustments.
From over here, I'd like to see the X11 dependence on the Macs go away. Pitch the GTK base and use QT, which is already efficiently cross-platform on Macs, Linux, and Windows.
As for the interface, so be it. If the other issues are fixed, the interface can be learned quickly enough. I used to use it for web images, and still have a certain fondness for 0.54, which ran on our SGI workstations. Maybe someone can ressurect that code-base and issue it as LIMP (Light Image Manipulation Program).
I'm glad to hear somebody speak up for XFS, since though I was less than fond of SGI in general (greasy salescritters and insecure OS), I never had a problem with XFS. I used that FS, and never lost data, on machines ranging from Indigos (which I had to tape over the power button to keep users from shutting them off like a PC), through Origin2000s. Same story for JFS on IBM hardware from RS/6000 320H through SP2. This was at locations with dirty power, where sometimes the workstations would get powered down unexpectedly. (once I got to see the cause; it's an amazing sight when you see a squirrel electrocute itself on a transformer, and go up in a ball of flame)
Therefore, I have to politely ask, "What the *BLEEP* have you Linux people *DONE* to those filesystems?"
He didn't say he was going to Windows. He could be a Frustrated AIX user. There's also the VM/CMS community which has trouble adapting to Linux as well. Something about not having a card punch and six IBM tech's in the box when they install it.
it's not stable enough to detect except by its decay chain. It would be nice if they would work on getting over the hump to the next island of stability, so that we could bag these things in an ion trap, and measure their mass directly. OTOH, if this keeps physicists occupied and out of the bars, I'm all for it.
More practically, you want to seal it off by itself (heavy curtains or folding partitions may be enough), the turn an AC on inside the mini-room, and threaten anyone who turns it off. I went through years of this at a former job, where the U maintained that cooling wasn't infrastructure, so our cluster's cooling was our problem. We used a portable unit for a while (and just vented the heat into the ceiling tiles, so the people above us had a warm floor), but eventually the answer was take over controllable space, and install a Liebert cold-water recirculating unit, as well as having the building airflow modified. Expensive, but we needed that headroom. Your situation is much smaller, so a closet with its own chiller and guaranteed air-circulation should do it. (Presuming, of course, that by 'server' you mean '1 to 2 proc Intel box pulling 500W max', rather than, 'I'm sharing an office with an E10K because we have nowhere else to put it.")
Rule 1 of Offices: the most expensive member sets the temperature
Rule 2 of Offices: the business data is more expensive than even a bunch of employees.
That plutonium might not be "missing". I heard a talk from one of the chemists working on remediation at Hanford, who said that at Oak Ridge they'd discovered a significant portion of the "missing" plutonium hanging out as drifts of barely sub-critical plutonium dust in the ventilation system. Not 'explode' subcritical, mind you, but 'a little more accumulation, and we'll have a real radiation event' subcritical.
So, rather than having been repurposed as weapons, it could still be polluting the facilities where it was used.
You have a foreign head of government you don't like. You "teleport" yourself to his position, and suddenly he's you. Half of me says, "way overdue" and the other half says, "erk... I'm not sure I'm ready for that kind of future".
The voice of thousands speaking in unison, "Hugo Chavez of Venezuala, We are the George Bush of America. Prepare to be assimilated."
The other side of why it's still there is would be to consider combustible hydrocarbons extracted from rocks. Otherwise known as oil, these materials still have biomolecules, such as hopanes and porphyrins, which are recognizable as being of terrestrial or marine origin. The hopanes, as I was told once by an organic geochemist working for a large, now merged, oil company, are used to tell the origin of the source of oil, and can give an indication of overall quality. (light versus heavy, etc). So, unless you believe that oil is spontaneously generated, or somehow makes complicated steroidal molecules which look exactly like modern plant steroids by purely inorganic processes, then it's reasonable to to presume that careful preservation in a sealed environment of a tough molecule, such as collagen or chitin, would also be possible.
So, while initially being surprised at the announcement, a moment's thought yielded, "as long as they're not claiming they have viable, sequenceable, dna, it's probably ok".
My theory about posts like this is that someone has an alternative-energy plan which is powered by political invective. Unfortunately, they needed another 10,000 units for this week's test.
Well, maybe not, but wouldn't you pay money to see Intel's CEO in his dark suit appear in a commercial singing, "Flash!...Ahhhhahhhhh....", with the surviving members of Queen backing him up?
Not much money, I know. But a 6" Subway vegetarian worth, probably.
Yes, but your free market is failing in one, crucial, aspect in the case of oil:
The money freely exchanged for the commodity is sent to repressive, fanatical, regimes that not only do not have our best interest at heart, but would frankly like to see us disappear in a satisfyingly violent manner from the face of the earth. While the free-market economics works in the abstract (and does fine if we're talking about Mexican, US-offshore, British North sea, oil), spending money to finance fundamentalist, intolerant, madrasses, rogue-state nuclear programs, alternative-combatants in any number of regional theatres, really doesn't seem like a good policy.
In an ideal world with rational actors, I'd agree that the FM is the way to go, but this is a case where doing the ground-work now so that we can still maintain some sort of modern economy then would seem to be prudent. At the same time, if it means that in the short-term we can tell any number of nut-job petty tyrants that they can go pound sand, then so much the better.
Slashdot Mirror
Which has always amazed me about the name sticking so long. Within six months of "Pulp Fiction" being released, a conservative Catholic woman of my acquaintence made the statement, "you used to be able to stop at people's house and ask to use the phone if you had car trouble, but now you don't know if they've got a Gimp in the basement or what!" At that point, any lingering questions about whether they were making fun of the disabled, or trying to make people squirm with that name, were pretty much settled.
I've really wondered about all the coy denials, attempts at cute mascots, etc, over the years, (such as Wilbur, the doggish icon), and why they didn't just roll with what they'd started. They could put up a leather-and-zipper logo on the home-page, or a caricature of Mappelthorpe with the bullwhip. Or they could change it to "Gnu Raster Image Processor", or follow the lead of the Squid team and just name it "Calimari".
Of course, if I have to make a choice, I'd still rather they added adjustment layers and 16-bit support rather than changed the name, but that's one more for the bug queue.
Because if you don't torture yourself, you'll have to make an appointment with the government, and they're terribly behind schedule these days. So, in order to have the job completed in a timely manner, we're outsourcing non-essential services, and the duty of torturing yourself has fallen to you.
Oh, but you're so close to the edge... Give in, feel the evangalism, follow the One True Path. In your heart of hearts, you know that Linux is only a transient state, a ripple in the pool of computing, before the Hurd blossoms forth in all its glory. It is time. Give up Evangalism, and become a Prophet of Hurd!
I think Edward Albee was the master of this.
"Poison; meditation; skiing; ants - nothing worked." -Edward Albee
I always thought the best way to secure a lab of XP machines was to take a epoxy kit, and fill in the network and USB ports. If you take that extra second to pull the CD-ROM flap down, you can secure those as well.
It tends to reduce the resale value, but it does cut down on a lot of nonsense.
A firesale Fuel could be a fun box. Probably outperformed by various commodity systems now, but so much cooler.
The NUMA architectures are nice, but doesn't IBM do those as well with the p690s? What I genuinely miss are the big vector machines. Some algorithms vectorized extremely well, and parallelized poorly. I remember a talk once by a crystallographer who said he was running 95% of his program on the CM5(?) at Pittsburgh, except for one subroutine which could only be vectorized. Every cycle through the refinement he'd pound the CM5 flat-out for an hour, then it would dump everything onto a Cray for one subroutine, pass all the data back to the CM5, and continue until done.
Outside of positronic ionization (low-performance, high cost) mass spectrometry, I thought that was the coolest abuse of technology by chemists to date.
Point taken about the SuperDrome pricing, and I'm still an Itanium fan (and beginning to understand Chingachgook at the end of "Last of the Mohicans" as a result), but the worry would be that for mid-sized problems, 2-4 proc boxes from HP, and interconnects from Myricom/Quadrics, will get you as far as most of SGI's line-up. PNNL built a massive Itanium cluster that way, and as opposed to buying SGI, you get a company (HP) which can still afford to invest in research in various technologies. Maybe SGI, sufficiently slimmed down, can keep developing their solutions and carve out a niche with Itanium-based Originish systems (if they hadn't gone Chapter 11, I was considering a mid-sized Altix at one point), but the competition isn't sitting on their hands either.
I wish them luck, but I think they spent too much time thrashing around in the wrong markets, and missed the boat. Pity, but so it goes.
Wierdly enough, if you drill hard enough you'll find their "Prism" system, which looks like the old Immersadesk, augmented by Reality Walls and Cave-type environments. OTOH, they won't tell you what it's running on, and the two Irix workstations still in the line-up are marked "EOL Dec 2006".
Therefore, they're going to compete in HPC with Itanium and Opteron systems, which seems to be a recipe for getting crushed by the Terra/Cray hybrid (under Cray's name), HP and their Itanium servers through SuperDrome systems, and IBM/Sun on the smaller Opteron boxes. Add to this that they've fired to many engineers, this has to be a delaying action before the real end: six guys running a consulting company out of a Mountain-View garage.
They really are a case of, "time to sell whatever assets are left, return the proceeds to the stockholders, and say, "it was fun"". However, since they just came out of bankruptcy, the stock is probably worth less than nothing, so time to sell whatever assets are left, order a pizza and six-pack of cheap beer with the proceeds, then turn out the lights.
You have never had to work with Physicists. Most of them I've known would have been happy if KWord (or OpenOffice, etc) had dropped the silly alphabet support, and just used MathML.
So, if you can improve MathML support, and remove support for words, you'll have a guaranteed user community.
Unfortunately, what best describes your desire is the distinctly Non-Free Aperture from Apple.
However, it's good to state those requirements and desires openly, as hopefully some amateur photographer who's a good programmer will take up the challenge. By now someone has shot his sister's wedding on a 4GB CF card in JPEG format, and is staring at 10K pictures that need tagged, identically post-processed, and sorted. This person just needs a little encouragement to start coding.
their OS is....?
... but also the security that protects those computers from viruses and other online threats,'
From the Original post: 'Microsoft seems to envision
Not to be picky, but on my Solaris boxes, I don't call up McAffee every time a security vulnerability is released, nor do I call them to protect my AIX systems from Crackers either. I expect that Sun and IBM, respectively, will secure their OS, issue patches, and provide the appropriate tools to manage security. We've been letting Microsoft get away with fobbing that duty off on third-parties for far too long. Pity if that impacts Symantec's business model, but Microsoft should have years ago either (a) fixed their OS or (b) taken the tcp/ip stack out and stuck a big, neon-orange, sticker on every box and install disk which reads, "This Products Is Terminally Insecure and If You Let It Connect to a Network, 12-Year Old Script Kiddies Will OWN Your Valuable Corporate DATA! Within 20 Minutes Or Less!"
It's hard in a case like this to know which one of them (Microsoft or Symantec) to have less sympathy for.
The GP meant the original comment about 16-bit in the context that most photographers use: 16-bits/channel for 48-bit total, versus 8-bit/channel for 24-bit total. True, you need more bits/channel to do these transitions smoothly, and PS does apparently use some sort of witchcraft (i.e. undisclosed algorithm) to interpolate values and make smoother curves.
It's honestly been years since I've heard of anyone running in 16-bit total color, though I still have a program or two around that warns you not to do it.
So, we're back where we started; I'm sure i'm not the only one who wants GIMP to take the (12-16bit/channel) output of their scanners or cameras, not throw any of it away, non-destructively edit said data, and then consistently output it to a variety of profiled devices. That's all. As I said, GIMP's interface is not one of the things that bothers me. The lack of higher than 8-bit/channel does, though hopefully the integration of GEGL will move that one out of the way as well.
Thank you. I do miss GIMP some times for quick web-graphics editing, or touching up a slide before a talk, and this would be nice to have.
I know, and it's thrown in there as a personal annoying feature, as I used gimp for quick picture hacking on my pre-mac laptop.
As for porting Photoshop to QT so it runs on Unices as well, sure. If they do that, then maybe they'll do pagemaker as well.
Well, there's 16-bit color support, non-destructive image editing (adjustment layers), CMYK and good profiling tools for RGB (lightjets) devices. Text mode could do a better job kerning as well (the third German example on their screenshots web page illustrates the issue), and some of the tools need a little polish. Maybe the problems will go away with 16-bit color, but it tends to posterize images easily if you do harsh curves adjustments.
From over here, I'd like to see the X11 dependence on the Macs go away. Pitch the GTK base and use QT, which is already efficiently cross-platform on Macs, Linux, and Windows.
As for the interface, so be it. If the other issues are fixed, the interface can be learned quickly enough. I used to use it for web images, and still have a certain fondness for 0.54, which ran on our SGI workstations. Maybe someone can ressurect that code-base and issue it as LIMP (Light Image Manipulation Program).
I'm glad to hear somebody speak up for XFS, since though I was less than fond of SGI in general (greasy salescritters and insecure OS), I never had a problem with XFS. I used that FS, and never lost data, on machines ranging from Indigos (which I had to tape over the power button to keep users from shutting them off like a PC), through Origin2000s. Same story for JFS on IBM hardware from RS/6000 320H through SP2. This was at locations with dirty power, where sometimes the workstations would get powered down unexpectedly. (once I got to see the cause; it's an amazing sight when you see a squirrel electrocute itself on a transformer, and go up in a ball of flame)
Therefore, I have to politely ask, "What the *BLEEP* have you Linux people *DONE* to those filesystems?"
He didn't say he was going to Windows. He could be a Frustrated AIX user. There's also the VM/CMS community which has trouble adapting to Linux as well. Something about not having a card punch and six IBM tech's in the box when they install it.
it's not stable enough to detect except by its decay chain. It would be nice if they would work on getting over the hump to the next island of stability, so that we could bag these things in an ion trap, and measure their mass directly. OTOH, if this keeps physicists occupied and out of the bars, I'm all for it.
as to cooling it, the answer is Yes.
More practically, you want to seal it off by itself (heavy curtains or folding partitions may be enough), the turn an AC on inside the mini-room, and threaten anyone who turns it off. I went through years of this at a former job, where the U maintained that cooling wasn't infrastructure, so our cluster's cooling was our problem. We used a portable unit for a while (and just vented the heat into the ceiling tiles, so the people above us had a warm floor), but eventually the answer was take over controllable space, and install a Liebert cold-water recirculating unit, as well as having the building airflow modified. Expensive, but we needed that headroom. Your situation is much smaller, so a closet with its own chiller and guaranteed air-circulation should do it. (Presuming, of course, that by 'server' you mean '1 to 2 proc Intel box pulling 500W max', rather than, 'I'm sharing an office with an E10K because we have nowhere else to put it.")
Rule 1 of Offices: the most expensive member sets the temperature
Rule 2 of Offices: the business data is more expensive than even a bunch of employees.
That plutonium might not be "missing". I heard a talk from one of the chemists working on remediation at Hanford, who said that at Oak Ridge they'd discovered a significant portion of the "missing" plutonium hanging out as drifts of barely sub-critical plutonium dust in the ventilation system. Not 'explode' subcritical, mind you, but 'a little more accumulation, and we'll have a real radiation event' subcritical.
So, rather than having been repurposed as weapons, it could still be polluting the facilities where it was used.
So, picture the following scenario:
You have a foreign head of government you don't like. You "teleport" yourself to his position, and suddenly he's you. Half of me says, "way overdue" and the other half says, "erk... I'm not sure I'm ready for that kind of future".
The voice of thousands speaking in unison, "Hugo Chavez of Venezuala, We are the George Bush of America. Prepare to be assimilated."
The other side of why it's still there is would be to consider combustible hydrocarbons extracted from rocks. Otherwise known as oil, these materials still have biomolecules, such as hopanes and porphyrins, which are recognizable as being of terrestrial or marine origin. The hopanes, as I was told once by an organic geochemist working for a large, now merged, oil company, are used to tell the origin of the source of oil, and can give an indication of overall quality. (light versus heavy, etc). So, unless you believe that oil is spontaneously generated, or somehow makes complicated steroidal molecules which look exactly like modern plant steroids by purely inorganic processes, then it's reasonable to to presume that careful preservation in a sealed environment of a tough molecule, such as collagen or chitin, would also be possible.
So, while initially being surprised at the announcement, a moment's thought yielded, "as long as they're not claiming they have viable, sequenceable, dna, it's probably ok".
My theory about posts like this is that someone has an alternative-energy plan which is powered by political invective. Unfortunately, they needed another 10,000 units for this week's test.
So... cue the Freepers at 3... 2... 1...
Well, maybe not, but wouldn't you pay money to see Intel's CEO in his dark suit appear in a commercial singing, "Flash!...Ahhhhahhhhh....", with the surviving members of Queen backing him up?
Not much money, I know. But a 6" Subway vegetarian worth, probably.
Yes, but your free market is failing in one, crucial, aspect in the case of oil:
The money freely exchanged for the commodity is sent to repressive, fanatical, regimes that not only do not have our best interest at heart, but would frankly like to see us disappear in a satisfyingly violent manner from the face of the earth. While the free-market economics works in the abstract (and does fine if we're talking about Mexican, US-offshore, British North sea, oil), spending money to finance fundamentalist, intolerant, madrasses, rogue-state nuclear programs, alternative-combatants in any number of regional theatres, really doesn't seem like a good policy.
In an ideal world with rational actors, I'd agree that the FM is the way to go, but this is a case where doing the ground-work now so that we can still maintain some sort of modern economy then would seem to be prudent. At the same time, if it means that in the short-term we can tell any number of nut-job petty tyrants that they can go pound sand, then so much the better.