Here in Honolulu, the rather extensive bus system (rated best in the nation this year) is slated to be augmented in the future with... electric trams. They won't take power from overhead lines in true trolley style, though, but from plates in the pavement that will only provide power when the tram passes over. (Or, I suppose, when some resourceful local youth figure out how to cause a short and win a Darwin Award.)
Oh, and those reel mowers? Houses here are packed so closely that most lawns are smaller than most rooms. The local Sears stocks at least two or three different models, plus a couple of the powered ones - commonly used for golf courses, since reel mowers actually cut the grass better than rotary ones.
--
I'm with pc486 on this one. If ICANN extends their "enforcement" to existing gTLD's and stops allowing registrations in.net from companies that aren't providing network services, and stops allowing registrations in.org from entities that belong in.com or.net, I will be a very happy camper.
I consider NSOL's practice of encouraging entities to register in all three domains to be absolute heresy, driven by pure greed. Just one more reason NSOL needs to be launched into the Sun.
Okay, let's see. Motorola sunk what, $5 billion into this, and said "Gosh, we can't turn a profit on it."
Mccaw looked at picking it up for $500 million, a 90% discount, then said, "Gosh, we can't turn a profit on it."
Castle Harlan looked at picking it up for $50 million, a 99% discount, then said, "Gosh, we can't turn a profit on it."
Now there's a $30 million offer. That's less than 1 penny on the dollar. Or, in blue-light special terms, more than 99% off! I wonder just how little someone would have to pay to turn a profit on Iridium? Can it even be done?
I'm tempted to offer them $1,000 just for the privilege of de-orbiting the birds when and where I want. The "have your own meteor shower" feature would be worth it. And I wouldn't back out like those other bastards!:)
--
Those of us who're old enough may remember GEOS on the C-64, C-128 and Apple II from Berkeley Softworks, back in the '80s. BSW became GeoWorks around 1990, and sold the OS and app suites based on it for a few years, as well as selling it for PDA's like the Tandy/Casio/Sharp "Zoomer" and some HP OmniGo models.
Around the mid-'90s, GeoWorks focused more on smart phones (the
Nokia 9000 family of smart phones run GEOS), and desktop stuff was taken over by New Deal, Inc..
On the x86 platform, GEOS offered pre-emptive multitasking, multithreading, and object-oriented design (coded in something resembling Objective C, if I recall from the days when I had the SDK and developer docs, and in assembler) - and it did this a full five years ahead of Windows 95. In 1990, it had shared UI code for all apps, like GNOME and KDE are now doing.
It was also very fast as a platform - it was originally designed to run on an 8086 with 640K, and even the most recent versions are quite happy on a '286 with a meg or two. On anything "recent" in the way of a CPU, it should outperform just about anything - unless, of course, it's loading stuff over a dialup...
I'm a slow typist by/. standards - max out around 100 wpm. Gotta get me a Dvorak keyboard. But there's no way I'd ever get anywhere close to that speed writing. Not even 60-70 wpm. I can't write more than one or two letters in a second, and in the middle of a decently long word, I can pretty easily hit 3-4 keys a second.
I'd also have some concerns about the editing capabilities of the interface. I've used pen-based systems before (GEOS-based Tandy Zoomer) and generally going back and making changes is a real pain. --
Those of us who are fortunate enough to live in Hawaii are accustomed to most of our traffic going via "the mainland," with the exception of things that peer through HIX (our own lil' exchange!). I don't know whether there are any undersea links direct from us to Asia or not, but there might be. --
As someone who first set up a HA "shared-nothing" _production_ server cluster for a major (household name) customer over a year ago, I have a thought or two about the DEC solution you're so ga-ga over.
Simply put, any time all your cluster nodes share a resource, wham-o, you have a single point of failure. Oops. So when the RAID array that your shared boot disk sits on undergoes Sudden Massive Existence Failure, all your nodes are up a creek.
Ditto for anything else you're sharing.
If you're serious about availability, you need to replicate _all_ your services. And that means spending big-ass bucks. And probably not using Linux.
Yes, "shared-nothing" approaches have their drawbacks too - most notably the incredible annoyance of data synchronization between non-shared drives. But the entry point for commercial shared-anything solutions last year was something on the order of 35 _times_ the price of a shared-nothing Linux cluster, so if your hardware budget is anything shy of 6 digits, I think the "shared-nothing" approach will give you better availability.
I'm a geek. I married a girl who by all indications is smarter than me (she comes up with the classical theories of philosophy on her own without studying philosophy). She's also athletic, has a degree in a performing art, has done modeling for a company with the initials C.K., uses vi as her editor-of-choice on her Linux laptop, and gave birth to a really cute baby girl who we named (in proper geek fashion) Tera (nickname Terabit).
And I'm not even a fabulously-rich-on-paper dot-com geek.
Linus is also a good example. Everything I've read and seen indicates that Tove's a cutie, kicks butt (literally, as a multi-time national martial arts champions) and is no dumb bunny. Kids there, too.
Hmmm. Maybe the girls are getting tired of the dumb jocks? Maybe? Or am I too hopeful?
--
Not really good news, but not really bad news.
on
TurboLinux Layoffs
·
· Score: 2
I was sad to see that TurboLinux was cutting folks, since I've gotten to know some people there over the last year, and like Lonn says, they have (or had) some really, really good talent. I'm glad, though, that they're not planning more cuts.
As far as the original person wondering what they meant by "enterprise" products, I can only guess they're unaware of TurboLinux's TurboCluster Server clustering solution, which I've been working with for the last year.
TurboLinux Workstation and TurboLinux Server always struck me as "just one more distro," while TurboCluster Server was definitely a horse (or box) of a different color, and for quite some time was the only server clustering (as opposed to computational clustering) solution available in packaged form for Linux.
(My other thoughts on TLTCS are somewhere right around here for anyone curious.)
Anyway, it looks like they'll be continuing to support and focus on my favorite product, so I guess I should be happy. I should probably also mail my friends and see if they're still around... maybe we can pick some of them up where I work!;) --
Indeed. I'm looking at shelling out some money for another CompactFlash card for my camera soon too. (Couldn't imagine using SmartMedia, with its 64-meg limit, when CF is now up to 192-meg...)
When I stop to think about it, though, this OUM stuff doesn't replace a certain form factor, or a certain interface or API - it replaces a memory technology. If it works, and can be used as a replacement for DRAM and Flash, I can't think of anything to stop them from making SmartMedia and CompactFlash cards with OUM technology inside them. And if it stores more data in the same space, that'd be a pretty nice feature for digital cameras, too. Keep the interface pinout and form factor and all that the same, just replace the storage innards. --
I especially liked the read/write cycle lifespan, since for anything to replace existing memory technologies (like flash ram and CD-RW), it needs to improve on that performance parameter.
I'm a little curious, though, whether this is best suited for compact applications like flash RAM, or as a means of replacing both RAM and fixed-disk technologies. If the size reduction is substantial enough, and it's as fast as RAM, it's possible that a system might ship with a single fixed rotating rewritable drive (like a hard drive) using OUM technology to provide a significant chunk of storage space (say, multiple gigs) with high access and transfer speeds. Imagine your swap partition running at the same speed as your RAM. That'd be pretty spiffy. --
Ironic, considering that in actuality, astronauts have to be below a certain height, due to the close quarters and whatnot. I think anybody over 5 foot 7 (let's see, that's about 170 cm) is automatically disqualified.
Of course, this means that given existing technology, any tourist trips to the ISS or the lunar theme park or whatever will also be off-limits to the taller folk among us.
--
Re:TurboLinux TurboCluster Server
on
Linux Failover?
·
· Score: 1
Yep, TLTCS has been doing this sort of stuff since last year. Load balancing and failover. No big deal. I ran a cluster using it for some corporate websites, and now I run a cluster of laptops using it at home.
The only place it gets tricky is if your e-business stuff involves a whole load of stateful things, like you're saving user data on the servers during transactions and whatnot. Then you're getting into database replication, which is a whole 'nother can of big expensive worms. Scale it beyond a couple nodes, and you're dealing with SANs, at which point the can and worms are even bigger and more expensive.:)
Maintain state by passing data in hidden form variables or cookies or whatnot, and you won't have these problems. But anyway.
TLTCS is my friend. There are plenty of other options at various levels of maturity, and I'd think your consultant is... what was the phrase? "biting himself in the foot" by not being aware of products that are widely available... --
Slashdot Mirror
Oh, and those reel mowers? Houses here are packed so closely that most lawns are smaller than most rooms. The local Sears stocks at least two or three different models, plus a couple of the powered ones - commonly used for golf courses, since reel mowers actually cut the grass better than rotary ones.
--
So... did I miss the release of Quicktime for Linux? :)
--
I'm with pc486 on this one. If ICANN extends their "enforcement" to existing gTLD's and stops allowing registrations in .net from companies that aren't providing network services, and stops allowing registrations in .org from entities that belong in .com or .net, I will be a very happy camper.
I consider NSOL's practice of encouraging entities to register in all three domains to be absolute heresy, driven by pure greed. Just one more reason NSOL needs to be launched into the Sun.
--
Mccaw looked at picking it up for $500 million, a 90% discount, then said, "Gosh, we can't turn a profit on it."
Castle Harlan looked at picking it up for $50 million, a 99% discount, then said, "Gosh, we can't turn a profit on it."
Now there's a $30 million offer. That's less than 1 penny on the dollar. Or, in blue-light special terms, more than 99% off! I wonder just how little someone would have to pay to turn a profit on Iridium? Can it even be done?
I'm tempted to offer them $1,000 just for the privilege of de-orbiting the birds when and where I want. The "have your own meteor shower" feature would be worth it. And I wouldn't back out like those other bastards! :)
--
Those of us who're old enough may remember GEOS on the C-64, C-128 and Apple II from Berkeley Softworks, back in the '80s. BSW became GeoWorks around 1990, and sold the OS and app suites based on it for a few years, as well as selling it for PDA's like the Tandy/Casio/Sharp "Zoomer" and some HP OmniGo models.
Around the mid-'90s, GeoWorks focused more on smart phones (the Nokia 9000 family of smart phones run GEOS), and desktop stuff was taken over by New Deal, Inc..
On the x86 platform, GEOS offered pre-emptive multitasking, multithreading, and object-oriented design (coded in something resembling Objective C, if I recall from the days when I had the SDK and developer docs, and in assembler) - and it did this a full five years ahead of Windows 95. In 1990, it had shared UI code for all apps, like GNOME and KDE are now doing.
It was also very fast as a platform - it was originally designed to run on an 8086 with 640K, and even the most recent versions are quite happy on a '286 with a meg or two. On anything "recent" in the way of a CPU, it should outperform just about anything - unless, of course, it's loading stuff over a dialup...
--
I'd also have some concerns about the editing capabilities of the interface. I've used pen-based systems before (GEOS-based Tandy Zoomer) and generally going back and making changes is a real pain.
--
Those of us who are fortunate enough to live in Hawaii are accustomed to most of our traffic going via "the mainland," with the exception of things that peer through HIX (our own lil' exchange!). I don't know whether there are any undersea links direct from us to Asia or not, but there might be.
--
Simply put, any time all your cluster nodes share a resource, wham-o, you have a single point of failure. Oops. So when the RAID array that your shared boot disk sits on undergoes Sudden Massive Existence Failure, all your nodes are up a creek.
Ditto for anything else you're sharing.
If you're serious about availability, you need to replicate _all_ your services. And that means spending big-ass bucks. And probably not using Linux.
Yes, "shared-nothing" approaches have their drawbacks too - most notably the incredible annoyance of data synchronization between non-shared drives. But the entry point for commercial shared-anything solutions last year was something on the order of 35 _times_ the price of a shared-nothing Linux cluster, so if your hardware budget is anything shy of 6 digits, I think the "shared-nothing" approach will give you better availability.
--
And I'm not even a fabulously-rich-on-paper dot-com geek.
Linus is also a good example. Everything I've read and seen indicates that Tove's a cutie, kicks butt (literally, as a multi-time national martial arts champions) and is no dumb bunny. Kids there, too.
Hmmm. Maybe the girls are getting tired of the dumb jocks? Maybe? Or am I too hopeful?
--
As far as the original person wondering what they meant by "enterprise" products, I can only guess they're unaware of TurboLinux's TurboCluster Server clustering solution, which I've been working with for the last year.
TurboLinux Workstation and TurboLinux Server always struck me as "just one more distro," while TurboCluster Server was definitely a horse (or box) of a different color, and for quite some time was the only server clustering (as opposed to computational clustering) solution available in packaged form for Linux.
(My other thoughts on TLTCS are somewhere right around here for anyone curious.)
Anyway, it looks like they'll be continuing to support and focus on my favorite product, so I guess I should be happy. I should probably also mail my friends and see if they're still around... maybe we can pick some of them up where I work! ;)
--
So will your GPS receiver or cell phone also tip the cops off to the fact that you're driving with an open beer in the car?
--
When I stop to think about it, though, this OUM stuff doesn't replace a certain form factor, or a certain interface or API - it replaces a memory technology. If it works, and can be used as a replacement for DRAM and Flash, I can't think of anything to stop them from making SmartMedia and CompactFlash cards with OUM technology inside them. And if it stores more data in the same space, that'd be a pretty nice feature for digital cameras, too. Keep the interface pinout and form factor and all that the same, just replace the storage innards.
--
As far as life span, the web site presentation said something about 10 trillion write cycles.
--
I'm a little curious, though, whether this is best suited for compact applications like flash RAM, or as a means of replacing both RAM and fixed-disk technologies. If the size reduction is substantial enough, and it's as fast as RAM, it's possible that a system might ship with a single fixed rotating rewritable drive (like a hard drive) using OUM technology to provide a significant chunk of storage space (say, multiple gigs) with high access and transfer speeds. Imagine your swap partition running at the same speed as your RAM. That'd be pretty spiffy.
--
Yep, international projects tend to have their growing pains, at the very least. And both sides in this deal know that - remember EuroDisney?
--
Of course, this means that given existing technology, any tourist trips to the ISS or the lunar theme park or whatever will also be off-limits to the taller folk among us.
--
The only place it gets tricky is if your e-business stuff involves a whole load of stateful things, like you're saving user data on the servers during transactions and whatnot. Then you're getting into database replication, which is a whole 'nother can of big expensive worms. Scale it beyond a couple nodes, and you're dealing with SANs, at which point the can and worms are even bigger and more expensive. :)
Maintain state by passing data in hidden form variables or cookies or whatnot, and you won't have these problems. But anyway.
TLTCS is my friend. There are plenty of other options at various levels of maturity, and I'd think your consultant is... what was the phrase? "biting himself in the foot" by not being aware of products that are widely available...
--