More important than hearts and minds, the W3C let the charter for the xhtml2 working group expire earlier this year in favour of allocating resources to the (x)html5 effort.
Their current business model - which is working out well - necessitates slick marketing and high margin products. Apple is no more going to change that than BMW is going to start selling economy cars.
Who cares when HDTV first came out? Only 13.7% of households had an HDTV when the Wii launched. Even now it's somewhere between 35-40%. They made a bet on targeting a broader market would pay off more than chasing specs, and it seems to have paid off for them.
Their stockholders seem to care a lot more about the fifty million people who did buy the console than people who go on websites to say they won't.
Ermmm, how about simply moving to a smaller form factor drive. Most servers have moved to 2.5" bays already and I see no reason to doubt SAN vendors will start offering SFF shelves as well. Another approach is to just throw enough cheap disk at the problem that long background rebuilds aren't a concern. Multiple RAID sets off redundant storage stacks have better DR characteristics regardless.
That isn't to say alternate approaches to data integrity aren't called for as well. It's clear that future filesystems are going to include some level of end-to-end checksumming and offer software approaches to data replication. Likewise, there are plenty of approaches to data replication that don't follow traditional filesystem conventions; consider Google FS or CouchDB.
Try a rebuild on a much larger aggregate running a dual parity array under load. Trust me, they can easily run days. Say you have a 16 disk aggregate using 1TB 7200RPM disks. Because you need every block in a stripe to reconstruct parity, you need to read from the other disks to reconstruct; so 14 reads and 1 write per block.
You're also misunderstanding how the SSD caching works for ZFS. Blocks are only pulled in after repeated requests, which isn't going to be the case for a resliver. There will be at least some benefit to read ahead caching in memory, but even that has sharply diminishing returns, particularly with the ZFS rebuild strategy of reconstructing at a file level rather than a linear block rebuild. That approach has significant benefits though. By walking through the metadata instead of blindly copying blocks you don't have to rebuild empty space, and if - god forbid - you lose more than one drive in a RAID-Z or two drives in a RAID-Z2 array, you still have a partial recovery to work with.
Umm. Looking on NewEgg there isn't a single board that doesn't have at least one PATA port and spot checking motherboard I couldn't find a single one that didn't include at least one serial port and two PS/2 ports. Legacy-free boards are available, but they're still nowhere near the bulk of the market.
A super computer is also a low volume product. The more of something you sell, the more the incremental costs matter. Let's say you sell a $1500 POS system that cost a half million to develop and costs about $500 to build. Sell a thousand and your hardware and labor costs are about even. Sell ten thousand and your development costs are an order of magnitude smaller than component costs.
There are multiple re-spin projects (the most prominent being Scientific Linux). Not hard to import new GPG keys and point at new repos. Or just build the SRPMS yourself. It's rarely more involved "rpmbuild --rebuild ". And of course there's always the last ditch option of ponying up for a RHEL license.
The project itself isn't a single point of failure.
Erm, the project wasn't relying on him. If he hadn't resurfaced about the only real repercussion would have been moving to a new domain name. Things have been running smoothly for a long time without his active participation.
Yeah, we've only save a few hundred thousand so far. Bad decision there.:) And now we'll be force to... nothing really, since the CentOS project isn't breaking up. Even if it was, we could pull updates from the Scientific Linux repos or rebuild an SRPM now and again or switch to Red Hat painlessly.
Bad examples. Solaris 8 doesn't reach the end of it's service life until March 2012, legacy versions of Irix will reach end of support no sooner than December 2013 and even the oldest versions of OpenVMS for the Alpha will be supported through 2012. And of course all three platforms have new versions coming out, so there's an upgrade path on current hardware platforms.
Really, enterprise vendors (including Red Hat) have an excellent history of supporting their paying customers for extended periods.
Your example of RHEL not getting an update in a timely manner is wrong. They issues RHSA-2009:1162-1 nine days ago to address the Firefox vulnerabilities. That's the same day the vulnerabilities were announced by the Mozilla foundation.
You're seriously on crack here. The SPARC instruction set has remained larely the same, but the underlying chip architecture is drastically different. The same can be said of Intel's x86 architecture, including exceptional backwards compatability; current 64-bit processors retain compatability with the original 8086 released over thirty years ago. For that matter it's pretty true for every major architecture; e.g. MIPS, POWER and ARM.
And Intel released their first 32-bit chip in 1985, which not only is significantly more than fifteen years ago, but also before there were any SPARC products on the market. You would have more of a point if you were talking about the 64-bit market, as the initial 64-bit parts from Sun hit the market in 1995 while Intel didn't release any 64-bit products until 2001 and didn't extend the x86 architecture until 2004, when they licensed AMD's implementation.
In the long run, FOSS converges to one winner, challenged by many (much smaller) creatures.
Yeah, just like vi and bash and pine and Gnome crowded out all their competitors!
Try to build a new browser or new *nix kernel and see how many people you project gets. Try to build a new browser or new *nix kernel and see how many people you project gets. Try to compete with Apache. Try to build a new OpenOffice (though one that had a major corp backing).
There are plenty of vibrant, active projects for all the categories you mentioned. Several BSD projects pre-date Linux and there are several successful newer kernels - Darwin, OpenSolaris, DragonFly BSD, etc. WebKit has matured to the point that it's a serious threat to the dominance to the ubiquity of Mozilla (and with an entire ecosystem of derived browsers rather than mostly one monolithic entity). And there are tons of projects competing (or complimenting) Apache - nginx, lighttpd, cherokee among the up and comers.
You also ignore that there are plenty of software categories that have no clear forerunner, such as text editors, development environments and music players.
Or using ridiculously easy passwords or variations on the same password (e.g. "Wordpass!", "w0rd-pass", "Wordpa$$", "WORDpass", "WORDpa55"). Not to mention 2500 accounts that go years between password changes because it too much of a pain in the ass to be thorough (assuming you even remember or track every account).
The new one (DSi) is a 133MHz ARM9 with a 33MHz ARM7 coprocessor. Main processor on the older ones was 67MHz. It also upgrades it from 4MB to 16MB main memory and includes a music player and web browser.
Not as capable as an ultraportable but it certainly sounds like it would fit the bill for the "browser in a box" market.
Slashdot Mirror
More important than hearts and minds, the W3C let the charter for the xhtml2 working group expire earlier this year in favour of allocating resources to the (x)html5 effort.
These little WRT's and such have the equivalent of 8-bit 200 Mhz CPU's.
They have what is a 32-bit 200 MHz processor. Specifically this one in the referenced Linksys model.
I was thinking George Bush should sue. They so totally ripped off the idea of a stylized W from his election campaign.
Their current business model - which is working out well - necessitates slick marketing and high margin products. Apple is no more going to change that than BMW is going to start selling economy cars.
There's a difference between something a company says and unsubstantiated rumours on the internet.
Who cares when HDTV first came out? Only 13.7% of households had an HDTV when the Wii launched. Even now it's somewhere between 35-40%. They made a bet on targeting a broader market would pay off more than chasing specs, and it seems to have paid off for them.
Their stockholders seem to care a lot more about the fifty million people who did buy the console than people who go on websites to say they won't.
Ermmm, how about simply moving to a smaller form factor drive. Most servers have moved to 2.5" bays already and I see no reason to doubt SAN vendors will start offering SFF shelves as well. Another approach is to just throw enough cheap disk at the problem that long background rebuilds aren't a concern. Multiple RAID sets off redundant storage stacks have better DR characteristics regardless.
That isn't to say alternate approaches to data integrity aren't called for as well. It's clear that future filesystems are going to include some level of end-to-end checksumming and offer software approaches to data replication. Likewise, there are plenty of approaches to data replication that don't follow traditional filesystem conventions; consider Google FS or CouchDB.
Try a rebuild on a much larger aggregate running a dual parity array under load. Trust me, they can easily run days. Say you have a 16 disk aggregate using 1TB 7200RPM disks. Because you need every block in a stripe to reconstruct parity, you need to read from the other disks to reconstruct; so 14 reads and 1 write per block.
You're also misunderstanding how the SSD caching works for ZFS. Blocks are only pulled in after repeated requests, which isn't going to be the case for a resliver. There will be at least some benefit to read ahead caching in memory, but even that has sharply diminishing returns, particularly with the ZFS rebuild strategy of reconstructing at a file level rather than a linear block rebuild. That approach has significant benefits though. By walking through the metadata instead of blindly copying blocks you don't have to rebuild empty space, and if - god forbid - you lose more than one drive in a RAID-Z or two drives in a RAID-Z2 array, you still have a partial recovery to work with.
Umm. Looking on NewEgg there isn't a single board that doesn't have at least one PATA port and spot checking motherboard I couldn't find a single one that didn't include at least one serial port and two PS/2 ports. Legacy-free boards are available, but they're still nowhere near the bulk of the market.
A super computer is also a low volume product. The more of something you sell, the more the incremental costs matter. Let's say you sell a $1500 POS system that cost a half million to develop and costs about $500 to build. Sell a thousand and your hardware and labor costs are about even. Sell ten thousand and your development costs are an order of magnitude smaller than component costs.
People don't come with instruction manuals. You might actually have to listen to one another. :)
There are multiple re-spin projects (the most prominent being Scientific Linux). Not hard to import new GPG keys and point at new repos. Or just build the SRPMS yourself. It's rarely more involved "rpmbuild --rebuild ". And of course there's always the last ditch option of ponying up for a RHEL license.
The project itself isn't a single point of failure.
Erm, the project wasn't relying on him. If he hadn't resurfaced about the only real repercussion would have been moving to a new domain name. Things have been running smoothly for a long time without his active participation.
It provides a stable server platform for organizations that don't require vendor support.
Yeah, we've only save a few hundred thousand so far. Bad decision there. :) And now we'll be force to ... nothing really, since the CentOS project isn't breaking up. Even if it was, we could pull updates from the Scientific Linux repos or rebuild an SRPM now and again or switch to Red Hat painlessly.
Bad examples. Solaris 8 doesn't reach the end of it's service life until March 2012, legacy versions of Irix will reach end of support no sooner than December 2013 and even the oldest versions of OpenVMS for the Alpha will be supported through 2012. And of course all three platforms have new versions coming out, so there's an upgrade path on current hardware platforms.
Really, enterprise vendors (including Red Hat) have an excellent history of supporting their paying customers for extended periods.
Your example of RHEL not getting an update in a timely manner is wrong. They issues RHSA-2009:1162-1 nine days ago to address the Firefox vulnerabilities. That's the same day the vulnerabilities were announced by the Mozilla foundation.
B-O-O H-O-O
So you have to add "without cheats" when bragging. Times is hard. :)
You're seriously on crack here. The SPARC instruction set has remained larely the same, but the underlying chip architecture is drastically different. The same can be said of Intel's x86 architecture, including exceptional backwards compatability; current 64-bit processors retain compatability with the original 8086 released over thirty years ago. For that matter it's pretty true for every major architecture; e.g. MIPS, POWER and ARM.
And Intel released their first 32-bit chip in 1985, which not only is significantly more than fifteen years ago, but also before there were any SPARC products on the market. You would have more of a point if you were talking about the 64-bit market, as the initial 64-bit parts from Sun hit the market in 1995 while Intel didn't release any 64-bit products until 2001 and didn't extend the x86 architecture until 2004, when they licensed AMD's implementation.
In the long run, FOSS converges to one winner, challenged by many (much smaller) creatures.
Yeah, just like vi and bash and pine and Gnome crowded out all their competitors!
Try to build a new browser or new *nix kernel and see how many people you project gets. Try to build a new browser or new *nix kernel and see how many people you project gets. Try to compete with Apache. Try to build a new OpenOffice (though one that had a major corp backing).
There are plenty of vibrant, active projects for all the categories you mentioned. Several BSD projects pre-date Linux and there are several successful newer kernels - Darwin, OpenSolaris, DragonFly BSD, etc. WebKit has matured to the point that it's a serious threat to the dominance to the ubiquity of Mozilla (and with an entire ecosystem of derived browsers rather than mostly one monolithic entity). And there are tons of projects competing (or complimenting) Apache - nginx, lighttpd, cherokee among the up and comers.
You also ignore that there are plenty of software categories that have no clear forerunner, such as text editors, development environments and music players.
Five hundred is "very low end"? Not for general purpose desktops. Think more like $250-300.
Or using ridiculously easy passwords or variations on the same password (e.g. "Wordpass!", "w0rd-pass", "Wordpa$$", "WORDpass", "WORDpa55"). Not to mention 2500 accounts that go years between password changes because it too much of a pain in the ass to be thorough (assuming you even remember or track every account).
Must be a nice world where the only cause for an unclean shutdown is power interupption. And where the power supply itself never goes tits up.
Mighty small western world you live in.
The HP 2140 offers a 10.1" 720p (1366x768) as will the upcoming Dell Mini 10 (unless early reports are wrong).
The new one (DSi) is a 133MHz ARM9 with a 33MHz ARM7 coprocessor. Main processor on the older ones was 67MHz. It also upgrades it from 4MB to 16MB main memory and includes a music player and web browser.
Not as capable as an ultraportable but it certainly sounds like it would fit the bill for the "browser in a box" market.