The _real_ power of the web is the eventual elimination of middle-men consolidators like Amazon. When you want something, go directly to the author or publisher (through their public relations systems, whatever they are) and cut out the middleman profit. We shouldn't have to keep paying for people to move stuff from one warehouse to another warehouse before it gets to us. Buy from the first warehouse.
IMNSHO, the real power of the web is that middlemen can become "thinner" and "broader". I.e. they take less overhead on a transaction and can support more suppliers and customers. Witness Amazon.com, eBay, Yahoo, mp3.com, mysap.com, etc. Most people do not want to go into the trouble of finding all suppliers, comparing prices, etc., etc. Middlemen didnot (and will not) go away; they will just get smarter.
All that disintermediation stuff was so much Wired talk way back in 1996... engineers never lie; we just approximate the truth.
I said this in the previous discussion, but I think it bears repeating: Give Bezos a break.
I assume that most of the/. crowd was on the Net before 1995. For those of you that were, which was the first website to conduct serious "e-commerce", selling real things, not just shareware or porn? Amazon.com.
And Amazon not only has become the de-facto Internet bookstore (when you post a link about some obscure book, how many times do you go looking for it on bn.com?) but also the gold standard for e-commerce sites.
Amazon may not be the world's best or most profitable Web company (incl. patent rant here) but noone can deny that Bezos had a vision and worked hard to make it a reality. And by doing so he legitimized commerce over the Internet which is really the primary reason, us techies are doing better now than any other professional group.
I was hoping for better memory from this crowd...
engineers never lie; we just approximate the truth.
Errr... KTEL. They announced they were moving their *web servers* to Red Hat and got like a ~25% stock jump. It's equivalent to say eToys announcing they're gonna ship via USPS instead of UPS from now on (I dunno what they're using, it's just an example!).
These days, throw "Linux" or "wireless" or "business to business" in a press release and you will get a jump on the stock...
engineers never lie; we just approximate the truth.
Hear, hear... I made almost the identical comment in another thread. The 1-click issue is significant, but only to the/. population. Off-line, a patent on software may as well be a patent on thingamagicks...
What counts is that Bezos/Amazon risked everything to start e-commerce *first*; they made e-shopping accessible to John Q Public, and they helped legitimize the Internet... And I cannot think of any other group of people that has profitted more out of this than the/. readership, yet here we are arguing over Amazon's bottom line. What would happen if Amazon went under? How many investors would take a second look at their Internet portfolio? How many would pull out of the second- and third-tier dot-coms out there? and whose bottom-line that would hurt? not John Q Public's, but certainly John Slash Dotter's...
Give the guy a break...
engineers never lie; we just approximate the truth.
Is it me or is everybody forgetting, when Amazon.com started selling books? Was it 1995? 1996? Who else was on-line back then, trying to sell real stuff (not just shareware or porn) over the Internet? And to top it off, Bezos created a very good site, that is akin to the Library of Congress of the Internet ("err... what was that book I heard about the other day? Lemme go to Amazon and take a look"). Isn't Amazon.com really the point of reference of any e-commerce site (much like eBay is for auctions and Yahoo is for portals)?
Now, I do agree that Amazon is over-valued, over-stretched and over-exposed, and their business concept is due for a SIGHUP, but that doesnt mean that Bezos isn't a visionary and a pioneer.
engineers never lie; we just approximate the truth.
... IMHO, one of the reasons for companies to choose wacky names is that you never know what the name means in another language. A completely made up name may have less chances of being offensive in another country (and there's also the URL situation, I've been there...).
Some great examples:
Nova - someone beat me to it. Car name (Chevy); in Spanish is ~ 'no go'.
Osram - lighting bulbs, German I think. Apparently the word means something close to 'shit' in Polish.
Malaga - Car (SEAT) and also a place in Spain. Unfortunately in Greek means 'wanker'/'jerk'.
SEAT - it's actually a car company, not a chair;-)...
Somewhat related: Charlton Heston. In Greek the film distrubutors call him ~'Easton' as the correct pronunciation in Greek means, literally, 'shit on him';-)...
engineers never lie; we just approximate the truth.
Maybe I should clarify what I meant by bad SMP performance: of course, an SMP machine will be faster than a singe-CPU machine, but: a) you're not gonna get anywhere near (n-1)*100% (n=number of processors) speed increase. b) Not all utilities, helper progs, etc. know about or take into consideration SMP. I could live with these problems...
But the real kicker is this: No CPU affinity. I.e. the kernel doesn't stick a job to one CPU and let it crunch away there; instead, it dynamically load-balances all jobs around all available CPUs. This may not be too bad on a single-user desktop machine, but for a Beowulf cluster is horrendous because you loose all the advantages of CPU cache. Especially if, like yours truly, you are working with Xeons. They might as well be Pentiums, or for that matter Celerons.
Of course, you can always wait for 2.4.x or just use single-CPU machines, if you have the space, air conditioning and patience;-)...
engineers never lie; we just approximate the truth.
First of all, as some other posts said, don't go off building a Beowulf (that name has been overused IMHO) unless you know what you are building it for: clusters are built around the software you're gonna run on them, not the other way around.
If you plan to port MPP applications from a Cray or an Origin 2K, a Beowulf with and MPI port will most likely do what you want. If you are interested in a HA cluster, then we're really not talking Beowulfs; take a look at TurboLinux's TurboCluster distro.
If you want to throw lotsa CPU power to a problem that's not already MPP'd, a port to Mosix might worth your while, but investigate cautiously: Mosix does a good job (I am told) of process migration, but it doesnt migrate sockets yet, so it may effectively double your network bandwidth --this may not be a problem if your interprocess comm is minimal, or it might be a show-stopper. Do consider a port to MPI in this case: MPI is an industry standard and it works almost as well on a Cray as it does on a Beowulf.
Network communication is not as big a deal as it used to be: besides SCI, there's Myrinet (with OSS drivers and software too), Gigabit ethernet (also OSS drivers from some companies) and they all more or less work with Linux. Or you can go with the original Beowulf solution and bond Ethernet channels (i.e. make 2 NICs look and feel like 1 to the OS, almost doubling your capacity). It all depends on your application's inter-process communication requirements.
If you do decide on a Beowulf, heed these words: be carefule of SMP machines, at least this early in the game. Linux SMP is deficient at best --hopefully 2.4.x will solve it, but I wouldn't hold my breath. If you decide on an SMP machine, stay away from Xeon's as the extra cost will be useless right now --because of SMP problems with Linux, you might as well have a regular Pentium in there, or even some Celerons (hey, it'll buy better networking equipment;-). Also, do not plan to rely on NFS: Linux NFS is spotty when stressed by some high-bandwidth processes.
I guess the best advice would be: Don't go spending all your NSF money right away. Get 2-3 machines with some Fast Ethernet, set the thing up, port your software, make sure it works as well as you expected it to, THEN go spend the big $$$ on SCI, more nodes, etc. The biggest advantage of Beowulfs is *Freedom* as in flexibility;-)...
As the old maps said: "Beware: Monsters Here";-)...
engineers never lie; we just approximate the truth.
If the last moves by RH (Cygnus, Mozilla) mean that RH is focusing on the desktop --which is Linux's weak point-- doesn't that mean that they're might be an opening for someone else to tackle the server market?
And since, that's where Linux's current strengths lie (cheap server boxes), doesnt that leave a window for RH's rivals to catch up? (I am thinking TurboLinux)...
engineers never lie; we just approximate the truth.
I read the Discovery magazine articles... I had never heard of the Arrow before (and I am an aeronautical engineer;-) but it certainly looks like an impressive machine. I dunno if it qualifies as a hack though: it had a full design cycle, whilte the SR-71 was almost thrown together in 6 months by Skunk Works and held records for four decades.
Stories about great machines that never were get me depressed. Maybe that should be another feature... The XB-70 Valkyrie comes to mind, the (original) B-1A (not the lamer B-1B), the Russian equivalent (NATO nicknamed "BlackJack" I think) of the B-1, the Tu-144, the Tu-244, the Boeing SST... any other votes?
engineers never lie; we just approximate the truth.
I guess my definition of a "hack" requires innovation. I donot want to put down the flight and ground crew of Apollo 13, but they didn't really innovate. They had balls the size of a Saturn V and achieved the impossible, but that CO2 scrubber was already on the spacecraft. A great McGuyver hack, but in the Top 10? I am sure any veteran Army/Navy technician has at least one story like that...
I do agree with you though on the F-104 and the U-2; even the F-117 could qualify --it's amazing that thing flies at all;-)... Maybe Skunk Works should be the #1 hackers of all time.
engineers never lie; we just approximate the truth.
The SR-71 Blackbird. It may not be a "classical" hack, 'coz Lockheed's Skunk Works had an unlimited budget to throw at the problem, but considering the technology at the time, it kicked some ass... Some stats, for the non-plane freaks out there: * Total time it took to design it and built a prototype: 6 (or maybe 8?) months. There are software programs out there that took a lot longer than that;-) * It still (~40 years later) holds the title for the fastest *production* aircraft out there (err... at least non-classified;-) Mach 3.62 is nothing to sneeze at...
If you don't dare consider an airplane (i.e. a complete system) as a hack, consider the following: * The damn thing was almost entirely built of titanium alloy --only material available back then that could handle the temperatures involved. Problem: noone before was able to machine titanium. The Lockheed guys built an entire machine shop from scratch. * Titanium, as any metal, expands when heated: the planes had to have 'seams' in the wings that were closed when the sheetmetal expanded: the SR-71 leaked fuel (120 octane fuel) while parked on the runway! * The Pratt&Whitney (I think) folks had to come up with an engine that could change modes of operation in mid-flight: they made the first and only combination turbojet-ramjet engine. The Lockheed people had to make them work at any angle of attack. Yeah, it's esoteric, but the implementation is a tour-de-force to this day. * The poor Russians had no way to intercept these aircraft although they knew they were flying overhead and photgraphing everything (at Mach 3.62 the SR-71 could outrun any rocket or bullet at the time, and I it still can). So they build the all-steel Mig 29 (another great aircraft). But the -29 was too damn heavy to fly as high as the titanium-only -71, so the Soviets flew formations of -29s *under* the -71 to obstruct its camera's view...
I highly reccommend the excellent "Skunk Works" book to anyone impressed by this... I just don't think most of the/.ers will care;-(...
I guess I have to put in a computer hack as well. Hmmm... : FSP (yeah, that's an 'S').
engineers never lie; we just approximate the truth.
You know, that could have been a T3E --I am not the one running the benchmarks;-)... It sounds like your applications are much more I/O intensive than the stuff I am used to (computational fluid dynamics).
In our applications, a Linux-powered Xeon III-based Beowulf can compete with (slightly older) O2Ks *at the same number of CPUs*. Again, YMMV: as you say, the pipes in these SGIs are much, much fatter/faster than anything we can throw in on a Beowulf.
But the important thing, IMHO, is that the price/performance ratios of Beowulfs are now enabling a new class of applications, with *dedicated* hardware built to fit the software requirements rather than the other way around.
With the amount of money and man-hours being thrown into Beowulf enabling technologies (fast networking, maintenance schemes, HA, process migration, etc., etc.) I think we're approaching a shift along the lines of the old workstation/mainframe schism: cheap dedicated machines (Beowulfs now, workstations back then) versus very expensive, generic heavy iron (supercomputers now, mainframes back then). In the end, the largest mindshare (number of applications/developers) and the better price/performance ratio will win.
I am siding with the 'wulfs;-)...
engineers never lie; we just approximate the truth.
You are right, but you're also wrong. I agree that Linux SMP sucks --at least the 2.2 series. But a Quad Xeon III can offer equivalent performance (~40%) of a Cray T3D anyday, at an order of magnitude less $$$. The 12 Xeon III cluster I sysadmin can go head to head with a T3 easily, over Base 100 Ethernet! (we just got Gigabit Ethernet --give us a week;-).
At the low end, a compact Beowulf can easily compete with a Cray. Most scientific codes these days use MPI anyways, which can be ported easily from UNICOS or Irix to Linux. At the high end, or for specialized applications, we have ways to go. But, with better SMP support, and better support for fast networking (SCI, Gigabit Ethernet, Fibre Channel) Linux is getting there, fast. I wouldn't put money into SGI --except maybe to short them;-)...
engineers never lie; we just approximate the truth.
Even if the courts do penalize Microsoft in an appropriate and significant way, what do you think should be done to stop other companies waiting in the sidelines from becoming the next monopoly (Sun and AOL come to mind?).
Can the courts force *everybody* to open up their APIs and/or file formats? To me, this seems the only way to stop companies from using the Network Effect for their own gain and fortune...
engineers never lie; we just approximate the truth.
Can someone give a link? I am sysadmin on a PowerEdge 6300/550 with this RAID controller running RH6. Haven't had any RAID problems so far though...
One note: the Linux SMP performance doesnt surprise me one bit: we're running FPU benchmarks on several 6300s in a Beowulf config and we're losing a big chunk of Xeon III power because of Linux's lack of CPU affinity (2.2.10). If anyone has tried a 2.3.x kernel and got good results, please lemme know;-)...
engineers never lie; we just approximate the truth.
You're right; the idea of charging support for something that is build to make the users' life easier is a paradox. I believe that GPL'ed software has a place in the SOHO/Home market with some other business model: either charge for specialized "components" (say an analysis package for Gnumeric) or per use of higher-level functions.
I.e. give away the base software for free and charge people to use extensions/rarely used functions/"plus services".
Just my $0.002
engineers never lie; we just approximate the truth.
I agree wholeheartedly with your comments: I'd just like to point out to windows users out there a really nice feature on CRT, my favorite terminal program: highlite something, right click and then select "Open URL";-)...
engineers never lie; we just approximate the truth.
Just a coupla quick comments: Beowulf clusters are by no means or stretch of the imagination a technology that can be brought into an enterprise environment. It can save some research operations (granted, some by private companies) some serious bucks, but it cannot be deployed for serious, production-ready applications. At least not yet.
"Beowulf Technology" is little more than the notion of moving algorithms based on heavy message-passing from large, dedicated and very expensive machines to a large dedicated *network* of inexpensive machines. And how this is done, is truly a "hack" --a great one at that, but a hack nonetheless.
To truly be able to move such a solution to the enterprise space (and some, including myself, are trying to) you need a transparent, robust operating environment, including but not limited to, failover, job and batch control, large improvements in kernel performance (particularly wrt to SMP*), and probably a hundred other things.
These will come in time from some clustering technology (HA, Beowulf or Mosix or probably all of them), but we're still not there yet.
(*) e.g. I find it ludicrous that our 12-Xeon III cluster can not take advantage of all the monstrous CPU cache because the SMP kernel (2.2.15) keeps rotating jobs around the 4 processors on each node... it still beats a Cray YMP hands down, but that's not the point...
engineers never lie; we just approximate the truth.
FYI, some friends of mine have been in Tandy's op center. Whenever you make a purchase at RS and give them your name, the info is uploaded *live* to Tandy. Supposedly, their op room looks like 'War Games'...
engineers never lie; we just approximate the truth.
Hehe... which reminds me... back when I was in school and I could play with all the nice toys (I studied CFD which needs as much power as you can throw at the problem), I compiled XTetris on a Cray Y-MP, renamed the executable something arcane, like 3flowx or something, and played against the darn thing for ~1 hr for *one* game, which you can do, since XTetris could only use one CPU at a time, the Y-MP had 256 (?) pretty slow ones and the prog can only get a time-slice of that one too...
Trust me, heavy iron does not a good Quake machine make...
Oh, did I mention I was on the console of a Reality Engine 2 at the time? Some guy sat next to me, stared at the screen and pronounced: "How come the Reality Engine thing is so slow?";-)...
engineers never lie; we just approximate the truth.
There is GSM and GSM... the US version of GSM (I am a happy customer of Powertel in the Southeast, which is a GSM carrier) is using the FCC-mandated PCS frequency (1900MHz). Although it's the same codec as Euro-GSM, I can't take my Nokia GSM phone to Europe and expect it to work, since in Europe we're (I am European too) using 900MHz (most common) or 1800Mhz (newer) GSM. I've heard of GSM phones that can do both 900 and 1800 MHz for pan-Euro travelers but not any that can do 900, 1800 and 1900 (anybody know of any? maybe the Ericsson i888?).
IMHO, what would be really cool (but almost imposible) is to have all nations agree on one universal cellular service frequency band. That would at least take that barrier to compatibility away... not to mention killing off Iriduim, Globalstar and that pesky Teledesic;-)...
engineers never lie; we just approximate the truth.
I just thought I'd chime in coz this post reminded me of the twin WD Caviar A340s in my old 486: been running intermittently for ~6 yrs. At some point the first drive's power plug got broken (my bad). Soldered power supply wiring directly to its board and insulated them with blue-tac. Still working, no bad sectors. And did I mention that this box was shipped cross-country *3* times?
...when the Web took off back in '95 or abouts, Perl was a robust, well-understood language with tons of ready, freely available scripts to process enormous amounts of text. And since HTTP is text-centric, it was a natural first choice for CGI implementations. But I doubt it is any longer.
The Web and the technologies around it have matured enough to enable us to use custom-made tools for Web applications. For straight Web-only applications I'd go with the very impressive PHP4. For deployments that need to be tied in to a larger infrastructure (DBs, legacy code, uncoventional hardware), my personal choice would be the Apache Java Tools. Apache JServ, Apache Cocoon, and soon, Apache Jakarta should cover all your needs.
Both PHP4 and Apache Java have session management, pre-compilation to portable byte code, platform independence, and work well with most popular DBs (JDBC covers more of them though).
YMMV. As one of my ole engineering profs says: "50% of successful engineering is picking the right tool for the job" or something like that;-)
Slashdot Mirror
IMNSHO, the real power of the web is that middlemen can become "thinner" and "broader". I.e. they take less overhead on a transaction and can support more suppliers and customers. Witness Amazon.com, eBay, Yahoo, mp3.com, mysap.com, etc. Most people do not want to go into the trouble of finding all suppliers, comparing prices, etc., etc. Middlemen didnot (and will not) go away; they will just get smarter.
All that disintermediation stuff was so much Wired talk way back in 1996...
engineers never lie; we just approximate the truth.
I said this in the previous discussion, but I think it bears repeating: Give Bezos a break.
/. crowd was on the Net before 1995. For those of you that were, which was the first website to conduct serious "e-commerce", selling real things, not just shareware or porn? Amazon.com.
I assume that most of the
And Amazon not only has become the de-facto Internet bookstore (when you post a link about some obscure book, how many times do you go looking for it on bn.com?) but also the gold standard for e-commerce sites.
Amazon may not be the world's best or most profitable Web company (incl. patent rant here) but noone can deny that Bezos had a vision and worked hard to make it a reality. And by doing so he legitimized commerce over the Internet which is really the primary reason, us techies are doing better now than any other professional group.
I was hoping for better memory from this crowd...
engineers never lie; we just approximate the truth.
Errr... KTEL. They announced they were moving their *web servers* to Red Hat and got like a ~25% stock jump. It's equivalent to say eToys announcing they're gonna ship via USPS instead of UPS from now on (I dunno what they're using, it's just an example!).
These days, throw "Linux" or "wireless" or "business to business" in a press release and you will get a jump on the stock...
engineers never lie; we just approximate the truth.
Hear, hear... I made almost the identical comment in another thread. The 1-click issue is significant, but only to the /. population. Off-line, a patent on software may as well be a patent on thingamagicks...
/. readership, yet here we are arguing over Amazon's bottom line. What would happen if Amazon went under? How many investors would take a second look at their Internet portfolio? How many would pull out of the second- and third-tier dot-coms out there? and whose bottom-line that would hurt? not John Q Public's, but certainly John Slash Dotter's...
What counts is that Bezos/Amazon risked everything to start e-commerce *first*; they made e-shopping accessible to John Q Public, and they helped legitimize the Internet... And I cannot think of any other group of people that has profitted more out of this than the
Give the guy a break...
engineers never lie; we just approximate the truth.
Is it me or is everybody forgetting, when Amazon.com started selling books? Was it 1995? 1996? Who else was on-line back then, trying to sell real stuff (not just shareware or porn) over the Internet? And to top it off, Bezos created a very good site, that is akin to the Library of Congress of the Internet ("err... what was that book I heard about the other day? Lemme go to Amazon and take a look"). Isn't Amazon.com really the point of reference of any e-commerce site (much like eBay is for auctions and Yahoo is for portals)?
Now, I do agree that Amazon is over-valued, over-stretched and over-exposed, and their business concept is due for a SIGHUP, but that doesnt mean that Bezos isn't a visionary and a pioneer.
engineers never lie; we just approximate the truth.
Some great examples:
Nova - someone beat me to it. Car name (Chevy); in Spanish is ~ 'no go'.
Osram - lighting bulbs, German I think. Apparently the word means something close to 'shit' in Polish.
Malaga - Car (SEAT) and also a place in Spain. Unfortunately in Greek means 'wanker'/'jerk'.
SEAT - it's actually a car company, not a chair ;-)...
;-)...
Somewhat related: Charlton Heston. In Greek the film distrubutors call him ~'Easton' as the correct pronunciation in Greek means, literally, 'shit on him'
engineers never lie; we just approximate the truth.
Maybe I should clarify what I meant by bad SMP performance: of course, an SMP machine will be faster than a singe-CPU machine, but: a) you're not gonna get anywhere near (n-1)*100% (n=number of processors) speed increase. b) Not all utilities, helper progs, etc. know about or take into consideration SMP. I could live with these problems...
;-)...
But the real kicker is this: No CPU affinity. I.e. the kernel doesn't stick a job to one CPU and let it crunch away there; instead, it dynamically load-balances all jobs around all available CPUs. This may not be too bad on a single-user desktop machine, but for a Beowulf cluster is horrendous because you loose all the advantages of CPU cache. Especially if, like yours truly, you are working with Xeons. They might as well be Pentiums, or for that matter Celerons.
Of course, you can always wait for 2.4.x or just use single-CPU machines, if you have the space, air conditioning and patience
engineers never lie; we just approximate the truth.
First of all, as some other posts said, don't go off building a Beowulf (that name has been overused IMHO) unless you know what you are building it for: clusters are built around the software you're gonna run on them, not the other way around.
;-). Also, do not plan to rely on NFS: Linux NFS is spotty when stressed by some high-bandwidth processes.
;-)...
;-)...
If you plan to port MPP applications from a Cray or an Origin 2K, a Beowulf with and MPI port will most likely do what you want. If you are interested in a HA cluster, then we're really not talking Beowulfs; take a look at TurboLinux's TurboCluster distro.
If you want to throw lotsa CPU power to a problem that's not already MPP'd, a port to Mosix might worth your while, but investigate cautiously: Mosix does a good job (I am told) of process migration, but it doesnt migrate sockets yet, so it may effectively double your network bandwidth --this may not be a problem if your interprocess comm is minimal, or it might be a show-stopper. Do consider a port to MPI in this case: MPI is an industry standard and it works almost as well on a Cray as it does on a Beowulf.
Network communication is not as big a deal as it used to be: besides SCI, there's Myrinet (with OSS drivers and software too), Gigabit ethernet (also OSS drivers from some companies) and they all more or less work with Linux. Or you can go with the original Beowulf solution and bond Ethernet channels (i.e. make 2 NICs look and feel like 1 to the OS, almost doubling your capacity). It all depends on your application's inter-process communication requirements.
If you do decide on a Beowulf, heed these words: be carefule of SMP machines, at least this early in the game. Linux SMP is deficient at best --hopefully 2.4.x will solve it, but I wouldn't hold my breath. If you decide on an SMP machine, stay away from Xeon's as the extra cost will be useless right now --because of SMP problems with Linux, you might as well have a regular Pentium in there, or even some Celerons (hey, it'll buy better networking equipment
I guess the best advice would be: Don't go spending all your NSF money right away. Get 2-3 machines with some Fast Ethernet, set the thing up, port your software, make sure it works as well as you expected it to, THEN go spend the big $$$ on SCI, more nodes, etc. The biggest advantage of Beowulfs is *Freedom* as in flexibility
As the old maps said: "Beware: Monsters Here"
engineers never lie; we just approximate the truth.
Don't forget the damn tick (') on the (older I think) Sun keyboards! Damn annoying design...
Solution: if you use Zsh:
setoption sunkeyboardhack
;-)
engineers never lie; we just approximate the truth.
If the last moves by RH (Cygnus, Mozilla) mean that RH is focusing on the desktop --which is Linux's weak point-- doesn't that mean that they're might be an opening for someone else to tackle the server market?
And since, that's where Linux's current strengths lie (cheap server boxes), doesnt that leave a window for RH's rivals to catch up? (I am thinking TurboLinux)...
engineers never lie; we just approximate the truth.
I read the Discovery magazine articles... I had never heard of the Arrow before (and I am an aeronautical engineer ;-) but it certainly looks like an impressive machine. I dunno if it qualifies as a hack though: it had a full design cycle, whilte the SR-71 was almost thrown together in 6 months by Skunk Works and held records for four decades.
Stories about great machines that never were get me depressed. Maybe that should be another feature... The XB-70 Valkyrie comes to mind, the (original) B-1A (not the lamer B-1B), the Russian equivalent (NATO nicknamed "BlackJack" I think) of the B-1, the Tu-144, the Tu-244, the Boeing SST... any other votes?
engineers never lie; we just approximate the truth.
I guess my definition of a "hack" requires innovation. I donot want to put down the flight and ground crew of Apollo 13, but they didn't really innovate. They had balls the size of a Saturn V and achieved the impossible, but that CO2 scrubber was already on the spacecraft. A great McGuyver hack, but in the Top 10? I am sure any veteran Army/Navy technician has at least one story like that...
;-)... Maybe Skunk Works should be the #1 hackers of all time.
I do agree with you though on the F-104 and the U-2; even the F-117 could qualify --it's amazing that thing flies at all
engineers never lie; we just approximate the truth.
...someone had to give at least one:
;-) ;-) Mach 3.62 is nothing to sneeze at...
/.ers will care ;-(...
The SR-71 Blackbird. It may not be a "classical" hack, 'coz Lockheed's Skunk Works had an unlimited budget to throw at the problem, but considering the technology at the time, it kicked some ass... Some stats, for the non-plane freaks out there:
* Total time it took to design it and built a prototype: 6 (or maybe 8?) months. There are software programs out there that took a lot longer than that
* It still (~40 years later) holds the title for the fastest *production* aircraft out there (err... at least non-classified
If you don't dare consider an airplane (i.e. a complete system) as a hack, consider the following:
* The damn thing was almost entirely built of titanium alloy --only material available back then that could handle the temperatures involved. Problem: noone before was able to machine titanium. The Lockheed guys built an entire machine shop from scratch.
* Titanium, as any metal, expands when heated: the planes had to have 'seams' in the wings that were closed when the sheetmetal expanded: the SR-71 leaked fuel (120 octane fuel) while parked on the runway!
* The Pratt&Whitney (I think) folks had to come up with an engine that could change modes of operation in mid-flight: they made the first and only combination turbojet-ramjet engine. The Lockheed people had to make them work at any angle of attack. Yeah, it's esoteric, but the implementation is a tour-de-force to this day.
* The poor Russians had no way to intercept these aircraft although they knew they were flying overhead and photgraphing everything (at Mach 3.62 the SR-71 could outrun any rocket or bullet at the time, and I it still can). So they build the all-steel Mig 29 (another great aircraft). But the -29 was too damn heavy to fly as high as the titanium-only -71, so the Soviets flew formations of -29s *under* the -71 to obstruct its camera's view...
I highly reccommend the excellent "Skunk Works" book to anyone impressed by this... I just don't think most of the
I guess I have to put in a computer hack as well. Hmmm... : FSP (yeah, that's an 'S').
engineers never lie; we just approximate the truth.
You know, that could have been a T3E --I am not the one running the benchmarks ;-)... It sounds like your applications are much more I/O intensive than the stuff I am used to (computational fluid dynamics).
;-)...
In our applications, a Linux-powered Xeon III-based Beowulf can compete with (slightly older) O2Ks *at the same number of CPUs*. Again, YMMV: as you say, the pipes in these SGIs are much, much fatter/faster than anything we can throw in on a Beowulf.
But the important thing, IMHO, is that the price/performance ratios of Beowulfs are now enabling a new class of applications, with *dedicated* hardware built to fit the software requirements rather than the other way around.
With the amount of money and man-hours being thrown into Beowulf enabling technologies (fast networking, maintenance schemes, HA, process migration, etc., etc.) I think we're approaching a shift along the lines of the old workstation/mainframe schism: cheap dedicated machines (Beowulfs now, workstations back then) versus very expensive, generic heavy iron (supercomputers now, mainframes back then). In the end, the largest mindshare (number of applications/developers) and the better price/performance ratio will win.
I am siding with the 'wulfs
engineers never lie; we just approximate the truth.
You are right, but you're also wrong. I agree that Linux SMP sucks --at least the 2.2 series. But a Quad Xeon III can offer equivalent performance (~40%) of a Cray T3D anyday, at an order of magnitude less $$$. The 12 Xeon III cluster I sysadmin can go head to head with a T3 easily, over Base 100 Ethernet! (we just got Gigabit Ethernet --give us a week ;-).
;-)...
At the low end, a compact Beowulf can easily compete with a Cray. Most scientific codes these days use MPI anyways, which can be ported easily from UNICOS or Irix to Linux. At the high end, or for specialized applications, we have ways to go. But, with better SMP support, and better support for fast networking (SCI, Gigabit Ethernet, Fibre Channel) Linux is getting there, fast. I wouldn't put money into SGI --except maybe to short them
engineers never lie; we just approximate the truth.
Even if the courts do penalize Microsoft in an appropriate and significant way, what do you think should be done to stop other companies waiting in the sidelines from becoming the next monopoly (Sun and AOL come to mind?).
Can the courts force *everybody* to open up their APIs and/or file formats? To me, this seems the only way to stop companies from using the Network Effect for their own gain and fortune...
engineers never lie; we just approximate the truth.
Can someone give a link? I am sysadmin on a PowerEdge 6300/550 with this RAID controller running RH6. Haven't had any RAID problems so far though...
;-)...
One note: the Linux SMP performance doesnt surprise me one bit: we're running FPU benchmarks on several 6300s in a Beowulf config and we're losing a big chunk of Xeon III power because of Linux's lack of CPU affinity (2.2.10). If anyone has tried a 2.3.x kernel and got good results, please lemme know
engineers never lie; we just approximate the truth.
You're right; the idea of charging support for something that is build to make the users' life easier is a paradox. I believe that GPL'ed software has a place in the SOHO/Home market with some other business model: either charge for specialized "components" (say an analysis package for Gnumeric) or per use of higher-level functions.
I.e. give away the base software for free and charge people to use extensions/rarely used functions/"plus services".
Just my $0.002
engineers never lie; we just approximate the truth.
I agree wholeheartedly with your comments: I'd just like to point out to windows users out there a really nice feature on CRT, my favorite terminal program: highlite something, right click and then select "Open URL" ;-)...
engineers never lie; we just approximate the truth.
Just a coupla quick comments: Beowulf clusters are by no means or stretch of the imagination a technology that can be brought into an enterprise environment. It can save some research operations (granted, some by private companies) some serious bucks, but it cannot be deployed for serious, production-ready applications. At least not yet.
"Beowulf Technology" is little more than the notion of moving algorithms based on heavy message-passing from large, dedicated and very expensive machines to a large dedicated *network* of inexpensive machines. And how this is done, is truly a "hack" --a great one at that, but a hack nonetheless.
To truly be able to move such a solution to the enterprise space (and some, including myself, are trying to) you need a transparent, robust operating environment, including but not limited to, failover, job and batch control, large improvements in kernel performance (particularly wrt to SMP*), and probably a hundred other things.
These will come in time from some clustering technology (HA, Beowulf or Mosix or probably all of them), but we're still not there yet.
(*) e.g. I find it ludicrous that our 12-Xeon III cluster can not take advantage of all the monstrous CPU cache because the SMP kernel (2.2.15) keeps rotating jobs around the 4 processors on each node... it still beats a Cray YMP hands down, but that's not the point...
engineers never lie; we just approximate the truth.
FYI, some friends of mine have been in Tandy's op center. Whenever you make a purchase at RS and give them your name, the info is uploaded *live* to Tandy. Supposedly, their op room looks like 'War Games'...
engineers never lie; we just approximate the truth.
Hehe... which reminds me... back when I was in school and I could play with all the nice toys (I studied CFD which needs as much power as you can throw at the problem), I compiled XTetris on a Cray Y-MP, renamed the executable something arcane, like 3flowx or something, and played against the darn thing for ~1 hr for *one* game, which you can do, since XTetris could only use one CPU at a time, the Y-MP had 256 (?) pretty slow ones and the prog can only get a time-slice of that one too...
;-)...
Trust me, heavy iron does not a good Quake machine make...
Oh, did I mention I was on the console of a Reality Engine 2 at the time? Some guy sat next to me, stared at the screen and pronounced: "How come the Reality Engine thing is so slow?"
engineers never lie; we just approximate the truth.
There is GSM and GSM... the US version of GSM (I am a happy customer of Powertel in the Southeast, which is a GSM carrier) is using the FCC-mandated PCS frequency (1900MHz). Although it's the same codec as Euro-GSM, I can't take my Nokia GSM phone to Europe and expect it to work, since in Europe we're (I am European too) using 900MHz (most common) or 1800Mhz (newer) GSM. I've heard of GSM phones that can do both 900 and 1800 MHz for pan-Euro travelers but not any that can do 900, 1800 and 1900 (anybody know of any? maybe the Ericsson i888?).
;-)...
IMHO, what would be really cool (but almost imposible) is to have all nations agree on one universal cellular service frequency band. That would at least take that barrier to compatibility away... not to mention killing off Iriduim, Globalstar and that pesky Teledesic
engineers never lie; we just approximate the truth.
I just thought I'd chime in coz this post reminded me of the twin WD Caviar A340s in my old 486: been running intermittently for ~6 yrs. At some point the first drive's power plug got broken (my bad). Soldered power supply wiring directly to its board and insulated them with blue-tac. Still working, no bad sectors. And did I mention that this box was shipped cross-country *3* times?
...when the Web took off back in '95 or abouts, Perl was a robust, well-understood language with tons of ready, freely available scripts to process enormous amounts of text. And since HTTP is text-centric, it was a natural first choice for CGI implementations. But I doubt it is any longer.
;-)
The Web and the technologies around it have matured enough to enable us to use custom-made tools for Web applications. For straight Web-only applications I'd go with the very impressive PHP4. For deployments that need to be tied in to a larger infrastructure (DBs, legacy code, uncoventional hardware), my personal choice would be the Apache Java Tools. Apache JServ, Apache Cocoon, and soon, Apache Jakarta should cover all your needs.
Both PHP4 and Apache Java have session management, pre-compilation to portable byte code, platform independence, and work well with most popular DBs (JDBC covers more of them though).
YMMV. As one of my ole engineering profs says: "50% of successful engineering is picking the right tool for the job" or something like that
n