Shutdown Computer vs. Shutdown the computer is exactly the same thing. The idiots who think this guy deserves +5 need to shit in the toilet, dunk their head in it, flush, and go back to the English classes they forgot since they've evidently been writing too much code.
*shrug* I didn't give it +5, and I wouldn't have. But I would have given it +2 easy. Maybe +3 if it cought me in the right mood.
If they don't, they're likely to end up writing their own menu options such as this and having a younger generation of/.ers insulting them.
I hope not. My only GUI apps are OSS. I expect a patch, not a bitch. Especally on things where even non-coders can find the quoted string and change it.
Do you really think the menus on the remote computer will change to say 'shutdown remote computer'?
Do? Or should? I think they should say "shutdown {computername}", or "shutdown {username}'s computer". I expect Microsoft just has them saying "shutdown computer".
The definite article is fairly obvious from the context - why waste screen space and user time on 'Shut Down *The* Computer', when it's clear that you are not shutting down any other computer?
Are you sure? What about Remote Desktop Connection? Can you be turning off/rebooting some else's computer?
If they sell the product to you, or present it to you as a gift, then no they have no rights to control your use of it whatsoever. This is pretty much the entire point.
I don't think they lose control just because they sell it to you. I think they lose control because they sell it to you without making you sign a terms of service contract. If you do sign such a contract I think you may be legally (or maybe just morally) bound to follow it.
I don't know about shrinkwrap agreements (where the terms are presented after you buy). In the past they were not legally binding, with the DCMA they might be. However I don't think they are morally binding. In fact I think they are morally repugnant.
And I definitly think that a shrinkwrap agreement on software you never ran (say, because the "forgot" to ship the Linux or FreeBSD version) isn't going to be legally binding at all!
O.B. trick. Don't like a shrinkwrap agreement? Change all the terms you disagree with, inital the changes, have it noterised, and send the terms back. This almost certonally won't get the terms change the way you want, but may legally void the shrinkwrap agrement and return you to the default rights of a UCC purchase (i.e. you probbably can't give copies to friends, but you can probbably reverse engenear, make backups, and all the other things people have been doing to software for 20 years, and machanical devices for centries).
If rackspace can cost $1000/U/month, then don't ya think it would be a good idea to find a better site for the boxen ?
Multiple OC-12s are rather expensave arn't they:-) Seriously that figure is 3 to 4 years old, and for a high bandwidth location with forced air cooling and 3 day battery backup. I don't know what current figues are, I would expect them to be much lower in many places.
My original point was merely that some places are really really expensave, and paying an extra $3500 to save 2U of rack space has a payback time mesured in months (two in this case).
Heck, if that's what rackspace is going for these days, I'm converting my apartment into a colo facility =)
You need a lot of bandwidth coming in to the facility, from multiple providers. Or at least on multiple fibre bundles. Better still if you are centrally located (to minimise lantency to both coasts). Lots of good places in the center of the USA to get land cheep, not so hard to find major switching centers (cheeper circuits), but kinda hard to find places close to multiple telcos.
The other route to being a high cost colo is to have lots of ISPs show up where you are (like MAE-east/MAE-west), but those are hard to start, you normally have to bring a bunch of bandwith to the party and peer cheep for a while...
And battery and/or gennerator back up is a must. Forced air cooling is very good. Humidity control is a must (normally the cooling does that for you, but just in case you want a alaska colo, you still need to control moisture even if cooling is taken care of by nature). Apartments tend not to provide much cooling, so ration out a big cunk of the bedroom for a chiller. They also don't normally have raised floor/drop ceiling, so chop off some more space to install them (or use all ladder racks and overhead cable runs, but you still need space for the cable runs).
You'll be better off renting wharehouse space... (esp if you can find it close to a CO).
Also rember Hurricane Eletric does way cheep rack space, and other places too...
I could build a 3U Linux or BSD machine with a couple Promise IDE cards in it and 4 80 gig IDE drives for about $1200 or so. Maxtor wants $4500!
Yep. It is cheper because doing it yourself is almost allways cheaper. Plus it is chepaer because it is three times the size. Sometimes that is a great tradeoff. Other times rack space is expensave (sometimes MORE then $1000/U/month). Frequently expensave enough to make a 1U vs. 3U diffrence worth many $1000s in a single year. Sometimes in just a month. Try pricing telco colos someday.
Of corse there are lots of places where rack space is dirt cheap, and using 3U vs. 1U to save a little money is a no-brainer.
The big problem is space tends to be an inelastic commodity. Run out of disk space? Buy more! Run out of space space? Oops, it takes months to set up a new lease and get it connected and... In fact just evecting people from their office/cubes so we can grow the machine room takes time...
So space starts out cheap, and gets incresingly costly as you consume it. Sometimes you have to go back and replace the 3U disk boxes with 1U disk boxes, and the 3U SPARCs with 1U SPARCs, and sell off the old ones (if they retain any value). At least the new stuff costs way less then it would have when you first did the build-out...
Unfortunately, I haven't seen anyone give an alternative. The original advisories just say "This is bad coding
practice, don't do it" but don't offer any alternatives and point to documents which RECOMMEND DOING
EXACTLY THIS.
Use snprintf(3), or the other "n" variants. Like snprintf(buf, sizeof(buf)-1, format,...). The -1 isn't needed, but it makes me feel better (whoever implments the libc function might not have read the spec close enough).
Or in C++ use a stringstream which will expand itself as needed. Or on 4.4BSD systems C programmers could use funopen(3) to get the same sort of effect, but not so portably.
And there is allways the "make my buffer one page big followed by a red zone", but that isn't hugely portable, and opens a DoS attack.
To be honest I use snprintf (and vsnprintf), and havn;t needed to do the like in C++ yet. But all of that stuff should work. And remember to put some overflow tests in your test vectors.
I've read that USB keyboards and mice are supported at least by linux, which is a good first step, but the vast majority of USB devices are things like cameras and scanners.
My FreeBSD box syncs with my Visor (Palm clone) over USB. I have a working USB Ethernet (intended for use on an i-opener, or similar unit which I havn't had the time to hack up). I expect my new digital camera will as well since I got sources to a Linux program that does it. I just have to find out the diffrences between raw USB access on Linux and FreeBSD.
According to the (BSD) usb mailing list I'm on many USB modems work. As do many USB mass storage devices. Not that USB1.1 is a fast access path to mass storage, still slower then 1987's async SCSI.
I don't know how closely sync'ed the BSD and Linux USB stacks are. I think the Linux one started with NetBSD's (a fair and good idea). FreeBSD's track's NetBSD's.
you're not using "working capital" accurately. Factories are long term investments. Working capital is short term.
Your right, at least several of the examples I gave were not working capital. I do beleve the rest of my post is correct. As is yours.
Most investors don't have a steadily growing stream of cash to keep pumping in.
Does that include VC houses like Kliner-Perkens or are VC houses actually a rareity? (I've only been trhough one IPO, and I wasn't even remotely involved with raising VC money, other then making the componies products work while that was going on)
I really don't think that companies should be filing for an IPO just to get out of debt, in fact I think that I'd be more comfortable investing in a company that has historically made some money, not lost it
One of the big reasons a compony goes public is to get some working capital to make investments (factories, expensave talent, marketing blitz) that will make an even bigger profit. A profitable compony frequently has less reason to IPO then a not-so-profitable one. One of the other big reasons is because tipically Venture Capitalists get to make some of those calls in exchange for money when the compony really really really needs it (like to go from a literal garage start-up to a underfunded low-rent office park start-up; or to go from "I got an idea!" to "It works in the lab!").
As for investing in a compony that has made money, well that doesn't mean they will make more in the furure. More importantly, it doesn't mean it will make more money then people are currently betting on which is what stock value really is. You would rather invest in a compony that is about to make a lot more money then people think it is. Maybe one that is just shipping a product that really is as good and revolutionary as people say (it's a good excuse to buy products the day they hit the market -- to check for good stock investments...a lession I should have learned with the Zip drive, and applyed to the AMD K7!).
P.S. I am not a profesional. This advice may lose you money. I'm not responsable. The risk is yours. Blah blah blah.
People don't even think of mobile phones as computers (mostly because...
they don't go wrong [...]
Except when they:
Reboot for no reason (Nokia 21xx, Nokia 9000, Samsung SCH-3500)
Decide that the phone number you just typed in should be deleted just because you havn't seen fit to stroke the phone's ego for two minutes (SCH-3500 -- called 411 on the lndline, typed the number in on the moble, made the call on the landline, talked very breifly and as I was about to tell the moble what to call the number it vanished...as I was watching)
Go in to extreame power suck mode and eat a day's worth of batts in a half hour (Nokia 9000)
Had a really really crappy interface for deleting text messages (SCH-3500 -- select message, press menu, press 2 OR 3 depending on the whim of the phone, then get kicked out of text message read mode, press 3 to get back, find next message, press OK, press menu, press 2 or 3...; no way to "delete all" or "delete all below here" or "delete any older then X days"
Did I mention the random reboots?
Or the just plain poweroffs (Nokia 21xx)
Or the reboots? During calls?
Or saying there is signal, but not being able to recieve a call?
Or the lame ergonomic of having the outside volume control turn the ringer off (or set to one beep mode) if it gets hit a random number of times (SCH-3500 -- and yes it was unlocked, but it's just dumb, and I normally lock it so it took five calls before I figured out why it had done that!)
Did I say anything about the reboots? I can't recall if I did. Better mention it again.
The lack of any really cool features a computer would have, like "tunnel simulation mode". On a long call you can't politely end? Press the tunnel button and your signal gradually degrades and cuts off. No need to find a elevator to escape!
Your chances of getting on as a director for a Star Wars game - or any PC game for that matter - is slim.
As a director? I have no idea. As a starting programer it didn't seem hard to get a starting programer-job at MP Games (Microprose's now defunct coin-op sister componay) almost ten years ago.
These companies put out great games, but you have to wonder about the sacrifice some of these programmers and artists have to make.
Actually I can rather assure you that I do:-)
It was a lot of work. And there was a lot of collabration with the artists and other coders, so everyone keeping similar hours was pretty important, which makes for painful mornings after an all-nighter.
Then again working for a little start up (UUNET) was also very demanding. It didn't have the same pain-in-the-ass be-in-in-the-morning requirements, but there was a lot of hard work there. I'm not sure there are worthwhile jobs that arn't demanding.
They always show programmers firing nerf guns around on Fridays in gaming mags, having fun, but they never show them sitting and staring at the same DirectX screen for 14 hours on end.
Well nobody shows that stuff. It isn't photogenic. If you ask to see it when you go interview you'll get too (you should at least walk by it even if you don't ask -- be extreamly wary of off-site interviews! Rember you are interviewing them as much as they are interviewing you...at least if you are doing it right!).
And every place has it's own little steam-blow-offs. Sometimes nerf wars (scheduled or not -- my Wildfire is loaded and ready monkey boy, bring it on). Sometimes gaming. At game componies there is a lot of gaming (the caf had lots of board games in it, and a few CoinOp clasics, plus our own of corse), inclusing "field trips" (either to see your game in test, or just to playtest the others...pity they didn't let me file an expense form for that, but at least it was gaming on work time).
But other componies have fun things as well. Nerf wars are popular. Cube decorating, and reconfiguring isn't bad. Nor is "lets see if we can get on the roof". Some places will have really long lunches (when far enough from deadlines at least).
Few componies are utterly devoid of any sort of fun at all. (well, maybe there are, I've only worked at about five places plus some consulting, so and other then consulting all in the MD/VA area)
So now you know why so much of free software is un-usuable.
No, much free software is unusable for the same reason much comercial software is unusable. It just that most software is unusable. If you disagree go use a peice of comercial software at random. Don't forget all the in-house apps that need a ton of instruction to use, stupid instructions like "don't hit 'all packages', that just crashes, select them all by letter".
Lots of comercial crap is worthless. Lots of free crap is worthless. It is probbably easier to find the worthless free crap, but only because it is free, not because there is more of it.
Don't forget lots of free stuff is pretty damm nice. Apache. OpenBSD. GCC. The Smithsonian. Love. And some claim Linux even.
No, no, no, god no. Read last months' C++ users Journel, or Strousup (damm those forginers and their hard to spell names!).
As "NULL References" are not defined, making one can do anything. They can crash you right as they are created. They can crash you when you try to take their address.
What you show works on some implmentations, but is a programming error, every bit as much as using memory after you free it is (which normally works until you allocate more memory). But any garbage collected implmetation, or an implmentation that does anything odd with references (perhaps because the CPU is odd, or the ABI) can show the breakage, even if you try to check for the NULL reference, because the bug is not the use of the NULL reference, but the mere existance, the bug starts when you create the reference, and an implmentation can crash at any moment after (or as) you do it, attempts to find and not use the NULL reference are just masking the bug, and not allways masking it.
Expect a future compiler rev to break such code. Expect porting the code to make the bug come live.
In other words, you are living the life of a VAX programmer that assumes '0 == *NULL' works, and will forever work (and that 'strlen(NULL) == 0' works as well). Soon everyon will want your code ported to that hot new Sun3 platform, and you will have to find and fix all those damm bugs.
It duplicates parts of the standard library. For programmers who already uses the standard library, it means you have to deal with two string classes, two vector classes, etc. If nothing else, that is ugly
Eh? I wrote a whole (modestly) big Gtk-- program and never used strings other the the standard C++ string, and standard C char*'s. Never used a vector other then the STL's. Gtk-- might (or might not) have it's own string and vector classes, but I definitly didn't do anything to avoid their use, and yet I'm also not using them.
Ah, but the 'func(*(int*)NULL);' part is illegal in and of itself (in C and C++), or more accuratly undefined. Most implmentations of C++ references just lazy evaulate the core dump:-)
There really is no such thing as a NULL reference really means
There is no language defined way to get them (although as you show, most implmentations will let you get one)
There is no language defined way of testing for them
When one is found the bug is clearly creating the reference (i.e. the dereference of the NULL pointer by the caller), and never ever in the code attempting to use the reference
I don't find those things to be a huge advantage for references. Nor do I find that they make my code cleaner looking (except as a return type so I can have lvalue functions). I find it a modest disadvantage that I don't have the "out of bound" signaling path of passing a NULL to mean something "specail" (default value, or skip that part, or something).
Left to my own devices I use pointers. Then again I kinda grew up on C (and APL, but we'll not mention that again). On the other hand I really really love the STL. One of the few saving graces of C++. And a huge one at that. Frequently even signifagantly faster then C.
Why should Gtk-- team and Gnomers have to study Qt to do their own C++ ?
Because if someone else has attacked the problem before you are rock stupid to not look closely at what they have done so you can see what they did right, or wrong. So what you know might be good to copy (if it fits into your fraework), and what to avoid at all costs.
Do you thnk AMD isn't one of the first buyers of Intel CPUs? That they don't cut them open with surgens saws as soon as possiable?
Do you think that Linux kernel hackers arn't looking at NetBSD's USB system?
Do you think McDonald's marketing teams don't eat at Burger King?
That GM doesn't buy Ford cars and take them apart?
How much of Qt have they copied? There are certain ethics to reverse enginering such as you don't look at the source even if it is available.
I doubt any code was copied. Qt is a toolkit. Gtk-- wraps an existing toolkit. The slot/connection model in Gtk-- is all done within C++ while Qt uses a preprocessor, which makes Qt programs "not quite C++", which I think would sometimes be a pain (not an insurmountable obsticle, but still a pain). Gtk-- makes templates fit the task, which seems to work quite well. I'm not sure if Qt could have done the same with the state of C++ compilers when they started. If I were to do it now I would definitly do it the Gtk-- or Inti way. Being second sometimes has huge advantages. The rest of the slot/connection model is similar in design between Qt and Gtk--, but it is also similar to Smalltalk and other systems that have come before (and I would claim both did a good thing copying a previous succsuful solution).
What's wrong with plain old C with the object system imposed by Gtk+? I thought that it was Stallman's decree to use plain old C allocating objects with pointers, which works for most people.
If you attack a strawman at least attack the right one. I don't think Stallman has a lot to do with Gtk+, I think it was Havroc or one of the other GTK+ develeprs who asserts that GTK+ being in C is a major huge good thing.
Personally I could give a crap less what language it is written in if it (a) works, and (b) has good bindings to the language I want to use. Now I realise that things never actually allways work, so I will avoid using a tool if it is written in Intercal, but just because I wouldn't pick C to do an OO program in doesn't mean I'll avoid it if, if it has good C++ (or Java, or whatever) bindings.
More importantly I think the "offical" argument has allways been that writing in C makes it easyer to have Python/Perl/Java/Whatever bindings then if it were in C++. I'm not positave I agree. I also don't think I care a whole lot. If I write a GUI app I'll probbably write it in C++ (or Java, but if I do it in Java I'll use Swing anyway). But the offical argument has never been "C rulz, C++ blows goats"
How many apps have been written with Gtk-- to date? Would some of those who have used Gtk-- in applications which have progressed beyond the pre-alpha stage care to comment?
Beats me how many have passed alpha. Bu I can say w3juke was really really far easyer to code up then any toolkit I had used in the past (note I havn't used Qt), including some non-Unix ones. But w3juke has not passed alpha. Not really because of the toolkit, but because of lack of documentation, and available time, and the abundent lazyness of the programmer.
And of corse now I want to come up with another smal task to go code up in Inti just so I know how it compairs...
both the hardware and software were revolutionary and represent one of the biggest missed opportunities in the industry
I never owned a NeXT, but the University I went to bought into it big time, so I did spend a bit of time playing with them.
The software was revolutionary. In many ways we are still catching up. Definitly in having a user friendly Unix we are still catching up. As little as I liked Objective C's performance, it did make things easy. Nice, nice, nice software. Good choice of bundled apps for an academic market too.
The hardware was not stunning at all.
If you sat it next to it's Sun's boxes of the era it was dog slow (if you ran SunTools at least -- if you ran X the display on the NeXT xould catch up). Both had equivolent resultion. The NeXT has 4bit (8bit?) grey. The Sun had either 1-bit deep mono, or 8bit color (which could do 8bit grey) depending on which graphics option you got. The NeXT had a 68030 (68040? 68020? mmmm, maybe the 68030) at something like 20Mhz. Sun had recently come out with the SPARC 1+ the follow on to the first desktop SPARC, 25Mhz I think. But much much much faster. Doing a whole lot more per cycle then the Moto part. The SPARC didn't feel a little faster it felt a lot faster.
The few hardware features the NeXT had and nobody else did were not particurlay well recieved. The "floptical" was a bit fragile, and most people only had the one and no HD so it wasn't removable media, it was just a slow hard drive. The DSP never got used for anything intresting, the promised high speed modem was extreamly late, and I not sure it ever worked. About the only inovatave hardware feature I remember on the NeXT was the cubes looked way cooler then even the new SPARC pizza boxes.
But the software, oh man was that software ahead of it's time...
Then again, it could have been running a java morphing layer too. If that's the case, then it looks like my
reservations have been answered and I'm going to start getting excited about this.
It was. Or at least the use of a normally reserved opcode to tell the morpher to switch instruction sets seemd, and the rest of the explination of the demo (like them saying there was a picoJava morpher) seemed more consistant with two morphers at once then running a x86 JVM.
There is (as far as I know) only the one production morpher. Which makes sense. Don't dilute your manpower until you have some mesure of success....
2, can the Crusoe chip execute code written in difference instruciton sets at the same time (well, as much as any other popular single-pipe processor executes multiple threads), eg can it run Crusoe-optimized x86 Debian running an application compiled for Alpha Linux?
One of the demo's they had at unveling was a copy of DOOM, and x86 executable with part of the inner loop replaced by picoJava. It seemed to run that pretty quick, so it can switch instruction sets dyanmically. There is undoubtly some penality, but it apparently isn't huge.
There is no evidence that it can run Alpha code particurally fast. It has 48 (or 40?) registers internally. More then enough to do software registr renaming, and have operatinog regesters left when emulating a 4 register (10 if you count segment registes, the PC and flags) machine. The Alpha has 32 user visable registers (or is it 32int plus 32FP?). The 21264 has 80 renaming registers (invisable to the programmer) in two banks. The Crosue would have far less, so could be expected not to be able to hide as much latency.
Oh, and since the Crosue is designed to emulate the 32bit x86, it is not likely to have 64 bit registers, or a 64bit ALU. And it's FPU certonally is geared towards the x86 quirks (80 bits, almost IEEE, but not quite).
I wouldn't want to try to emulate the Alpha on a Crosue. It could probbably do a great 68000 (or the CPU32 in the palm, which is almost a 68000). Maybe even a good ARM, or V8 SPARC (these are pushing it though). Could definitly do a good JVM. Would do a great 6502, or Z80.:-)
But not a Alpha. Not a V9 SPARC. Not the 64bit MIPS. Not the IA64.
Um, their Cray division did alot of the work for the O2000! In fact at release the >64CPU configs were only avail from cray. Oh, and the frame to frame comm channel? It's named the "CrayLink".
Nice machines though, even if a bit long in the tooth (the O2000 is fourish years old, the O3000's should be announced anytime now, go look at comp.arch)
If you want economical, "Beowulf" clusters are the way to go now a days.
Sure, if you need very little communication between machines Beowulf is great, and the O2000's expensave comms (the xbow and craylink) are waisted. If you need a lot of comm, but not a lot of com a O2000 is great. If you need a lot-lot of comm maybe you are out of luck until the O3000, HP SuperDome, or IBM Power5 show up.
Quick MP break down:
NORMA - NO Remote Memory Access - Beowulf is a NORMA, to get at memory on other systems you need to make OS calls, or at least use really expensave mmap'ed (NFS) files (i.e. non local memory costs 1000x more then local memory to access).
NUMA - Non-Uniform Memory Access - remote memory costs maybe 10x more then local memory access, and caching has to be handled specally (i.e. each system has to know when to flush it's cache "magically"). Very few examples, some IBM research machines do this.
ccNUMA - Cache Coherent Non-Uniform Memory Access. Remote memory access costs maybe 10x local, but the cache's work. Most large multip-processor machines work this way. The O2000, the Sun E10000. The E10000 has much less then a 10x penality for remote memory, but local memory costs more then the O2000's local accesses. On both the OS can move pages from CPU board to CPU board depending on access. The E10000 comes closest to giving the impresion that it is a UMA machine (and the O2000 isn't bad at it)
UMA - Uniform Memory Access. There is no remote memory, or no penality for accessing it. A tipical multiprocessor PC works this way. Tipically easy to build for small numbers of CPUs, incresingly impossable for larger numbers (or useless - you could make a UMA for 1024 CPUs by having extra shitty memory access times, but there is no known way to make one with good access times, even if you had a literally unlimited budget!).
A NORMA (or better) is great for raytracing, crypto cracking, and the like. A UMA is great for N-body simulation (with large N). I wouldn't want to track the flow of air molicules over a wing with a Beowulf, but I wouldn't want to pay for a ccNUMA if I was "just" running PR-Renderman.
I don't know what you remember about your conservation of energy, but any energy that goes into your computer comes out as heat. (Exceptions: the monitor & speakers make heat somewhere other than inside the computer, though not as much as they themselves have to disipate.)
Clearly I don't remember enough. But I was actually intending to talk about how the CPU's waste heat is in a very small area (even the "hot" desktop CPUs don't disapate as much heat as the lightbulbs in my room, but I don't need cooling fans for my bulbs because they have more surface area...). The drive generates it's heat over a large enough area that it may need no fan. And as you said the LCD (or really it's backlight) make a lot of heat, but not anywhere that the laptop has to worry about it.
Slashdot Mirror
*shrug* I didn't give it +5, and I wouldn't have. But I would have given it +2 easy. Maybe +3 if it cought me in the right mood.
I hope not. My only GUI apps are OSS. I expect a patch, not a bitch. Especally on things where even non-coders can find the quoted string and change it.
Do? Or should? I think they should say "shutdown {computername}", or "shutdown {username}'s computer". I expect Microsoft just has them saying "shutdown computer".
Are you sure? What about Remote Desktop Connection? Can you be turning off/rebooting some else's computer?
I don't think they lose control just because they sell it to you. I think they lose control because they sell it to you without making you sign a terms of service contract. If you do sign such a contract I think you may be legally (or maybe just morally) bound to follow it.
I don't know about shrinkwrap agreements (where the terms are presented after you buy). In the past they were not legally binding, with the DCMA they might be. However I don't think they are morally binding. In fact I think they are morally repugnant.
And I definitly think that a shrinkwrap agreement on software you never ran (say, because the "forgot" to ship the Linux or FreeBSD version) isn't going to be legally binding at all!
O.B. trick. Don't like a shrinkwrap agreement? Change all the terms you disagree with, inital the changes, have it noterised, and send the terms back. This almost certonally won't get the terms change the way you want, but may legally void the shrinkwrap agrement and return you to the default rights of a UCC purchase (i.e. you probbably can't give copies to friends, but you can probbably reverse engenear, make backups, and all the other things people have been doing to software for 20 years, and machanical devices for centries).
Multiple OC-12s are rather expensave arn't they :-) Seriously that figure is 3 to 4 years old, and for a high bandwidth location with forced air cooling and 3 day battery backup. I don't know what current figues are, I would expect them to be much lower in many places.
My original point was merely that some places are really really expensave, and paying an extra $3500 to save 2U of rack space has a payback time mesured in months (two in this case).
You need a lot of bandwidth coming in to the facility, from multiple providers. Or at least on multiple fibre bundles. Better still if you are centrally located (to minimise lantency to both coasts). Lots of good places in the center of the USA to get land cheep, not so hard to find major switching centers (cheeper circuits), but kinda hard to find places close to multiple telcos.
The other route to being a high cost colo is to have lots of ISPs show up where you are (like MAE-east/MAE-west), but those are hard to start, you normally have to bring a bunch of bandwith to the party and peer cheep for a while...
And battery and/or gennerator back up is a must. Forced air cooling is very good. Humidity control is a must (normally the cooling does that for you, but just in case you want a alaska colo, you still need to control moisture even if cooling is taken care of by nature). Apartments tend not to provide much cooling, so ration out a big cunk of the bedroom for a chiller. They also don't normally have raised floor/drop ceiling, so chop off some more space to install them (or use all ladder racks and overhead cable runs, but you still need space for the cable runs).
You'll be better off renting wharehouse space... (esp if you can find it close to a CO).
Also rember Hurricane Eletric does way cheep rack space, and other places too...
Yep. It is cheper because doing it yourself is almost allways cheaper. Plus it is chepaer because it is three times the size. Sometimes that is a great tradeoff. Other times rack space is expensave (sometimes MORE then $1000/U/month). Frequently expensave enough to make a 1U vs. 3U diffrence worth many $1000s in a single year. Sometimes in just a month. Try pricing telco colos someday.
Of corse there are lots of places where rack space is dirt cheap, and using 3U vs. 1U to save a little money is a no-brainer.
The big problem is space tends to be an inelastic commodity. Run out of disk space? Buy more! Run out of space space? Oops, it takes months to set up a new lease and get it connected and... In fact just evecting people from their office/cubes so we can grow the machine room takes time...
So space starts out cheap, and gets incresingly costly as you consume it. Sometimes you have to go back and replace the 3U disk boxes with 1U disk boxes, and the 3U SPARCs with 1U SPARCs, and sell off the old ones (if they retain any value). At least the new stuff costs way less then it would have when you first did the build-out...
When did that start? Why? Did I miss a meeting?
Use snprintf(3), or the other "n" variants. Like snprintf(buf, sizeof(buf)-1, format, ...). The -1 isn't needed, but it makes me feel better (whoever implments the libc function might not have read the spec close enough).
Or in C++ use a stringstream which will expand itself as needed. Or on 4.4BSD systems C programmers could use funopen(3) to get the same sort of effect, but not so portably.
And there is allways the "make my buffer one page big followed by a red zone", but that isn't hugely portable, and opens a DoS attack.
To be honest I use snprintf (and vsnprintf), and havn;t needed to do the like in C++ yet. But all of that stuff should work. And remember to put some overflow tests in your test vectors.
You mean it isn't?
Ever see Dogma?
My FreeBSD box syncs with my Visor (Palm clone) over USB. I have a working USB Ethernet (intended for use on an i-opener, or similar unit which I havn't had the time to hack up). I expect my new digital camera will as well since I got sources to a Linux program that does it. I just have to find out the diffrences between raw USB access on Linux and FreeBSD.
According to the (BSD) usb mailing list I'm on many USB modems work. As do many USB mass storage devices. Not that USB1.1 is a fast access path to mass storage, still slower then 1987's async SCSI.
I don't know how closely sync'ed the BSD and Linux USB stacks are. I think the Linux one started with NetBSD's (a fair and good idea). FreeBSD's track's NetBSD's.
I'de say more but the dog wants attention...
Your right, at least several of the examples I gave were not working capital. I do beleve the rest of my post is correct. As is yours.
Does that include VC houses like Kliner-Perkens or are VC houses actually a rareity? (I've only been trhough one IPO, and I wasn't even remotely involved with raising VC money, other then making the componies products work while that was going on)
One of the big reasons a compony goes public is to get some working capital to make investments (factories, expensave talent, marketing blitz) that will make an even bigger profit. A profitable compony frequently has less reason to IPO then a not-so-profitable one. One of the other big reasons is because tipically Venture Capitalists get to make some of those calls in exchange for money when the compony really really really needs it (like to go from a literal garage start-up to a underfunded low-rent office park start-up; or to go from "I got an idea!" to "It works in the lab!").
As for investing in a compony that has made money, well that doesn't mean they will make more in the furure. More importantly, it doesn't mean it will make more money then people are currently betting on which is what stock value really is. You would rather invest in a compony that is about to make a lot more money then people think it is. Maybe one that is just shipping a product that really is as good and revolutionary as people say (it's a good excuse to buy products the day they hit the market -- to check for good stock investments...a lession I should have learned with the Zip drive, and applyed to the AMD K7!).
P.S. I am not a profesional. This advice may lose you money. I'm not responsable. The risk is yours. Blah blah blah.
Except when they:
As a director? I have no idea. As a starting programer it didn't seem hard to get a starting programer-job at MP Games (Microprose's now defunct coin-op sister componay) almost ten years ago.
Actually I can rather assure you that I do :-)
It was a lot of work. And there was a lot of collabration with the artists and other coders, so everyone keeping similar hours was pretty important, which makes for painful mornings after an all-nighter.
Then again working for a little start up (UUNET) was also very demanding. It didn't have the same pain-in-the-ass be-in-in-the-morning requirements, but there was a lot of hard work there. I'm not sure there are worthwhile jobs that arn't demanding.
Well nobody shows that stuff. It isn't photogenic. If you ask to see it when you go interview you'll get too (you should at least walk by it even if you don't ask -- be extreamly wary of off-site interviews! Rember you are interviewing them as much as they are interviewing you...at least if you are doing it right!).
And every place has it's own little steam-blow-offs. Sometimes nerf wars (scheduled or not -- my Wildfire is loaded and ready monkey boy, bring it on). Sometimes gaming. At game componies there is a lot of gaming (the caf had lots of board games in it, and a few CoinOp clasics, plus our own of corse), inclusing "field trips" (either to see your game in test, or just to playtest the others...pity they didn't let me file an expense form for that, but at least it was gaming on work time).
But other componies have fun things as well. Nerf wars are popular. Cube decorating, and reconfiguring isn't bad. Nor is "lets see if we can get on the roof". Some places will have really long lunches (when far enough from deadlines at least).
Few componies are utterly devoid of any sort of fun at all. (well, maybe there are, I've only worked at about five places plus some consulting, so and other then consulting all in the MD/VA area)
No, much free software is unusable for the same reason much comercial software is unusable. It just that most software is unusable. If you disagree go use a peice of comercial software at random. Don't forget all the in-house apps that need a ton of instruction to use, stupid instructions like "don't hit 'all packages', that just crashes, select them all by letter".
Lots of comercial crap is worthless. Lots of free crap is worthless. It is probbably easier to find the worthless free crap, but only because it is free, not because there is more of it.
Don't forget lots of free stuff is pretty damm nice. Apache. OpenBSD. GCC. The Smithsonian. Love. And some claim Linux even.
Eh? How is that relivent?
No, no, no, god no. Read last months' C++ users Journel, or Strousup (damm those forginers and their hard to spell names!).
As "NULL References" are not defined, making one can do anything. They can crash you right as they are created. They can crash you when you try to take their address.
What you show works on some implmentations, but is a programming error, every bit as much as using memory after you free it is (which normally works until you allocate more memory). But any garbage collected implmetation, or an implmentation that does anything odd with references (perhaps because the CPU is odd, or the ABI) can show the breakage, even if you try to check for the NULL reference, because the bug is not the use of the NULL reference, but the mere existance, the bug starts when you create the reference, and an implmentation can crash at any moment after (or as) you do it, attempts to find and not use the NULL reference are just masking the bug, and not allways masking it.
Expect a future compiler rev to break such code. Expect porting the code to make the bug come live.
In other words, you are living the life of a VAX programmer that assumes '0 == *NULL' works, and will forever work (and that 'strlen(NULL) == 0' works as well). Soon everyon will want your code ported to that hot new Sun3 platform, and you will have to find and fix all those damm bugs.
Eh? I wrote a whole (modestly) big Gtk-- program and never used strings other the the standard C++ string, and standard C char*'s. Never used a vector other then the STL's. Gtk-- might (or might not) have it's own string and vector classes, but I definitly didn't do anything to avoid their use, and yet I'm also not using them.
Ah, but the 'func(*(int*)NULL);' part is illegal in and of itself (in C and C++), or more accuratly undefined. Most implmentations of C++ references just lazy evaulate the core dump :-)
There really is no such thing as a NULL reference really means
I don't find those things to be a huge advantage for references. Nor do I find that they make my code cleaner looking (except as a return type so I can have lvalue functions). I find it a modest disadvantage that I don't have the "out of bound" signaling path of passing a NULL to mean something "specail" (default value, or skip that part, or something).
Left to my own devices I use pointers. Then again I kinda grew up on C (and APL, but we'll not mention that again). On the other hand I really really love the STL. One of the few saving graces of C++. And a huge one at that. Frequently even signifagantly faster then C.
Because if someone else has attacked the problem before you are rock stupid to not look closely at what they have done so you can see what they did right, or wrong. So what you know might be good to copy (if it fits into your fraework), and what to avoid at all costs.
Do you thnk AMD isn't one of the first buyers of Intel CPUs? That they don't cut them open with surgens saws as soon as possiable?
Do you think that Linux kernel hackers arn't looking at NetBSD's USB system?
Do you think McDonald's marketing teams don't eat at Burger King?
That GM doesn't buy Ford cars and take them apart?
I doubt any code was copied. Qt is a toolkit. Gtk-- wraps an existing toolkit. The slot/connection model in Gtk-- is all done within C++ while Qt uses a preprocessor, which makes Qt programs "not quite C++", which I think would sometimes be a pain (not an insurmountable obsticle, but still a pain). Gtk-- makes templates fit the task, which seems to work quite well. I'm not sure if Qt could have done the same with the state of C++ compilers when they started. If I were to do it now I would definitly do it the Gtk-- or Inti way. Being second sometimes has huge advantages. The rest of the slot/connection model is similar in design between Qt and Gtk--, but it is also similar to Smalltalk and other systems that have come before (and I would claim both did a good thing copying a previous succsuful solution).
If you attack a strawman at least attack the right one. I don't think Stallman has a lot to do with Gtk+, I think it was Havroc or one of the other GTK+ develeprs who asserts that GTK+ being in C is a major huge good thing.
Personally I could give a crap less what language it is written in if it (a) works, and (b) has good bindings to the language I want to use. Now I realise that things never actually allways work, so I will avoid using a tool if it is written in Intercal, but just because I wouldn't pick C to do an OO program in doesn't mean I'll avoid it if, if it has good C++ (or Java, or whatever) bindings.
More importantly I think the "offical" argument has allways been that writing in C makes it easyer to have Python/Perl/Java/Whatever bindings then if it were in C++. I'm not positave I agree. I also don't think I care a whole lot. If I write a GUI app I'll probbably write it in C++ (or Java, but if I do it in Java I'll use Swing anyway). But the offical argument has never been "C rulz, C++ blows goats"
Beats me how many have passed alpha. Bu I can say w3juke was really really far easyer to code up then any toolkit I had used in the past (note I havn't used Qt), including some non-Unix ones. But w3juke has not passed alpha. Not really because of the toolkit, but because of lack of documentation, and available time, and the abundent lazyness of the programmer.
And of corse now I want to come up with another smal task to go code up in Inti just so I know how it compairs...
I never owned a NeXT, but the University I went to bought into it big time, so I did spend a bit of time playing with them.
The software was revolutionary. In many ways we are still catching up. Definitly in having a user friendly Unix we are still catching up. As little as I liked Objective C's performance, it did make things easy. Nice, nice, nice software. Good choice of bundled apps for an academic market too.
The hardware was not stunning at all.
If you sat it next to it's Sun's boxes of the era it was dog slow (if you ran SunTools at least -- if you ran X the display on the NeXT xould catch up). Both had equivolent resultion. The NeXT has 4bit (8bit?) grey. The Sun had either 1-bit deep mono, or 8bit color (which could do 8bit grey) depending on which graphics option you got. The NeXT had a 68030 (68040? 68020? mmmm, maybe the 68030) at something like 20Mhz. Sun had recently come out with the SPARC 1+ the follow on to the first desktop SPARC, 25Mhz I think. But much much much faster. Doing a whole lot more per cycle then the Moto part. The SPARC didn't feel a little faster it felt a lot faster.
The few hardware features the NeXT had and nobody else did were not particurlay well recieved. The "floptical" was a bit fragile, and most people only had the one and no HD so it wasn't removable media, it was just a slow hard drive. The DSP never got used for anything intresting, the promised high speed modem was extreamly late, and I not sure it ever worked. About the only inovatave hardware feature I remember on the NeXT was the cubes looked way cooler then even the new SPARC pizza boxes.
But the software, oh man was that software ahead of it's time...
It was. Or at least the use of a normally reserved opcode to tell the morpher to switch instruction sets seemd, and the rest of the explination of the demo (like them saying there was a picoJava morpher) seemed more consistant with two morphers at once then running a x86 JVM.
There is (as far as I know) only the one production morpher. Which makes sense. Don't dilute your manpower until you have some mesure of success....
One of the demo's they had at unveling was a copy of DOOM, and x86 executable with part of the inner loop replaced by picoJava. It seemed to run that pretty quick, so it can switch instruction sets dyanmically. There is undoubtly some penality, but it apparently isn't huge.
There is no evidence that it can run Alpha code particurally fast. It has 48 (or 40?) registers internally. More then enough to do software registr renaming, and have operatinog regesters left when emulating a 4 register (10 if you count segment registes, the PC and flags) machine. The Alpha has 32 user visable registers (or is it 32int plus 32FP?). The 21264 has 80 renaming registers (invisable to the programmer) in two banks. The Crosue would have far less, so could be expected not to be able to hide as much latency.
Oh, and since the Crosue is designed to emulate the 32bit x86, it is not likely to have 64 bit registers, or a 64bit ALU. And it's FPU certonally is geared towards the x86 quirks (80 bits, almost IEEE, but not quite).
I wouldn't want to try to emulate the Alpha on a Crosue. It could probbably do a great 68000 (or the CPU32 in the palm, which is almost a 68000). Maybe even a good ARM, or V8 SPARC (these are pushing it though). Could definitly do a good JVM. Would do a great 6502, or Z80. :-)
But not a Alpha. Not a V9 SPARC. Not the 64bit MIPS. Not the IA64.
Um, their Cray division did alot of the work for the O2000! In fact at release the >64CPU configs were only avail from cray. Oh, and the frame to frame comm channel? It's named the "CrayLink".
Nice machines though, even if a bit long in the tooth (the O2000 is fourish years old, the O3000's should be announced anytime now, go look at comp.arch)
Sure, if you need very little communication between machines Beowulf is great, and the O2000's expensave comms (the xbow and craylink) are waisted. If you need a lot of comm, but not a lot of com a O2000 is great. If you need a lot-lot of comm maybe you are out of luck until the O3000, HP SuperDome, or IBM Power5 show up.
Quick MP break down:
A NORMA (or better) is great for raytracing, crypto cracking, and the like. A UMA is great for N-body simulation (with large N). I wouldn't want to track the flow of air molicules over a wing with a Beowulf, but I wouldn't want to pay for a ccNUMA if I was "just" running PR-Renderman.
Clearly I don't remember enough. But I was actually intending to talk about how the CPU's waste heat is in a very small area (even the "hot" desktop CPUs don't disapate as much heat as the lightbulbs in my room, but I don't need cooling fans for my bulbs because they have more surface area...). The drive generates it's heat over a large enough area that it may need no fan. And as you said the LCD (or really it's backlight) make a lot of heat, but not anywhere that the laptop has to worry about it.