Mine is clocked normally. As is its bus, and the
aging Matrox Millenium 1 I have in there. And yet
every now and then 4.0.1 will segfault on me. And it ain't random either: I looked at the core dumps with gdb, and it's always in the same place.
On the plus side, 4.0 pulled this routine on me on a daily basis (same stack trace), 4.0.1 does it every so many weeks.
Before anyone says I'm whining: I'm not. I'm only observing, and I submitted an "official" bug report. Back in August already. And about
May or June concerning 4.0.
Whether I'm doing a good gob is something my boss
has to judge. Wether I'm using e-mail to talk to
my bank manager (hey, it's just an example), is
somehing that the IT department might decide to
look into, whether my boss thinks it's a good idea or not. By the time I can reply with "My boss is extremely happy with my work performance and even stated that he'd like me to have a twin brother such that he could hire him as well.", the damage may already have been done.
Not that the above matters much in my particular case, but in principle it does.
Warning: I'm not speaking for my employer. On the other hand, I am speaking as someone who professionally has to support a piece of software that, at least on Linux, depends on having a g++ that, shall we say, behaves itself. Part of this product is a small runtime library, that the customer has to link against when compiling the C++ code that we
generated for him/her.
and so we have a gcc snapshot as the complier. Somehow this makes us evil, whatever.
I'm sorry, but I do think it is evil to use a gcc
snapshot. The various g++ compilers have a history of not being binary compatible. While there are technically good reasons for these changes, they still are a nightmare to support. If companies like RedHat start distributing snapshots of something as essential as gcc, then before we know it, every single release of every single distribution will need
its own set of libraries and bug workarounds. If we're not careful, we'll end up in Windows country, where each application ships with is
own incompatible version of some DLL that, when installed, breaks at least two other applications. This can't be right.
For the zealots: Open sourcing is not an answer
in our particular case. Don't ask me why, it just isn't. If you ask anyway, I will not reply, for it would be off topic.
For the RedHat people: Please note that I do not
complain about 7.0 as a whole. Since I have not used it yet, doing so would not be fair. My only
gripe is with the general idea of shipping gcc snapshots as teh default compiler for a major distribution.
Ballot stuffing on the web happens all the time and by all sorts of people with some agenda to defend. It's also quite easy to do, actually.
In this particular case, I have seen no evidence whatsoever that Micro$oft did it. There are NT fanatics outside Micro$oft, you know. There are also people that will do this just for fun, so as to get M$ to be bashed again on Slashdot.
Regarding being an expert, I have the quote from some party unknown hanging on the wall of my office:
Be warned that being an expert requires more than understanding how a system works or is supposed
to work. Real expertise is gained by investigating
why a system does not work.
Maybe with training you can put out ~100WPM while dictating, but then the question becomes whether the machine can understand them at that speed. And remember, it has to do that at least 95% or so of the time, otherwise you have to continuously go back to check what it's doing.
Also, suppose you spot a misspelled word (or worse, something like a person's name) that you want fixed. How long does it take to move the
cursor there by means of a keyboard and/or mouse and fix it? How long does it take for you to say (and the machine to understand and execute): "third paragraph, eleventh line, replace McIntire with McIntyre"? Hmmm, seems like you will even have to count and specify the individual characters to get it to do that right...
Out of the mere 24 hours that an earthly day has, about 8 are supposed to be spent in the office. That's one third. Considering that another 8 are spent in bed or therabout, one can consider that the average office worker spends half his concious working life (hmmm, does such a thing actually exist?:-) in his/her office. Add in the commuting bits etc, and it turns out such these people spend a lot more of his/her waking hours in the office than at home.
Thus, office space accomodation should be extremely important to any company that takes its workers seriously. Companies that do not treat their workers seriously, generally aren't worth working for. Definitely not in the high tech business, where people's brains are the company's real assets, and where working days of 8 hours or less actually are the exception.
Here's what I want from my employer in terms of
office space (I'm a software engineer, and spend more than the required amount of time in the office):
A say in who I share an office with. This can make a huge difference. The number of people should be kept low (2 or 3), such as to minimise disturbances. On the other hand, a project team of 4 should have the ability to share an office if they want to do so, even if 3 is the company rule. Having to exchange phone calls or e-mails, or even to walk over to another room to discuss some detailed coding issue, also are disturbances.
A room with a view. Or at least a outside window. I want something to look at while I'm thinking. Besides, it's better on one's eyes if one can shift focus every so often.
Plenty of space to stack papers, manuals, etc., both on my desk and in the form of filing cabinets. Note that this is not in conflict with a clean desk policy for those who want the latter.
The ability to have some privacy when using my computer. That is, the ability to not have people sneak up on me from behind while I'm doing so called "less productive" stuff such as reading stuff on the web or even posting on Slashdot.
A door that can be opened or closed as I or those who I share an office with see fit.
Decent lighting, over which I actually have control.
The total absence of any kind of furniture police that goes about angrily pointing out that it is not allowed to turn one's desk around, that it is illegal for a senior person to share an office with a non-senior one, and who knows what other nonsense.
Nope. When they say dual licensed, they mean dual licensed. The pessimistic bit is that even the dual bit may still fail to happen if too many contributors would object. OK, the latter is unlikely, but not impossible.
It's not a transition phase towards GPL only. Here's why: Netscape. If they put Mozilla under GPL only, Netscape has to be GPLed as well, which they cannot do and/or don't want to do.
If you just continue to inhale 100 percent of CO2, you die (we indeed can't take the oxygen out of CO2 and use that instead). But that's mostly due to lack of oxygen in pure molecular form, not due to the CO2 itself. It's similar to inhaling pure nitrogen: you'll die from that just as well, even though normal air contains 78 (IIRC) percent nitrogen that does you no harm whatsoever.
As others have mentioned, excess C02 will result in symptoms of suffocation (even when enouugh oxygen is around), which means that you will be warned that something's up.
CO2 isn't dangerous to humans. Well, at least not as such. Enormous quantities of it are continously released into the atmosphere by cars, electricity plants, volcanoes, your local heater (unless its electric), etc, And by... exhaling humans.
CO, on the other hand, is extremely dangerous: it's odorless, invisible and very toxic. Also initial symptoms of CO poisoning are such that victims tend to react in completely the wrong ways.
Forcing moderators to read at -1 is something I could live with. it's what I do anyway. But forcing them to read "newest first", or hiding other information from them is a good idea. Moderators also are readers and posters.
To me it menas that he didn't turn it on and considers this one of the things that can still be tuned. Look at the other tuning items on the list on slide 42, and you will see that it is so. IMHO, at least.
Go to http://www.deja.com/forms/nuke.shtml and search the entire database for the e-mail address you want to kill. Then go through the posts one by one (Yes, one by one. Sigh...), clicking on the "Nuke" button. Then wait for the confirmation mails to arrive, and forward each of them back. Note that you must still be able to read and send mail at/from the address in question.
Not really feasible if you have hundreds of articles, me thinks, but it worked for me. If I had hundreds of posts, I'd process a few of them manually to figure out all the details of how the urls work, and then write a script to send of the remaining requests.
Yes, there is a need for a good usenet archive searcher. I use(d) Deja quite often to search for stuff in the past, but it's going down the drain real fast. Yet, until I find a replacement I will still go there every first monday of the month (cron reminds me) to find out if someone somewhere posted something about a piece of software that I maintain. By the way, this also suggests a feature that I would like to have: the ability to completely automate this search.
But be sure to make your stuff better than Deja and to better respect people's rights and feelings. I.e. NO editing of the messages, please, and certainly not to insert ads. As a matter of principle, I'm busy nuking all my posts from Deja because of the ad issue. The home ones are gone already, the work ones will be as soon as I get back to the office.
Ah, but then consider this: someone takes gcc, improves its code generation for SPARC a lot and compiles it on a x86 machine. Then (s)he makes or takes an non-GPL-ed x86 emulator for SPARC, and sells the combination of both as a proprietary C compiler for SPARC machines. Assume that it is done is such a way that the end user cannot just extract the gcc binaries from the distribution and run them on a real x86.
By the reasoning you mention, this would not be a violation of the GPL, since the actualy gcc code is being executed by an intermediate program.
If the above really is not a violation, then let's go one step further: assume the product is also sold with an x86 emulator running on... x86.
Also consider the fact that when the prayer book was written, this probably was not the only remaining copy of the Archimedes text. It can very well have been one a several known to still exist. So, if they were short on raw materials for their important new task, why not reuse this one?
I agree in general, but would still like to know: how often does one need to swap a CPU in what are supposed to be high quality machines in the first place?
I've seen hundreds a UNIX workstations and (SMP) servers (from Apollo, Sony, Digital, Sun, and HP) pass through our offices over the last 12 years. Not one CPU has failed. I did see several memory modules fail, but that is easily explained: when we bought that batch of machines, we decided to save some money by going for third party modules instead of buying the real stuff from HP. Quite a few of these cheap beasts failed, none of the HP ones ever did. The stuff that fails most are the power supplies, the disks, etc.
Indeed. And for completeness: once you have made all the right descisions about the security process, you still need to get the implementation of the whole lot right (which in turn is a process issue, by the way). And with right I mean exactly right. It wouldn't be the first time that a security problem is accidently inserted during the implementation phase (e.g. use only half of the bits at some point, or use the wrong bits such that the result becomes easier to analyse, or...).
Unix was designed on the basis of not printing spurious chatty messages, so it's easy to tell when something really does require your attention.
Sounds like Windows to me. Windows does not tell me anything while booting, even if I instruct it to be verbose. Linux will chat away at top speed. In fact, that's what I like: I want to know what the box is doing and whether the CPU/BIOS/... suffers from known bugs.
Re:But is this really for the better?
on
Microsoft Loses
·
· Score: 1
Yes, some money will really be lost (i.e. will disappear). But not all of it. And before you know it some new vapour money will be created again elsewhere.
--
Re:But is this really for the better?
on
Microsoft Loses
·
· Score: 1
I can understand your point of view, but one shouldn't forget that there are lots of people who own Micro$oft stock indirectly through all sorts of investment and pension funds. And most of these have no idea what's supposed to be so bad about Micro$oft either.
Hell, even I (a computer professional and long time MircoSoft hater (and for all the right reasons at that:-)) will probably loose a few dollars these days because of such a fund over which I have no direct control (other than to get out, but that also costs). But franckly, I personally don't really care. I knew full well what I was doing when I got in, and have gained considerably more from it than I will loose now. Besides, the loss will be temporary anyway, and most of the money that is leaving Micro$oft will be moving elsewhere where I can again hope to profit from it.
Unfortunately I don't remember the source, but I once read that Bill has a rule that Micro$oft must always have enough "cash" readily available to survive a full year without income.
Slashdot Mirror
On the plus side, 4.0 pulled this routine on me on a daily basis (same stack trace), 4.0.1 does it every so many weeks.
Before anyone says I'm whining: I'm not. I'm only observing, and I submitted an "official" bug report. Back in August already. And about May or June concerning 4.0.
--
Whether I'm doing a good gob is something my boss has to judge. Wether I'm using e-mail to talk to my bank manager (hey, it's just an example), is somehing that the IT department might decide to look into, whether my boss thinks it's a good idea or not. By the time I can reply with "My boss is extremely happy with my work performance and even stated that he'd like me to have a twin brother such that he could hire him as well.", the damage may already have been done.
Not that the above matters much in my particular case, but in principle it does.
--
and so we have a gcc snapshot as the complier. Somehow this makes us evil, whatever.
I'm sorry, but I do think it is evil to use a gcc snapshot. The various g++ compilers have a history of not being binary compatible. While there are technically good reasons for these changes, they still are a nightmare to support. If companies like RedHat start distributing snapshots of something as essential as gcc, then before we know it, every single release of every single distribution will need its own set of libraries and bug workarounds. If we're not careful, we'll end up in Windows country, where each application ships with is own incompatible version of some DLL that, when installed, breaks at least two other applications. This can't be right.
For the zealots: Open sourcing is not an answer in our particular case. Don't ask me why, it just isn't. If you ask anyway, I will not reply, for it would be off topic.
For the RedHat people: Please note that I do not complain about 7.0 as a whole. Since I have not used it yet, doing so would not be fair. My only gripe is with the general idea of shipping gcc snapshots as teh default compiler for a major distribution.
--
In this particular case, I have seen no evidence whatsoever that Micro$oft did it. There are NT fanatics outside Micro$oft, you know. There are also people that will do this just for fun, so as to get M$ to be bashed again on Slashdot.
--
Be warned that being an expert requires more than understanding how a system works or is supposed to work. Real expertise is gained by investigating why a system does not work.
--
Also, suppose you spot a misspelled word (or worse, something like a person's name) that you want fixed. How long does it take to move the cursor there by means of a keyboard and/or mouse and fix it? How long does it take for you to say (and the machine to understand and execute): "third paragraph, eleventh line, replace McIntire with McIntyre"? Hmmm, seems like you will even have to count and specify the individual characters to get it to do that right...
--
Thus, office space accomodation should be extremely important to any company that takes its workers seriously. Companies that do not treat their workers seriously, generally aren't worth working for. Definitely not in the high tech business, where people's brains are the company's real assets, and where working days of 8 hours or less actually are the exception.
Here's what I want from my employer in terms of office space (I'm a software engineer, and spend more than the required amount of time in the office):
--
It's not a transition phase towards GPL only. Here's why: Netscape. If they put Mozilla under GPL only, Netscape has to be GPLed as well, which they cannot do and/or don't want to do.
--
As others have mentioned, excess C02 will result in symptoms of suffocation (even when enouugh oxygen is around), which means that you will be warned that something's up.
--
--
CO, on the other hand, is extremely dangerous: it's odorless, invisible and very toxic. Also initial symptoms of CO poisoning are such that victims tend to react in completely the wrong ways.
--
--
--
Not really feasible if you have hundreds of articles, me thinks, but it worked for me. If I had hundreds of posts, I'd process a few of them manually to figure out all the details of how the urls work, and then write a script to send of the remaining requests.
--
But be sure to make your stuff better than Deja and to better respect people's rights and feelings. I.e. NO editing of the messages, please, and certainly not to insert ads. As a matter of principle, I'm busy nuking all my posts from Deja because of the ad issue. The home ones are gone already, the work ones will be as soon as I get back to the office.
--
By the reasoning you mention, this would not be a violation of the GPL, since the actualy gcc code is being executed by an intermediate program.
If the above really is not a violation, then let's go one step further: assume the product is also sold with an x86 emulator running on... x86.
--
--
I've seen hundreds a UNIX workstations and (SMP) servers (from Apollo, Sony, Digital, Sun, and HP) pass through our offices over the last 12 years. Not one CPU has failed. I did see several memory modules fail, but that is easily explained: when we bought that batch of machines, we decided to save some money by going for third party modules instead of buying the real stuff from HP. Quite a few of these cheap beasts failed, none of the HP ones ever did. The stuff that fails most are the power supplies, the disks, etc.
--
--
--
Sounds like Windows to me. Windows does not tell me anything while booting, even if I instruct it to be verbose. Linux will chat away at top speed. In fact, that's what I like: I want to know what the box is doing and whether the CPU/BIOS/... suffers from known bugs.
--
--
--
Hell, even I (a computer professional and long time MircoSoft hater (and for all the right reasons at that :-)) will probably loose a few dollars these days because of such a fund over which I have no direct control (other than to get out, but that also costs). But franckly, I personally don't really care. I knew full well what I was doing when I got in, and have gained considerably more from it than I will loose now. Besides, the loss will be temporary anyway, and most of the money that is leaving Micro$oft will be moving elsewhere where I can again hope to profit from it.
--
--