TGVs run on standard tracks. They use specially constructed railway lines because running at those speeds requires the bends to be less sharp and all that, but the gauge is the same.
I would never give up total control over my home machine, if only because playing around with kernels is a hobby, but at work things are different.
So different in fact, that I've been one of the forces behind the introduction of X terminals over there 3 years ago and have never regretted it. Oh, and I'm not one of those idi.. ehrmm users who log into screens on a backend server (look at my signature for a related hint). I work in the EDA business, and develop ECAD software for a living. Other people at the place where I work actually design quite big chips using this software and working in the same computing environment. Apart from a few Micro$oft fans, they are all pretty happy with the general architecture of our infrastructure.
Besides using computers I also happen to know about UNIX system administration and all that lot, and so will claim that I know why I pushed for thin clients. I can only say to Sun that they should push the Sun Ray concept as much as possible, e.g. by opening up or licensing the comunication protocol (e.g. to HP, hint, hint;-). Moving the X server out of the X terminal solves the last remaining problem caused by distributed management. And it has other nice advantages as well. Especially the ability to move to another terminal, without logging out, and still having all of one's open windows around, together with all the associated state information, independent of how well the tools are written. This is very yummy stuff.
Also, while reading the responses so far, I get the impression that the importance of OS kernel stuff for CS students is being overrated here. OS courses are important to CS students, but there are many other aspects to CS that are just as important in real life. Just look at where CS graduates end up in real life: How many end up doing OS (or OS-level) development and how many end up in application development?
Its sad, but the CS department I'm refering used to be a UNIX house, while nowadays they seem to be mostly using M$ in their courses as well as for their internal workings.
The reasons are easy to see: M$ is what a considerable section of the industry asks for, and while teachings specific practical skills is not their main focus, why should they not have the students use the stuff that they will most likely have to use later on anyway? It used to be UNIX, but it now seems to be M$. Also, M$ actually is the cheaper platform if they want to expose their students to Object Modelling products such as Rational Rose et al. (I've seen the Rational Rose UNIX vs. NT pricing differences first hand.)
By the way, one of the better known Linux kernel guys works at said department, but that makes not much of a difference.
Whether the NSA/FBI/... are now giving in because they have found a new trick is "interesting speculation", but not all that relevant in practice. Even if it is true, nothing changes for users of "exportable" encryption software when dealing with the NSA: the NSA was able to break such crypto before, and it still is. What's more, nothing much should change for users who distrust the NSA and use stronger crypto either: based on their distrust and assumption that the NSA already had a secret trick up their sleeves, they will (should) already have opted for stuff that is a lot harder to break anyway.
What is relevant in practice, is that one can now export (i.e. use) stronger crypto that (for the time being) only the likes of NSA, but not Joe Random Cracker, would be able to break.
Notice that I'm not saying "so what". All I'm saying is that the important thing is that it's a move in the right direction, whatever reason made it possible.
Has anyone out there run in to any real problems while doing Y2K testing?
Yes. The HP-UX version of sccs failed on me on january 1, 1999 at 00:13 because for some insane reason it uses (used?) a date 1 year into the future when checking in a modification. The Solaris sccs was fine, though, and the HP-UX problem was fixed within hours by installing a patch that had been out for nearly half a year. But still...
PS: Moderate this as funny if you like, but it really happened as described. The good thing about this was that it made a few people listen to my requests not to wait till now to update things.
--
Moderate this up! It's good medicine.
on
Moderation Ideas
·
· Score: 1
Cfr subject. Sorry for not doing it myself. I would if I could.
Maybe I'm mistaken in my terminology (sorry, I'm a railway fan, but English is not my mother tongue), but as far as I know, Britain and France (as well as the rest of Europe, except for Spain and Russia) all use the same gauge for a large majority of their railway lines: 1.435 meter.
This is, however not to mean that they use the same rules when other measurements are concerned, such as how close to the tracks a signal is allowed to be etc. These things are indeed different in the UK.
HP-UX 11, being a 64-bit OS, breaks the 4 Gig limit, but IIRC the actual limit varies a bit between "patch levels". They're moving it upwards all the time, I seem to recall.
I don't agree with the bit about not "needing" the names.
Especially when I'm entering a discussion that has been going on for a while, I like to know how many active parties there are, and what other posts are by the person who wrote something that I found interesting (or revolting) in some way. This kind of information can save a lot of time (and in Europe: telephone charges).
Also, this effect actually is more important to me while being a moderator. Not so as to locate preferred favorites or victims, but so as to "analyse" the treads taht I read more efficiently. Being a moderator takes more time than just being a reader, so let's no waste time trying to read while one accidently also happens to be a moderator.
It feels a bit silly to reply to my own comment, but: in the time it took me to type it, the ortiginal one got moderated up from 2 to 5. I guess that shows that moderation can work.
Well, the post obviously is not being moderated through the floor. And if I would have points right now, I'd moderate it up from 2 to 3, even though I do not agree that the moderation idea is bad. Slashdot improved a lot when moderation was added, and has not lost that again since then.
I do, however, agree that the meta-moderation idea is a looser. It's overly complicated to use, and based on my experience addresses a problem that does not need (an attempt at) this kind of "drastic" action.
I also think/agree that this is a case of the system taking on a life of its own. Moderation problems? Simple: moderate the moderators. Meta-moderation problems? Also simple: moderate the meta moderators. Not. This is one big step on the way to something similar to a bureaucracy, where every little action needs to be motivated on a paper form (in triplicate) and approved by a comittee that is not aware of anything but that fact that one of those forms was filled out. And probably not even that.
Sorry to post a "me too", but in this case I really think I should, for this is a really bad idea that will not benefit Slashdot at all.
Let's not forget that there is life outside Slashdot (yes, even Slashdot readers have jobs, studies, families, and all that funcky stuff), and that we need people who have one to be an active part of the community in order to prevent a case of mental incest. Please do not expect them to spend lots of time on doing quality moderation that is not going to count. Not only because they deserve some respect for their efforts, but also because they won't do it.
Always showing the controls would quickly drive away the good, thoughtfull, moderators and deliver Slashdot into the hands of the "randomly firing loose canons". And sure, meta-moderation might help to address that, but somehow I'm getting a strong feeling that this whole meta-moderation idea is a case of a solution that's looking for a matching problem. Besides, who would still care to meta-moderate after a while?
By definition of `average' they almost certainly do. The number of them that sort of has nothing else to think about is far larger than the number of techies or others with alternative programs compiled and running on them.
But... who, cares about the average? Who even wants to be average?:-)
The funny part is that the definition of "being normal" is "to fit the norm" and that the norm is defined as "that what people in general do, like, accept,....". Hence, being normal translates into being average. Fortunately I have a record of claiming to be glad to be abnormal.:-)
This would find too many false links. Here's one reason why: often when I edit and view an existing page, I edit a temporary copy instead and replace the real page only when I'm satisfied with the changes. Clearly, you don't want the temporary version to be indexed...
I'm having a real hard time imagining why anyone would want to run a web browser any wider than 800 pixels anyway
My default size is 800x880, but occasionally I'll make a window a bit wider than that.
But, speaking as a long-time programmer and a big fan of genericity: it is not the issue how big people's windows are (on average) and/or what one can or cannot imagine that others will want to do (either now or 5 or so years from now). The issue is that when it takes no extra effort at all to get the tables right, there is no reason to get them wrong. Hell, even if it did require some effort, it should still be done right.
Computers are meant to automate things for us, so why, oh why, do so many of their users, and even so many programers, insist on doing things manually and getting them wrong in the process? Let that stupid piece of silicon take care of things like the layout of WWW pages automatically. That's what it's "good" at, or that's at least what others already have t old it how to do.
Anyway, if I remember correctly, some Micro$erf (possibly even the Big Micro$hief himself) once admitted that the biggest mi$take they made when concocting DOS, was to use \ instead of the then (already;-) "widely" accepted/. Hmmm, while I hate the \ with a passion, I very much doubt the accuracy of the assesment, but that's another issue.
I don't remember where I read that, though, only that it was many many moons ago.
Not really. They got slashdotted simply because they made the headlines on slashdot and every other slashdotter and his/her dog headed over there just to have a look, whether actually interested or not. Multiple times actually, since they weren't getting through.
The things that Borland want to know, are:
what is it that they should build;
for whom;
can they make money out of it?
Those are very valid questions that go beyond the mere fact that lots of people visit the poll. There already is something like gcc, after all, which is free (like in speech and beer) and definitely good enough (maybe not to some people, but to most).
If NT can show Linux to suck 3 times in a 2 months period,...
It can do this 100 times in 2 months, because even with Linux's openness only so much can get done 2 months. So basically each of those 3 times are one and the same. In other words, your reasoning flawed and the number of times is irrelevant.
But the raw data are not irrelevant. They show Linux catching up fast. Rematch in 6 months?
--
Re:Warning! Rocky shores ahead!
on
Linux 2.3.0
·
· Score: 4
Indeed. I know I don't stand much of a chance with this (and will probably be accused of promoting sensorship by some), but I would plead for not announcing kernels in the development series here on/. Those who know what these kernels are for, will know how to be informed about them via other means.
A year ago, I still though different, but let's face the facts: the more Linux reaches the masses, the more this will become an issue. The 2.1.44 kernel stands out as a "shining" example of what can happen if things go real bad, but there have been others that were not quite as bad, but still unfinished and/or shaky enough to undermine Linux' reputation if too many newbies would have gotten hold of them.
sometime in the near future there will be a gcc 3.0
What definition of `near future' is being used here? A few weeks? Something like a month or 2? More than 2 months still? Can anybody tell already?
I have glibc 2.1 and egcs 1.1.2 waiting to be installed over here, but need to find the time to do it. If the `near future' is sufficiently near, I might want to wait for it to come.
Cfr. subject. After all, what does it say? It simply says to M$: if you do it, do it properly, and even then don't expect too much from it. Big news. I for one, while welcoming Netscape's move, back in '98, have said from the start that simply "going open" is no solution. For instance: it's supposed to attract developers, but hey, if everybody does it, then what? How many good developers that have both time and motivation to contribute are there? And then there is the problem of having a clean design that supports the approach well enough. M$, know this just as well, I think.
In any case, it's up to M$ to figure out what to do. If they mess up, too bad for them. If they don't, well, read the response: "we will welcome it", which obviously is the only possible polite answer the authors could provide to that scenario, so it could just as well have been left unsaid. Which also holds for the other part of the response. Especially since it can be read to imply that the design of M$ is a mess. That may very well be so, but there is no way to convince any pro-M$ (or even neutral) reader of this without proof. So the response ends up looking like a wild attack on M$ that is being made for no other reason than that M$ might be looking into knocking down one the people's favourite anti-M$ arguments: that it has to be bad because it is not open. This is of no use.
So why was this response needed? Or at least, why did some people think so? Think about it! I've given my answer already somewhere in the previous paragraph.
Note that this is not an anti-ESR or anti-whoever posting. In fact, I tend to be somewhat pro-ESR. In any case, it is definitely not intended to be a pro-M$ posting.
We've been using HP stuff for about 5 years now, and this is the 4th time since then that they changed their debugger strategy: softdebug version 1, softdebug version 2, good old dde brought in from Apollo and sort of ruined by HP, wdb,...
Taking into account that this wdb change is indeed at least a year old by now, I wonder whether they have already decided about their next strategic debugger...
Slashdot Mirror
--
I would never give up total control over my home machine, if only because playing around with kernels is a hobby, but at work things are different.
So different in fact, that I've been one of the forces behind the introduction of X terminals over there 3 years ago and have never regretted it. Oh, and I'm not one of those idi.. ehrmm users who log into screens on a backend server (look at my signature for a related hint). I work in the EDA business, and develop ECAD software for a living. Other people at the place where I work actually design quite big chips using this software and working in the same computing environment. Apart from a few Micro$oft fans, they are all pretty happy with the general architecture of our infrastructure.
Besides using computers I also happen to know about UNIX system administration and all that lot, and so will claim that I know why I pushed for thin clients. I can only say to Sun that they should push the Sun Ray concept as much as possible, e.g. by opening up or licensing the comunication protocol (e.g. to HP, hint, hint ;-). Moving the X server out of the X terminal solves the last remaining problem caused by distributed management. And it has other nice advantages as well. Especially the ability to move to another terminal, without logging out, and still having all of one's open windows around, together with all the associated state information, independent of how well the tools are written. This is very yummy stuff.
--
Also, while reading the responses so far, I get the impression that the importance of OS kernel stuff for CS students is being overrated here. OS courses are important to CS students, but there are many other aspects to CS that are just as important in real life. Just look at where CS graduates end up in real life: How many end up doing OS (or OS-level) development and how many end up in application development?
Its sad, but the CS department I'm refering used to be a UNIX house, while nowadays they seem to be mostly using M$ in their courses as well as for their internal workings.
The reasons are easy to see: M$ is what a considerable section of the industry asks for, and while teachings specific practical skills is not their main focus, why should they not have the students use the stuff that they will most likely have to use later on anyway? It used to be UNIX, but it now seems to be M$. Also, M$ actually is the cheaper platform if they want to expose their students to Object Modelling products such as Rational Rose et al. (I've seen the Rational Rose UNIX vs. NT pricing differences first hand.)
By the way, one of the better known Linux kernel guys works at said department, but that makes not much of a difference.
--
Whether the NSA/FBI/... are now giving in because they have found a new trick is "interesting speculation", but not all that relevant in practice. Even if it is true, nothing changes for users of "exportable" encryption software when dealing with the NSA: the NSA was able to break such crypto before, and it still is. What's more, nothing much should change for users who distrust the NSA and use stronger crypto either: based on their distrust and assumption that the NSA already had a secret trick up their sleeves, they will (should) already have opted for stuff that is a lot harder to break anyway.
What is relevant in practice, is that one can now export (i.e. use) stronger crypto that (for the time being) only the likes of NSA, but not Joe Random Cracker, would be able to break.
Notice that I'm not saying "so what". All I'm saying is that the important thing is that it's a move in the right direction, whatever reason made it possible.
--
Yes. The HP-UX version of sccs failed on me on january 1, 1999 at 00:13 because for some insane reason it uses (used?) a date 1 year into the future when checking in a modification. The Solaris sccs was fine, though, and the HP-UX problem was fixed within hours by installing a patch that had been out for nearly half a year. But still...
PS: Moderate this as funny if you like, but it really happened as described. The good thing about this was that it made a few people listen to my requests not to wait till now to update things.
--
--
This is, however not to mean that they use the same rules when other measurements are concerned, such as how close to the tracks a signal is allowed to be etc. These things are indeed different in the UK.
--
HP-UX 11, being a 64-bit OS, breaks the 4 Gig limit, but IIRC the actual limit varies a bit between "patch levels". They're moving it upwards all the time, I seem to recall.
--
Especially when I'm entering a discussion that has been going on for a while, I like to know how many active parties there are, and what other posts are by the person who wrote something that I found interesting (or revolting) in some way. This kind of information can save a lot of time (and in Europe: telephone charges).
Also, this effect actually is more important to me while being a moderator. Not so as to locate preferred favorites or victims, but so as to "analyse" the treads taht I read more efficiently. Being a moderator takes more time than just being a reader, so let's no waste time trying to read while one accidently also happens to be a moderator.
--
--
I do, however, agree that the meta-moderation idea is a looser. It's overly complicated to use, and based on my experience addresses a problem that does not need (an attempt at) this kind of "drastic" action.
I also think/agree that this is a case of the system taking on a life of its own. Moderation problems? Simple: moderate the moderators. Meta-moderation problems? Also simple: moderate the meta moderators. Not. This is one big step on the way to something similar to a bureaucracy, where every little action needs to be motivated on a paper form (in triplicate) and approved by a comittee that is not aware of anything but that fact that one of those forms was filled out. And probably not even that.
--
I vote for the "reaching enlightenment requires leading balanced life" idea.
--
Let's not forget that there is life outside Slashdot (yes, even Slashdot readers have jobs, studies, families, and all that funcky stuff), and that we need people who have one to be an active part of the community in order to prevent a case of mental incest. Please do not expect them to spend lots of time on doing quality moderation that is not going to count. Not only because they deserve some respect for their efforts, but also because they won't do it.
Always showing the controls would quickly drive away the good, thoughtfull, moderators and deliver Slashdot into the hands of the "randomly firing loose canons". And sure, meta-moderation might help to address that, but somehow I'm getting a strong feeling that this whole meta-moderation idea is a case of a solution that's looking for a matching problem. Besides, who would still care to meta-moderate after a while?
--
--
By definition of `average' they almost certainly do. The number of them that sort of has nothing else to think about is far larger than the number of techies or others with alternative programs compiled and running on them.
But... who, cares about the average? Who even wants to be average? :-)
The funny part is that the definition of "being normal" is "to fit the norm" and that the norm is defined as "that what people in general do, like, accept, ....". Hence, being normal translates into being average. Fortunately I have a record of claiming to be glad to be abnormal. :-)
--
And then there are all the privacy concerns...
--
There you have it: hardwiring 600 pixels is not the right thing to do!
PS: I'm a big fan of TeX. Guess why...
--
My default size is 800x880, but occasionally I'll make a window a bit wider than that.
But, speaking as a long-time programmer and a big fan of genericity: it is not the issue how big people's windows are (on average) and/or what one can or cannot imagine that others will want to do (either now or 5 or so years from now). The issue is that when it takes no extra effort at all to get the tables right, there is no reason to get them wrong. Hell, even if it did require some effort, it should still be done right.
Computers are meant to automate things for us, so why, oh why, do so many of their users, and even so many programers, insist on doing things manually and getting them wrong in the process? Let that stupid piece of silicon take care of things like the layout of WWW pages automatically. That's what it's "good" at, or that's at least what others already have t old it how to do.
--
Anyway, if I remember correctly, some Micro$erf (possibly even the Big Micro$hief himself) once admitted that the biggest mi$take they made when concocting DOS, was to use \ instead of the then (already ;-) "widely" accepted /. Hmmm, while I hate the \ with a passion, I very much doubt the accuracy of the assesment, but that's another issue.
I don't remember where I read that, though, only that it was many many moons ago.
--
The things that Borland want to know, are:
Those are very valid questions that go beyond the mere fact that lots of people visit the poll. There already is something like gcc, after all, which is free (like in speech and beer) and definitely good enough (maybe not to some people, but to most).
--
It can do this 100 times in 2 months, because even with Linux's openness only so much can get done 2 months. So basically each of those 3 times are one and the same. In other words, your reasoning flawed and the number of times is irrelevant.
But the raw data are not irrelevant. They show Linux catching up fast. Rematch in 6 months?
--
A year ago, I still though different, but let's face the facts: the more Linux reaches the masses, the more this will become an issue. The 2.1.44 kernel stands out as a "shining" example of what can happen if things go real bad, but there have been others that were not quite as bad, but still unfinished and/or shaky enough to undermine Linux' reputation if too many newbies would have gotten hold of them.
--
What definition of `near future' is being used here? A few weeks? Something like a month or 2? More than 2 months still? Can anybody tell already?
I have glibc 2.1 and egcs 1.1.2 waiting to be installed over here, but need to find the time to do it. If the `near future' is sufficiently near, I might want to wait for it to come.
--
In any case, it's up to M$ to figure out what to do. If they mess up, too bad for them. If they don't, well, read the response: "we will welcome it", which obviously is the only possible polite answer the authors could provide to that scenario, so it could just as well have been left unsaid. Which also holds for the other part of the response. Especially since it can be read to imply that the design of M$ is a mess. That may very well be so, but there is no way to convince any pro-M$ (or even neutral) reader of this without proof. So the response ends up looking like a wild attack on M$ that is being made for no other reason than that M$ might be looking into knocking down one the people's favourite anti-M$ arguments: that it has to be bad because it is not open. This is of no use.
So why was this response needed? Or at least, why did some people think so? Think about it! I've given my answer already somewhere in the previous paragraph.
Note that this is not an anti-ESR or anti-whoever posting. In fact, I tend to be somewhat pro-ESR. In any case, it is definitely not intended to be a pro-M$ posting.
--
Taking into account that this wdb change is indeed at least a year old by now, I wonder whether they have already decided about their next strategic debugger...
--