I like text files too [ grep:-) ] - and for bookmarks, maybe that's sane.
But bear in mind you can bookmark a group of tabs, and it no longer looks quite so simple. There's quite a lot of metadata: the URL, the user's name for it, the sort order, last accessed date etc.
The place this *really* matters though is the mozilla-mail address book. Have you ever tried to fix that? change one character, and the whole thing won't even parse.
Anyway, sqlite is easily used by other applications - and that's what really matters. OK, I can't grep it anymore, but on the otherhand, I don't need to learn the file format.
I am delighted to see this. Some of the mozilla stuff still uses Mork, which is truly and utterly horrid.
I recommend reading this delightful code by Jamie Zawinski, which has a brilliant rant about it:
Had you considered PHP? It has the huge advantage of simplicity (you can get started very very fast); it's also very powerful and widely useful (many many functions are available). In many ways, it is similar to C or Perl, so it's easy to migrate. The online documentation is brilliant. Also, you can use it as a scripting language (php-cli), for a simple GUI (php-gtk), or even for network-socket programming.
1)Set up the old machines as thin clients on a network. (eg LTSP). That means you can still use them; you'll have good performance, and the upgrade will be much easier. Then buy one high-end ($2k) server.
2)Keep a windows machine or 2. Run the windows software on it - and connect with VNC. Costs nothing; 100% compatible. [Wine is quite good these days, but if you already have the MS licenses...]
3)Don't buy any new windows apps - you'll only feed the addiction!
4)Rip the fonts out of your existing installs. The MS corefonts are still the best for legibility.
See: http://pcb.sourceforge.net/ I've used it for some fairly complex multi-layer boards. I prefer to do the schematic on paper and manually route it, but autorouting is supported. One warning: some of the symbols are wrong. In particular, the (default) holes in the pads are too small for most resistors.
Fair point. But I've actually seen what Ross Anderson does - even talking at the No2ID conference. His work is extremely good, and I've read quite a few papers. Also, your point fails, becuase if he were really in "counter-intelligence", he wouldn't make such a statement in public!
That article surprised me rather - because I know Ross Anderson to be one of the Good Guys. He is opposed to DRM, Trusted Computing, (see here) and ID cards.
Furthermore, even if he has had a change of heart, he's far too smart to advocate a backdoor into encryption.
Under Mandriva 2006, you get dmix by default. Set all outputs to alsa. Then use aoss (eg "aoss play file.mp3") around any older apps. Lastly, disable Arts totally, and set the KDE system notifications to "use external player". I play the file with sox | aplay.
The problem is that the UK card is intended (by stealth) to become a card without which you cannot live (it will be checked whatever service you wish to access), and which will contain an enormous amount of data - both on the card and the database. Furthermore, it will increase the risk of terrorism, make database abuse much easier, and may render civil disobedience (which is vital to our democracy) impossible for most people.
Of course, the BBC is (largely) immune to this. It is generally very good at resisting political pressure, and the funding body in this case mainly acts to keep the quality *up*. The BBC is usually pretty unbiased (which, of course, means that both sides accuse it of being biased against them).
> It may be very good for you, but to tell me that I need and want it, and therefore have to pay for it...
Indeed. That's why *I* support it. In other words, it gets my vote. However, even people who don't watch the BBC TV do get a benefit: the BBC keeps some pressure on the quality of the competition, and Radio 4 helps to provide an informed electorate, thereby contributing to our democracy.
For example, had the goodfortune to attend a private school. Therefore, I have never used the state education system - should I withdraw my support?
I agree that the soundcards ought to be numbered/dev/dsp0 etc. It wouldn't be that hard to do so if needed, since udev can allow you to make whichever entries you like!/dev/dsp is by convention - and anyway, alsa has a new design altogether. I think that the special devices (null,zero,random,stdin,stdout...) can remain unique. But I disagree about non-lettered HDDs. hda1 is much clearer to me than hd00. Besides which, things like hd110 would be ambiguous: is it hdb11 or hdk1 ?
I see what you mean. But still, no one is expected to believe in Pure maths except because it is beautiful and useful.
Religions requires you to believe them because they are each "The Absolute Truth". [They are also sufficiently inconsistent that this is not possible!]
Lastly, would any religious person be willing to consider God as an "axiom"? Sure..."assume God, and the rest follows..." - but that's not a very helpful theological position!
I disagree - your are talking about "trust". Sure, you can't personally verify everything - it would be impractical to do so. But the fundamental science has been verified by others, and importantly, you could verify it yourself if you needed to.
Religious faith is different in that the claims which are made are extraordinary (thereby needing strong justification), and yet nobody has ever verified them! Furthermore, you can't test them yourself.
For example, I've never done the experiment of putting lithium into water. But I trust the science teacher who told me about it because the claim is reasonable, and because, if I wanted to, I could test it for myself.
Lastly, not all physics is proven beyond reasonable doubt. But it never claims otherwise! Physics doesn't say "Accept this on faith"; it says "Here is the best theory we have - and so far, it fits all the experimental data perfectly. This explanation may not be the absolute truth, but it is pretty close - and we're still trying to get closer".
I see the subtlety of this argument - but I think it's a little spurious. Pure maths *is* testable within its axioms (Russell paradox notwithstanding), and it's also pretty well established via applied maths.
Not quite as illogical as it looks. The HDD partitions are numbered from 1 for consistency with the partition numbers.
The use of letters (hda,hdb...) isn't necessarily logical, but it does, I think, make it easier to make sense of. Eg "hda,hda1,hda2" are clearly the same device. Whereas "hd0,hd00,hd01" as you seem to propose would, I think, be harder to follow.
As for things like dsp and audio, these do essentially begin from 0, but the 0 is often omitted. That's probably due to the original assumption that no sane person would ever need more than one soundcard. Of course, nowadays, multiple soundcards are a reasonable proposition - but for illustration, consider having/dev/null0 and/dev/null1... !
It seems to me that Religious people need not disagree with science (since they are by virtue of faith, capable of the "doublethink" needed to maintain two incompatible world views simultaneously).
However, scientists must, if they are competent and sincere in their convictions, be opposed to any form of Faith. Faith cannot be experimentally tested/falsified, and therefore has no place in rational thought.
The only exception to this would be the "Gnostic" religions (often thought of as "Eastern" or "Mystical", although the original Christians eg St Paul were Gnostics too). These do not literally believe in a Personal Saviour God, but in that, by finding Enlightenment, one actually *becomes* a god. [Please excuse the simplification of Gnosticism here!]
I do in principle support the TV license, because the BBC (especially the Radio - World Service + Radio 3,4) is extremely good. However, I think that the TV license is a bad way to do it, and it ought to be included in general taxation. Reasons:
1)The license is there as a "tax of choice". So, if you don't have a TV, then you don't pay (not even if you do listen to the radio). This made sense in 1960 - but not so much now, when virtually everyone has a television.
2)The license collection is extremely inefficient. It involves hassle for the licensor, a draconian TV licensing authority (who make an enormous nuisance of themselves if you don't actually own a TV), and you cannot legally purchase any TV-capable equipment without giving a name and address to the retailer. [Yes, this is outrageous.] Enforcement and collection must cost a significant proportion of the total fee!
3)With the exception of pensioners, the TV license is the same for everyone. Yet, some can afford to pay more than others.
4)On principle: As a citizen, I have a natural right to my share of the RF spectrum - and to operate a Radio receiver!
However, the idea of a centrally funded broadcaster is a good one: it means that the quality of output need not go into freefall in the pursuit of ratings.
Why didn't you use Knoppix or something to get the music off it? Apple enforce a policy that you can only sync an iPod with one computer, but GTKPod doesn't. Or even use iPod Linux.
P.S. The iPod Linux GUI for the nano beats the native one by some way:-)
What I still wonder is why Ffox has got so much Gnome in it. Nothing against GTK (mostly), but a few things really really bug me:
1)Some of the options are controlled by prefs within the gnome-control center. There is no way to set/override them from within Firefox itself, nor is there even a hint as to where to find the control. By default, Firefox sends mailto:s to Evolution, not Thunderbird!
2)Why, oh why did they abandon the rather good native file-widget in favour of the horrendous abomination that is the new GTK2 filepicker? This is very limited, and worse, it takes 30 seconds to open, during which time, the application stalls!
Indeed. I cut a slightly long story short there - the machine in question was an expensive desktop (high-end parts, reputable manufacturers...) - but it got pressed into emergency service as a server...
> I think your eyes might have put the word "don't" into my post. I was joking that an informed post that actually helps > answer the question is quite rare on an 'ask slashdot'. It's usually a knee-jerk recommendation for $TECH_DU_JOUR;)
No, I'm not lost, and I think I do know what I'm talking about. What I'm basically trying to say is this:
Q: How do I get from A to B ? A: You don't want to start from here.
In other words, the original poster is asking about a RAID solution for a particular problem. However, I don't think that RAID is, in fact, the best way to go.
It costs roughly the same to have either 1 RAID box, or 2 separate servers without RAID. The former does give better performance, but this will probably be dominated by the network performance anyway: a single SATA drive can provide a bitrate faster than 100Mbit ethernet can transmit. The latter provides several extra advantages. In essence, it is "A Redundant Array of Inexpensive Servers".
> 2) Don't use a $15 PSU, you pay for what you get.
Correct - and indeed, I do buy expensive PSUs (and have a pile of never-used ones that came free with the case). However, even $70 PSUs can fail. RAID is banking on the odds of not loosing >1 HDDs simultaneously, so you are already looking at rare events.
> 3) If you live in a lightning prone area you should invest in some good protection (line conditioner for example).
Yes - but everywhere has *some* lightening, and you might just be unlucky.
> 5) What do you think the R is RAID is for? Redundancy. You'd need some horribly catastrophic > odds to lose all the redundant hard drives before you replace any.
Of course. But it might be cheaper (and more redundant) to have 2 separate machines, each with 1 HDD in it. My view is that this gives you lower overall risk. Sure, if you have a failure, it will take a bit more time to recover, but not much. All RAID protects you from is age-related disk failure.
> 6) Your data is still there.
No, it might not be. A recent problem I had on an (expensive) server was that a stick of RAM developed a fault, causing the OS to trash the filesystem. I was lucky: ReiserFS can re-build the journal from the files themselves. But if you use RAID, then this sort of crash will give you 6 identical copies of destroyed data!
Lastly, I would recommend nightly rotating snapshots to protect against user-error. That link I gave has an *extremely* clever solution.
Slashdot Mirror
I like text files too [ grep :-) ] - and for bookmarks, maybe that's sane.
But bear in mind you can bookmark a group of tabs, and it no longer looks quite so simple. There's quite a lot of metadata: the URL, the user's name for it, the sort order, last accessed date etc.
The place this *really* matters though is the mozilla-mail address book. Have you ever tried to fix that? change one character, and the whole thing won't even parse.
Anyway, sqlite is easily used by other applications - and that's what really matters. OK, I can't grep it anymore, but on the otherhand, I don't need to learn the file format.
I am delighted to see this. Some of the mozilla stuff still uses Mork, which is truly and utterly horrid. I recommend reading this delightful code by Jamie Zawinski, which has a brilliant rant about it:
The Rio Karma is an excellent machine - if you can still get them. Also, the newer ipods + ipodlinux aren't so bad.
Had you considered PHP? It has the huge advantage of simplicity (you can get started very very fast); it's also very powerful and widely useful (many many functions are available). In many ways, it is similar to C or Perl, so it's easy to migrate. The online documentation is brilliant. Also, you can use it as a scripting language (php-cli), for a simple GUI (php-gtk), or even for network-socket programming.
0 0.htm
P.S. Visual Basic is, and always was, utterly horrid.
http://www.ddj.com/documents/s=1503/ddj0001vs/jan
1)Set up the old machines as thin clients on a network. (eg LTSP). That means you can still use them; you'll have good performance, and the upgrade will be much easier. Then buy one high-end ($2k) server.
2)Keep a windows machine or 2. Run the windows software on it - and connect with VNC. Costs nothing; 100% compatible.
[Wine is quite good these days, but if you already have the MS licenses...]
3)Don't buy any new windows apps - you'll only feed the addiction!
4)Rip the fonts out of your existing installs. The MS corefonts are still the best for legibility.
See: http://pcb.sourceforge.net/
I've used it for some fairly complex multi-layer boards. I prefer to do the schematic on paper and manually route it, but autorouting is supported. One warning: some of the symbols are wrong. In particular, the (default) holes in the pads are too small for most resistors.
Fair point. But I've actually seen what Ross Anderson does - even talking at the No2ID conference. His work is extremely good, and I've read quite a few papers. Also, your point fails, becuase if he were really in "counter-intelligence", he wouldn't make such a statement in public!
That article surprised me rather - because I know Ross Anderson to be one of the Good Guys. He is opposed to DRM, Trusted Computing, (see here) and ID cards. Furthermore, even if he has had a change of heart, he's far too smart to advocate a backdoor into encryption.
Under Mandriva 2006, you get dmix by default. Set all outputs to alsa. Then use aoss (eg "aoss play file.mp3") around any older apps. Lastly, disable Arts totally, and set the KDE system notifications to "use external player". I play the file with sox | aplay.
The problem is that the UK card is intended (by stealth) to become a card without which you cannot live (it will be checked whatever service you wish to access), and which will contain an enormous amount of data - both on the card and the database. Furthermore, it will increase the risk of terrorism, make database abuse much easier, and may render civil disobedience (which is vital to our democracy) impossible for most people.
Of course, the BBC is (largely) immune to this. It is generally very good at resisting political pressure, and the funding body in this case mainly acts to keep the quality *up*. The BBC is usually pretty unbiased (which, of course, means that both sides accuse it of being biased against them).
> It may be very good for you, but to tell me that I need and want it, and therefore have to pay for it...
Indeed. That's why *I* support it. In other words, it gets my vote. However, even people who don't watch the BBC TV do get a benefit: the BBC keeps some pressure on the quality of the competition, and Radio 4 helps to provide an informed electorate, thereby contributing to our democracy.
For example, had the goodfortune to attend a private school. Therefore, I have never used the state education system - should I withdraw my support?
I agree that the soundcards ought to be numbered /dev/dsp0 etc. It wouldn't be that hard to do so if needed, since udev can allow you to make whichever entries you like! /dev/dsp is by convention - and anyway, alsa has a new design altogether. I think that the special devices (null,zero,random,stdin,stdout...) can remain unique. But I disagree about non-lettered HDDs.
hda1 is much clearer to me than hd00. Besides which, things like hd110 would be ambiguous: is it hdb11 or hdk1 ?
I see what you mean. But still, no one is expected to believe in Pure maths except because it is beautiful and useful.
Religions requires you to believe them because they are each "The Absolute Truth". [They are also sufficiently inconsistent that this is not possible!]
Lastly, would any religious person be willing to consider God as an "axiom"? Sure..."assume God, and the rest follows..." - but that's not a very helpful theological position!
I disagree - your are talking about "trust". Sure, you can't personally verify everything - it would be impractical to do so.
But the fundamental science has been verified by others, and importantly, you could verify it yourself if you needed to.
Religious faith is different in that the claims which are made are extraordinary (thereby needing strong justification), and yet nobody has ever verified them! Furthermore, you can't test them yourself.
For example, I've never done the experiment of putting lithium into water. But I trust the science teacher who told me about it because the claim is reasonable, and because, if I wanted to, I could test it for myself.
Lastly, not all physics is proven beyond reasonable doubt. But it never claims otherwise! Physics doesn't say "Accept this on faith"; it says "Here is the best theory we have - and so far, it fits all the experimental data perfectly. This explanation may not be the absolute truth, but it is pretty close - and we're still trying to get closer".
I see the subtlety of this argument - but I think it's a little spurious. Pure maths *is* testable within its axioms (Russell paradox notwithstanding), and it's also pretty well established via applied maths.
Not quite as illogical as it looks. The HDD partitions are numbered from 1 for consistency with the partition numbers.
/dev/null0 and /dev/null1 ... !
The use of letters (hda,hdb...) isn't necessarily logical, but it does, I think, make it easier to make sense of. Eg "hda,hda1,hda2" are clearly the same device. Whereas "hd0,hd00,hd01" as you seem to propose would, I think, be harder to follow.
As for things like dsp and audio, these do essentially begin from 0, but the 0 is often omitted. That's probably due to the original assumption that no sane person would ever need more than one soundcard. Of course, nowadays, multiple soundcards are a reasonable proposition - but for illustration, consider having
It seems to me that Religious people need not disagree with science (since they are by virtue of faith, capable of the "doublethink" needed to maintain two incompatible world views simultaneously).
However, scientists must, if they are competent and sincere in their convictions, be opposed to any form of Faith. Faith cannot be experimentally tested/falsified, and therefore has no place in rational thought.
The only exception to this would be the "Gnostic" religions (often thought of as "Eastern" or "Mystical", although the original Christians eg St Paul were Gnostics too). These do not literally believe in a Personal Saviour God, but in that, by finding Enlightenment, one actually *becomes* a god. [Please excuse the simplification of Gnosticism here!]
I do in principle support the TV license, because the BBC (especially the Radio - World Service + Radio 3,4) is extremely good. However, I think that the TV license is a bad way to do it, and it ought to be included in general taxation. Reasons:
1)The license is there as a "tax of choice". So, if you don't have a TV, then you don't pay (not even if you do listen to the radio). This made sense in 1960 - but not so much now, when virtually everyone has a television.
2)The license collection is extremely inefficient. It involves hassle for the licensor, a draconian TV licensing authority (who make an enormous nuisance of themselves if you don't actually own a TV), and you cannot legally purchase any TV-capable equipment without giving a name and address to the retailer. [Yes, this is outrageous.] Enforcement and collection must cost a significant proportion of the total fee!
3)With the exception of pensioners, the TV license is the same for everyone. Yet, some can afford to pay more than others.
4)On principle: As a citizen, I have a natural right to my share of the RF spectrum - and to operate a Radio receiver!
However, the idea of a centrally funded broadcaster is a good one: it means that the quality of output need not go into freefall in the pursuit of ratings.
Why didn't you use Knoppix or something to get the music off it? Apple enforce a policy that you can only sync an iPod with one computer, but GTKPod doesn't. Or even use iPod Linux.
:-)
P.S. The iPod Linux GUI for the nano beats the native one by some way
What I still wonder is why Ffox has got so much Gnome in it. Nothing against GTK (mostly), but a few things really really bug me:
1)Some of the options are controlled by prefs within the gnome-control center. There is no way to set/override them from within Firefox itself, nor is there even a hint as to where to find the control. By default, Firefox sends mailto:s to Evolution, not Thunderbird!
2)Why, oh why did they abandon the rather good native file-widget in favour of the horrendous abomination that is the new GTK2 filepicker? This is very limited, and worse, it takes 30 seconds to open, during which time, the application stalls!
Indeed. I cut a slightly long story short there - the machine in question was an expensive desktop (high-end parts, reputable manufacturers...) - but it got pressed into emergency service as a server...
> I think your eyes might have put the word "don't" into my post. I was joking that an informed post that actually helps ;)
> answer the question is quite rare on an 'ask slashdot'. It's usually a knee-jerk recommendation for $TECH_DU_JOUR
D'oh! I apologise.
Richard
No, I'm not lost, and I think I do know what I'm talking about. What I'm basically trying to say is this:
Q: How do I get from A to B ?
A: You don't want to start from here.
In other words, the original poster is asking about a RAID solution for a particular problem. However, I don't think that RAID is, in fact, the best way to go.
It costs roughly the same to have either 1 RAID box, or 2 separate servers without RAID. The former does give better performance, but this will probably be dominated by the network performance anyway: a single SATA drive can provide a bitrate faster than 100Mbit ethernet can transmit. The latter provides several extra advantages. In essence, it is "A Redundant Array of Inexpensive Servers".
> 2) Don't use a $15 PSU, you pay for what you get.
Correct - and indeed, I do buy expensive PSUs (and have a pile of never-used ones that came free with the case). However,
even $70 PSUs can fail. RAID is banking on the odds of not loosing >1 HDDs simultaneously, so you are already looking at rare events.
> 3) If you live in a lightning prone area you should invest in some good protection (line conditioner for example).
Yes - but everywhere has *some* lightening, and you might just be unlucky.
> 5) What do you think the R is RAID is for? Redundancy. You'd need some horribly catastrophic
> odds to lose all the redundant hard drives before you replace any.
Of course. But it might be cheaper (and more redundant) to have 2 separate machines, each with 1 HDD in it. My view is that this gives you lower overall risk. Sure, if you have a failure, it will take a bit more time to recover, but not much. All RAID protects you from is age-related disk failure.
> 6) Your data is still there.
No, it might not be. A recent problem I had on an (expensive) server was that a stick of RAM developed a fault, causing the OS to trash the filesystem. I was lucky: ReiserFS can re-build the journal from the files themselves. But if you use RAID, then this sort of crash will give you 6 identical copies of destroyed data!
Lastly, I would recommend nightly rotating snapshots to protect against user-error. That link I gave has an *extremely* clever solution.