Time to move to Savannah
on
Linuxworld Fun
·
· Score: 4, Informative
For those of you that use sourceforge for their free software projects, it looks like it's time to move to savannah.gnu.org.
In case you're wondering, the gnu.org in there does not imply that your project needs to be under the GPL/LGPL --- any Free Software projects are welcome.
Why would you want to move? Well, from what I hear, extracting some of your meta-data is already hard/imposible from Sourceforge --- this seems like a trend that is likely to continue, so perhaps you should get out while you still can.
At least you can be sure that the Free Software Foundation won't pull any similar tricks.
Why does Debian always come with such old stuff? Fine, maybe KDE 2.2 is more stable than 3 but still...
Well, we test it until we're happy with it. This takes time. Time during which newer packages get released. Packages that generally get uploaded to Debian unstable (or in the case of KDE3, Debian experimental), then people like yourself that want to run that package have to go through the "trauma" of editing one line in one file (or using a cute point and shoot front end), and then they can pick and mix what software they want. Is that really so difficult to understand?
Other people may release things because their marketing department tells them to. Debian has the luxury of not having a marketing department, so we don't need to do that. That's why we use the word stable, to mean stable.
OK, it takes longer than it might if we were all being paid to do this, but who cares. It's so easy to select the versions of software you want, and select the level of instability you can live with, that there is no issue to be resolved here.
For example, you might want to run a known good version of postgreSQL on you machine to look after you accounts database (don't want to loose that) but be willing to run a cute KDE3 based database access tool to view that data, on the assumption that if it crashes, it probably won't have chance to ask postgreSQL to do anything too bad --- your choice, go right ahead. apt-get will even keep track of that, adn continue to upgrade postgeSQL from the stable branch, and kdata (or whatever) from unstable, or even http://freds.kde.emporium.com/debian/ say.
In summary, give Debian a try if you fancy it, but please stop coming up with spurious excuses not to, if you don't.
so, you want to recompile package foo, well first you need some source
apt-get source foo
or maybe some newer source
apt-get source foo/unstable
or maybe you need the source that became available 5 minutes ago, in which case you do one of the above, drop the new tarball in the current directory, go into the old Debian source directory, and run uupdate and maybe fix some patching problems in the new directory that got created for you.
Next, you need to build this stuff, so let's get in the source directory:
cd foo-1.2.3
oh, but we might need some other development libraries to build this, so lets grab what we need
apt-get build-dep foo
that's better, now we can tweak some source or options maybe, and make ourselves a package
debuild
right, so now we have a new package, so we install it
sudo dpkg -i../foo-1.2.3*deb
and it gets installed (or maybe it has some dependencies, if you got this far, you can work it out yourself) just like it was an official Debian package, which means you get to remove it cleanly, etc. if the need arrises.
Who ever said source was difficult to play with in Debian? Debian is by developers, for developers -- we like source. That's why we're into Free Software.
I've just been informed that apt-get dist-upgrade is in fact not recomended, because if you don't know exactly what you're doing it has a tendency to remove half the packages on your system, and not bother upgrading. So you are left with a perfectly valid, but somewhat emaciated instalation at the end.
dselect on the other hand makes smarter decissions about things like "replaces" and "suggests" package interdependancies, and lets you resolve conflicts before going for the upgrade, so that is the recomended route, unles you happen to know better.
Of course, I didn't know that, because I know how to avoid getting bitten by apt-get, so don't tend to notice its teeth.
Sorry about the previous mis-information, please igore it (feel free to mod it into oblivion)
Most Debian users have been running almost exactly this for ages, as "woody" or "testing" --- It works rather nicely.
That's the point with Debian, you get to choose when to abandon rock solid reliability in favour of shiney new features, and most of the time it doesn't even degrade your reliability.
People seem to miss the point that Debian stable is meant to be STABLE. If you want to install a box in the middle of the desert, 100s of miles from anywhere, then you want it stable, and you don't give a damn about which version of KDE you use. If you want the latest KDE, just grab it out of unstable, or of off the KDE Debian package maintainer's bleeding edge archive --- It's your chouce. If it kills your machine, you get to press the reset button, but generally it won't, and if it does, it's what you decided you wanted, and you get the joy of filing a bug report, and helping to fix the problem.
Would I recomend a switch? Well, if you're happy with SuSE, and you don't care about Software Freedom, then SuSE is a fine distribution, unfortunately YaST is non-free software, so I'll never use it, and SuSE doesn't have apt-get, which makes any other feature they might have pale into insignificance in my opinion. Obviously, I am biased though:-)
50 a year, I agree, is nothing to get upset about.
If on the other hand you had received 35 spams in the last 24 hours, you might like me be forced to spend time fighting it, which is time that you cannot spend doing something useful.
Spamassassin has caught 3917 spams for me since mid April.
Obviously, the advent of spamassassin helps, but it is also teaching the spammers to be more devious, so will eventually result in the equivalent of multi-resistant bacteria.
Actually, looking at this, rather too many of those 35 spams got through the net. Must be time to apt-get install spamassassin/unstable
Anyway, if someone was stealing your time, in significant quantities, you would get rapidly upset about it. If you told me that you'd been mugged, I'd be sympathetic despite the fact that I've never been mugged.
The fact that I probably spend a few hours a month dealing with this, as do many others means that the spammers are depriving people of many lifetimes a year. This is mass murder, just spread out over a wider population, so the average loss is lower.
Postal junk mail goes on my fire, and warms my house, so doesn't really worry me (apart from the damage to the environment)
Windependance != Freedom, it seems
on
Windependence Day
·
· Score: 4, Insightful
I think it's a shame that most of the prizes are partly or completely non-free software.
It brings to mind what I believe is a translated polish proverb:
A change of leaders is the joy of fools.
To exchange one set of proprietary terms & conditions for another, even if the new ones are a lot less aggravating, when one could instead have grasped Freedom in perpetuity, is a wasted opportunity in my opinion.
Well since you don't buy Free Software from the developers, you only need to determine if the people selling you support, and/or the CDs are criminals, so there is no problem.
Every other industry in the world has had to do exactly this. The software industry is no different.
You have just decided to call software creation an industry, and thereby lump it together with other aspects of human endeavour covered by patents.
How about if we decide that programmers are a profession, like lawyers? I would say that you could draw fairly close parallels between the behaviours, and working conditions of both professions.
Perhaps you should consider what patents would do to the practice of law, if the originators of new legal tactics were allowed to patent and license them to lawyers who wished to use them in future.
Computer algorithms are NOT the same as mathematical formulae.
That looks suspiciously like an unsupported assertion to me. Progarmming languages are a form of mathemetical notation. That's why people go to the effort of proving various things, like Turing Completeness, about them, using mathematical techniques.
Somewhere you have to acknowledge that software is unique and nothing like it has ever existed before.
Before when? Before Ada Byron, Lady Lovelace, wrote programs for a Babbage's analytical engine? How about the Russian Peasant Multiplication Method (as used in most RSA implementations) which you can probably guess was originally used by Russian peasants, before they'd even considered the possibility of computers? Do the poems used by Ancient Chinese abacus operators to perform complex calculations, count as programs? (I'd say yes)
In other words, no I don't think I have to acknowledge anything of the sort.
Mathematical formulae are just ways of communicating ideas. Algorithms are a subset of those ideas that tell you how to make a machine perform a useful function.
Glad we managed to agree on that at least. "Algorithms are a form of mathatical formula."
Are you therefore saying that mathematics should be patentable as long as it's useful? In other words "Only useless mathematics should be exempt" ?
Do you think that patenting large primes is a reasonable thing to do? How about small primes?
To sumarise, I don't think mathematics should be subject to patents, and I think computer programs can be considered to be a branch of mathematics.
[P.S. you don't seem to have come up with a "good software patent" yet]
Boy there is a lot to respond to in this one, but here goes...
Oops, I think I just did it again, sorry. Hope you're not getting too bored.
However, if you recall that patents protect functionality and that a good (and valid) patent must disclose the functionality in sufficient detail so that a person of ordinary skill in the art can make the invention by reading the patent,
I do recall that, but I am yet to see a patent that is both comprehensible, and not for something blindingly obvious.
That is from the point of view of someone with a Computing Science degree (including a minor module on UK IP law) with 16 years subsequent professional experience as a programmer.
I'm yet to see a good software patent. That being the case, it's difficult to be sympathetic with arguments that assume that good patents are the norm, or even likely.
An example of a "good" software patent would be really helpful, to contrast with the legion of bad patents that have been widely aired.
The point I was trying to make about non-patenting territories is that if I live in such a territory (which I sort of do, but probably not for much longer), and if I don't care if people use my ideas, as long as they don't try to stop me using theirs (which I may or may not have arrived at independently) then I've got no reason to be interested in the patent system, have I?
Let's say I come up with something very clever, and I don't bother to patent it. Someone in the US downloads my implementation of my idea, writes it up as a patent, and applies for it, and then starts enforcing it, possibly even pursuing people that are using my implementation. Is that a realistic scenario? What should I then do to rectify the situation? Attempt to overturn the patent, by showing prior art? Am I right in thinking that overturning a bad patent costs about $1.5M these days? Don't you think that scenario highlights some fundamental problems with the US patent system?
But -- think of the possibility of a closed-source company liking an Open Source idea so much that they put resources into implementing the patented idea. The OPL would require them to then release the source.
You're thinking like a Patent Lawyer, while I'm thinking like a Computer Scientist. I don't think anyone should have the right to stop me using the best solution to a problem that I can think of, even if they thought of it first, because I consider algorithms to be equivalent to mathematical formulae, and as such discoveries, not inventions. Equally, I don't care if I thought of something first, I don't feel I have the right to stop other people using that idea. I certainly don't want to think of something only to discover I'm not allowed to use it because some government decided to issue a patent on it. I REALLY don't want to have to cross-check every idea I have and then have to choose between licensing the patent or producing a (possibly technically inferior) alternative solution, just because someone else thinks like me, but was quicker at getting to the patent office.
The current law in Europe states something like "programs for computers are not patenable as such", which the folks at the EPO cheerfully ignore and go ahead and issue software patents all the same (BTW, you couldn't explain what "as such" actually means there could you?). What nobody has succeeded in explaining to me is why we should contemplate allowing Software Patents at all. Obviously some (but not all) Lawyers like the idea, but the patent system is supposed to be for the benefit of society, or the industry, or something, but we Eurpean computer scientists have been cheerfully innovating without them, so why burden us with them now?
OK, I shouldn't have dived straight in with the inflammatory language, sorry about that. I'm afraid that software patents have always upset me quite deeply, so I tend to lose my rag about them.
I don't see that patenting Open Source ideas will encourage the discovery of a single algorithm that would not otherwise have been thought of, so all that would occur is that some of the limited monetary resources of the Free Software community would be diverted to propping up the structure of the patent office.
You say that $350 is a small price to pay. How many patentable ideas do you think are contained in the body of Free Software code Debian GNU/Linux (which only integrates a fraction of the software available) currently runs to about 5GB of compressed source code --- there are a lot of ideas in there, and that's only the tip of the iceberg.
Assuming that someone decided to go through that and patent all the as yet unpatented ideas, what purpose would that serve?
How about ideas that were thought up by people in teritories that do not recognise sotware patents?
As a non-lawyer, I'd assume that that fact that all that source has been published would mean that it would act as prior art, but I have a feeling that the USPTO allows after the event registration (feel free to enlighten me).
If it is prior art, haw can one ever register a Free Software idea, given that the publication is inherent in the development process.
If it is not prior art, what is to stop someone else from claiming to have originated these ideas, and patenting them themselves?
Assuming that it turns out that the an was patentable, and what you suggest was done, in what sense would it be more useful that simply publishing the code under the GPL (apart perhaps from the fact of preempting someone else's patenting of the same idea)?
I can see that it would allow one to prevent others from using that idea, but that is pretty much entirely against the principles of Free Software (if someone else wants to use an idea from one of my programs, and they go to the effort of reimplementing it, they're welcome to it).
I can also see that it is likely to be much more difficult to win a case relying on patent law, rather than copyright law, which means that the likelihood is that the party with the deeper pockets has an unfair advantage, which is not likely to be the Free Software developer.
Was that better? (I've calmed down, now that they've given me one of my dried frog tablets:-)
As to your point about the patent system being there to stay, it seems that there are likely to be parts of the world where that will not be the case for some time, and in places like South Africa and India they've been moving in the opposite direction (at least as far as pharmaceuticals are concerned). Reform of the USPTO has even been rumoured, and parts of Europe seem less than keen to get involved in this form of silliness.
Yes I have (to my shame;-) I'm an RHCE & an RHCX (i.e. I get to invigilate RHCE exams).
The RedHat test is a pretty decent test, unlike the one you describe. There are one or two bits that I found anoying (as a Debian user) because they appear to have decided that rather than fixing some things that I'd consider to be usability bugs, they'll just add the work-around proceedure into the test, but that only accounts for a couple of percent.
I'd be unhappy about letting someone that failed the RHCE be a sysadmin on anything I cared about.
Most people that pass it, and also get a year or so's experience around the same time, can probably be trusted to have a clue. They also would have little trouble administrating a Debian system, or any other GNU/Linux system, once they got over their confidence problem about it not being RedHat.
After all, I got 99.6% on the RHCE exam, and the first time I had installed RedHat was 4 days earlier at the start of the fast-track course. It's really not that different from Debian (/etc/rc.d/nit.d vs/etc/init.d, rpm vs dpg, um, what else?)
As a lawyer, how would you like to have to check each tactic you were planing to use in defending one of your clients, before actually using it, in order to check that it had not been patented by another lawyer?
That what the patent industry is trying to do to us. They (you?) pretend they're are doing us a favour (chanting "Innovation", "Protection of Property" etc), but in fact you are burdening us with the extra workload of (if anyone could be bothered) having to check every line of code against a patent database, or in the absence of that, getting sued for thinking of an idea after (of sometimes several years before) someone else.
Not only that, but the patents are worded to ensure that they provide almost no information whatsoever to someone interested in the technique they describe, so the claimed goal of driving forward the state of the art is total nonsense (can you cite a single instance of a Computer Scientist referring to patents in order to learn a novel technique? I doubt it).
Software patents are a government authorised tax on the software industry to make monopolistic corporations and patent lawyers rich. They have no positive effect on the state of the art in the field of computing whatsoever.
Unfortunately the patent lawyers are in charge of the patent offices, and those arms of government that are supposed to regulate them, so we're likely to end up as thoroughly shafted in Europe as is the current situation in the USA.
Having said all that, patents on other, material inventions seem totally fine to me, so I'm not saying patents or patent lawyers are evil per se, just the ones that try to take my (software) toolbox away, when I made my toolbox myself.
There will be no restrictions on the use of TE by the Linux open source community
Before a colo company, or anyone else uses the technology commercially, it will be necessary to negotiate a license with Secure Computing
Those two statements are incompatible.
If there are no restrictions for the open source community then that includes there being no restriction on subsequently using the software for commercial use.
I make my living out of selling support on Free Software, but I'm a Debian Developer, so is my usage restricted (commercial use) or unrestricted (member of the community)?
If they decide to try to enforce this patent, they will reap the whirlwind.
Reading section 7 of the GPL:
For example, if a patent license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program.
[emphasis mine]
As I see it, if they give me a copy of their patched code under the GPL, as a member of the community, and I pass it on to a commercial entity that they were trying to talk into paying for the patent, they lose, becasue they are only allowed to distribute Linux as long as they abide by the GPL.
I'm pretty certain that that means that if they attempt to restrict some people's (i.e. commerce's) right to use their code with their patent, they lose the right to distribute Linux (and any other GPLed code they've patched, with or without their code), permanently, until they repent and are then forgiven by all the relevant copyright holders.
Seeking that forgiveness may take a little time (say, forever) during which period they my find their precious Intellectual Property will be worth a little less than they had hoped, given that it will have become a patch looking for a kernel to be applied to.
Because I'm always installing Linux for clients, RedHat is always specified, so I have no choice,...
I too regularly install GNU/Linux for clients, and more often than not they specify RedHat (occasionaly SuSE), but I've not installed anything other than Debian for years (except during RHCE/RHCX courses;-), so you do have a choice.
The trick is to ask them why they specified RedHat. Most of them will cheerfully admit that they said that because it's the only distro.they've heard of so they were saying "RedHat Linux" in the same way they might say "Microsoft DOS", never realising there was a "DR-DOS" (once upon a time;-).
It doesn't normally take too much effort to convince them that they are paying me for my expertise, so if I recommend a particular flavour they might as well listen.
The only time I'd take such a request seriously would be if they were already a RedHat shop, and had a lot of in-house RedHat expertise.
The last time that was claimed the "expert" turned out to be clueless, and the existing RedHat systems so broken (9GB swap, 800MB root, permissions all +rwx) that I ended up having to reinstall them anyway --- they're a Debian shop now;-)
I've since decided that any RedHat shop that decides to hire me in, is probably not full of experts (otherwise they'd do it themselves), so take a lot of convincing that going with the flow is the wise thing to do.
If you have a good technical reason not to install RedHat, and you can justify it, give it a try.
The worst that happens is they say no (and you get to look smug if their decission bites them).
The best that might happen is that they decide to respect your opinion, which bodes well for the future business relationship, and means you get to work on the system you feel best fits the problem, which avoids stress and frustration.
Re:software developers v users - the battle contin
on
United Linux is Here
·
· Score: 2
OK, I can see that the gawk thing pissed you off, but it's actually pretty simple to either make your package depend on gawk (if you package it) or put a note in the README saying "This package depends on gawk" if that's the case.
You were saying that Debian should have provided a symlink from awk to gawk though, which is just plain wrong I'm afraid. If the user chose not to install gawk, then you don't get gawk. End of story.
Were you really using gawk specific features? or would awk have done the job for you? If the latter, you should have been invoking it as awk.
Having said all that, I'd imagine that it's pretty easy to miss that subtlety, in which case it would be pretty bloody frustrating, so I do understand where you're coming from. Just a shame nobody bothered to point this out to you earlier really.
Cheers, Phil.
Re:software developers v users - the battle contin
on
United Linux is Here
·
· Score: 2
That's what 'de facto' means, or didn't you get the memo?;-) De facto standards come from the most popular Linux distributions.
[aside: Debian is one of the most popular distros --- does that mean that Debian is the de facto standard?]
So, let's take something that you presumably would have considered to be a de facto standard at the time. Linuxconf was the standard configuration tool on RedHat for a while. So, do RedHat get to change that "standard" just because they originated it, or were they non-standard then, or are they non-standard now?
I'm not criticising RedHat specifically, since the other commercial distros do similar things, but if they change their defaults over time, then is the de facto standard limited to the logo on the startup screen?
Debian on the other hand does not have the organisational structure to allow it to chop & change rapidly, so any change ends up having to take account of all the people that will choose not to follow the change, resulting in a more robust distribution.
Chaos brings forth order.
Consistently difficult to install. Consistently Balkanized. Great. No wonder Debian is so popular. It is the Archie Bunker of Linux distributions: you can always rely on it to behave predictably.
Well, thanks for that ringing endorsement. I didn't ask you to use it did I?
Perhaps you should see the dozens of e-mails I received from Debian users reporting the "'which gawk' returned an error" message from my software. Perhaps you should tell those users that gawk was indeed there when each one specifically told me it wasn't. It wasn't there when I installed Debian in late 2001
Ah, sorry, you're complaining that gawk wasn't present when awk was? I read you to mean the reverse.
If you want gawk on Debian you can run:
apt-get install gawk
but you have the freedom not to, and there are other alternative implementations (i.e. mawk) so there's no need to do so.
If you are saying that someone should provide a link/usr/bin/gawk that points at/usr/bin/awk, which may in fact be a link to/usr/bin/mawk, then you're wrong.
If you are invoking/usr/bin/gawk because you are relying on features unique to gawk, then you would be disappointed to find that you were actually running mawk, true?
If on the other hand you are not using gawk specific features, you should be calling/usr/bin/awk.
Either way the problem you describe is not a fault in Debian, it is either a bug in your code, because you were saying gawk when you meant awk, or it's a failure in your package dependencies (if you packaged it) or in you readme, which should have indicated that your program depended on gawk.
I notice you didn't bother tackling me on cramfs.
You were saying that Debian was non-standard --- I don't think that what a particular version of a particular other distribution is doing counts as standard, so I didn't bother considering it. I don't like initrd/cramfs anyway --- it's just another thing to go wrong at boot time, so I don't use it on any of my machines. We use it on the install floppies/CDs to make things fit, but I don't see that that's something to get overly stressed about, and why anyone should care, I don't know.
If you're so excited about this de facto standard of yours, what do you think of using file system labels in fstab, instead of devices? The majority of RedHat systems that are running don't use that scheme, so are RedHat breaking the standard by introducing change, or are all historical RedHat installs now non-standard? How about Grub vs. Lilo? Do you want me to go on?
P.S. Read your Gun control post. --- glad we can agree on something, anyway:-)
Re:software developers v users - the battle contin
on
United Linux is Here
·
· Score: 4, Insightful
So you're saying that Debian doesn't follow "standards" (presumably meaning "whatever RedHat decided to do this week"), and that you don't like moving targets?
Debian can hardly be described as a moving target , given it's 18-month release cycle.;-)
If you want to measure consitency of behaviour over time, I think you'll find that Debian would win hands down.
If you want to measure some sort of "least surprise quotient" when a random *nix user comes across a distro for the first time, I think Debian would win again.
As for the awk link, mawk has been providing such a link since Mar 1997, and gawk since before Dec 1995, so I don't know when you last looked, but perhaps you should look again.
isn't the rainbow actually working on the dropplet scale, rather than the molecular scale.
In other words, is it not the fact that each raindrop is a sphere with a refractive index different from that of air, and that each one projects a light cone where the colour is determined by the angle to the original incoming light.
Also, you might want to mention that there is a second rainbow, resulting from the light internaly reflecting in each raindrop, with the colours in reverse order to the main rainbow, at a wider angle than the main one. (there's probably a third, resulting from a double internal bounce, but I don't think it's visible to the human eye in daylight, so you'll never see it)
In fact, when you analyse most of these things, it's often surprising how few coincidences happen.
Once you calculate in the number of songs you think of in a day, and the number of opportunities you have to hear songs, and the fact that you will tend to think of songs that are in some way prompted by some external influence, and the people that choose the songs to play on the radio are likely to have some of those inputs in common with you, you might come to the conclusion that it's not that surprising after all.
Also, we tend to forget all the times when the coincidence didn't occur, because that is simply not memorable, so over time we are left with a biased series of memorable coincidences.
It is not evident (to me, anyway) that theories about evolution and creation are really even scientific theories, because they're not directly testable, though they are based on scientific understanding of underlying physical processes which are separately testable.
I presume that you're saying that evolution is not testable, on the basis that it's all already happened, and we're just looking at the evidence left behind.
That would be fair enough, if it were not for the fact that we are still witnessing evolution in action, and so it can be observed.
There is the example of moths during the Industrial revolution, where the soot produced by the factories killed the lichen on the trees, and stained the bark. The moths had been white prior to this happening, which was good camoflage when sitting on the light coloured bark --- once the trees changed colour, the lighter moths were easy targets for the birds, and so the moths very quickly evolved to be black --- same species, different colour.
The factories have since cleaned up their act (or shut down) and the trees have lichen again, and the moths have evolved back to being white.
Slashdot Mirror
For those of you that use sourceforge for their free software projects, it looks like it's time to move to savannah.gnu.org.
In case you're wondering, the gnu.org in there does not imply that your project needs to be under the GPL/LGPL --- any Free Software projects are welcome.
Why would you want to move? Well, from what I hear, extracting some of your meta-data is already hard/imposible from Sourceforge --- this seems like a trend that is likely to continue, so perhaps you should get out while you still can.
At least you can be sure that the Free Software Foundation won't pull any similar tricks.
Why does Debian always come with such old stuff? Fine, maybe KDE 2.2 is more stable than 3 but still...
Well, we test it until we're happy with it. This takes time. Time during which newer packages get released. Packages that generally get uploaded to Debian unstable (or in the case of KDE3, Debian experimental), then people like yourself that want to run that package have to go through the "trauma" of editing one line in one file (or using a cute point and shoot front end), and then they can pick and mix what software they want. Is that really so difficult to understand?
Other people may release things because their marketing department tells them to. Debian has the luxury of not having a marketing department, so we don't need to do that. That's why we use the word stable, to mean stable.
OK, it takes longer than it might if we were all being paid to do this, but who cares. It's so easy to select the versions of software you want, and select the level of instability you can live with, that there is no issue to be resolved here.
For example, you might want to run a known good version of postgreSQL on you machine to look after you accounts database (don't want to loose that) but be willing to run a cute KDE3 based database access tool to view that data, on the assumption that if it crashes, it probably won't have chance to ask postgreSQL to do anything too bad --- your choice, go right ahead. apt-get will even keep track of that, adn continue to upgrade postgeSQL from the stable branch, and kdata (or whatever) from unstable, or even http://freds.kde.emporium.com/debian/ say.
In summary, give Debian a try if you fancy it, but please stop coming up with spurious excuses not to, if you don't.
so, you want to recompile package foo, well first you need some source
../foo-1.2.3*deb
apt-get source foo
or maybe some newer source
apt-get source foo/unstable
or maybe you need the source that became available 5 minutes ago, in which case you do one of the above, drop the new tarball in the current directory, go into the old Debian source directory, and run uupdate and maybe fix some patching problems in the new directory that got created for you.
Next, you need to build this stuff, so let's get in the source directory:
cd foo-1.2.3
oh, but we might need some other development libraries to build this, so lets grab what we need
apt-get build-dep foo
that's better, now we can tweak some source or options maybe, and make ourselves a package
debuild
right, so now we have a new package, so we install it
sudo dpkg -i
and it gets installed (or maybe it has some dependencies, if you got this far, you can work it out yourself) just like it was an official Debian package, which means you get to remove it cleanly, etc. if the need arrises.
Who ever said source was difficult to play with in Debian? Debian is by developers, for developers -- we like source. That's why we're into Free Software.
Er, correction to that.
I've just been informed that apt-get dist-upgrade is in fact not recomended, because if you don't know exactly what you're doing it has a tendency to remove half the packages on your system, and not bother upgrading. So you are left with a perfectly valid, but somewhat emaciated instalation at the end.
dselect on the other hand makes smarter decissions about things like "replaces" and "suggests" package interdependancies, and lets you resolve conflicts before going for the upgrade, so that is the recomended route, unles you happen to know better.
Of course, I didn't know that, because I know how to avoid getting bitten by apt-get, so don't tend to notice its teeth.
Sorry about the previous mis-information, please igore it (feel free to mod it into oblivion)
3.5 is a thowback to the 2.2 release notes, at which time apt-get was still experimental (although it was rock solid back then too to be honest)
Just use the 3.3 procedure, or if you're worried, wait until someone gets round to fixing it (later today I expect)
Most Debian users have been running almost exactly this for ages, as "woody" or "testing" --- It works rather nicely.
:-)
That's the point with Debian, you get to choose when to abandon rock solid reliability in favour of shiney new features, and most of the time it doesn't even degrade your reliability.
People seem to miss the point that Debian stable is meant to be STABLE. If you want to install a box in the middle of the desert, 100s of miles from anywhere, then you want it stable, and you don't give a damn about which version of KDE you use. If you want the latest KDE, just grab it out of unstable, or of off the KDE Debian package maintainer's bleeding edge archive --- It's your chouce. If it kills your machine, you get to press the reset button, but generally it won't, and if it does, it's what you decided you wanted, and you get the joy of filing a bug report, and helping to fix the problem.
Would I recomend a switch? Well, if you're happy with SuSE, and you don't care about Software Freedom, then SuSE is a fine distribution, unfortunately YaST is non-free software, so I'll never use it, and SuSE doesn't have apt-get, which makes any other feature they might have pale into insignificance in my opinion. Obviously, I am biased though
50 a year, I agree, is nothing to get upset about.
If on the other hand you had received 35 spams in the last 24 hours, you might like me be forced to spend time fighting it, which is time that you cannot spend doing something useful.
Spamassassin has caught 3917 spams for me since mid April.
Obviously, the advent of spamassassin helps, but it is also teaching the spammers to be more devious, so will eventually result in the equivalent of multi-resistant bacteria.
Actually, looking at this, rather too many of those 35 spams got through the net. Must be time to apt-get install spamassassin/unstable
Anyway, if someone was stealing your time, in significant quantities, you would get rapidly upset about it. If you told me that you'd been mugged, I'd be sympathetic despite the fact that I've never been mugged.
The fact that I probably spend a few hours a month dealing with this, as do many others means that the spammers are depriving people of many lifetimes a year. This is mass murder, just spread out over a wider population, so the average loss is lower.
Postal junk mail goes on my fire, and warms my house, so doesn't really worry me (apart from the damage to the environment)
I think it's a shame that most of the prizes are partly or completely non-free software.
It brings to mind what I believe is a translated polish proverb:
A change of leaders is the joy of fools.
To exchange one set of proprietary terms & conditions for another, even if the new ones are a lot less aggravating, when one could instead have grasped Freedom in perpetuity, is a wasted opportunity in my opinion.
True, but since there are many reasons, people are switching:
as is demonstrated by the fact that some people are now using GNU/Linux, and before they were not, and that that number is going up, not down.
Well since you don't buy Free Software from the developers, you only need to determine if the people selling you support, and/or the CDs are criminals, so there is no problem.
From the interview:
IDG: How about just the idea of having an open source browser, the Opera Web browser for instance.!
Doh!
Every other industry in the world has had to do exactly this. The software industry is no different.
You have just decided to call software creation an industry, and thereby lump it together with other aspects of human endeavour covered by patents.
How about if we decide that programmers are a profession, like lawyers? I would say that you could draw fairly close parallels between the behaviours, and working conditions of both professions.
Perhaps you should consider what patents would do to the practice of law, if the originators of new legal tactics were allowed to patent and license them to lawyers who wished to use them in future.
Computer algorithms are NOT the same as mathematical formulae.
That looks suspiciously like an unsupported assertion to me. Progarmming languages are a form of mathemetical notation. That's why people go to the effort of proving various things, like Turing Completeness, about them, using mathematical techniques.
Somewhere you have to acknowledge that software is unique and nothing like it has ever existed before.
Before when? Before Ada Byron, Lady Lovelace, wrote programs for a Babbage's analytical engine? How about the Russian Peasant Multiplication Method (as used in most RSA implementations) which you can probably guess was originally used by Russian peasants, before they'd even considered the possibility of computers? Do the poems used by Ancient Chinese abacus operators to perform complex calculations, count as programs? (I'd say yes)
In other words, no I don't think I have to acknowledge anything of the sort.
Mathematical formulae are just ways of communicating ideas. Algorithms are a subset of those ideas that tell you how to make a machine perform a useful function.
Glad we managed to agree on that at least. "Algorithms are a form of mathatical formula."
Are you therefore saying that mathematics should be patentable as long as it's useful? In other words "Only useless mathematics should be exempt" ?
Do you think that patenting large primes is a reasonable thing to do? How about small primes?
To sumarise, I don't think mathematics should be subject to patents, and I think computer programs can be considered to be a branch of mathematics.
[P.S. you don't seem to have come up with a "good software patent" yet]
Boy there is a lot to respond to in this one, but here goes ...
Oops, I think I just did it again, sorry. Hope you're not getting too bored.
However, if you recall that patents protect functionality and that a good (and valid) patent must disclose the functionality in sufficient detail so that a person of ordinary skill in the art can make the invention by reading the patent,
I do recall that, but I am yet to see a patent that is both comprehensible, and not for something blindingly obvious.
That is from the point of view of someone with a Computing Science degree (including a minor module on UK IP law) with 16 years subsequent professional experience as a programmer.
I'm yet to see a good software patent. That being the case, it's difficult to be sympathetic with arguments that assume that good patents are the norm, or even likely.
An example of a "good" software patent would be really helpful, to contrast with the legion of bad patents that have been widely aired.
The point I was trying to make about non-patenting territories is that if I live in such a territory (which I sort of do, but probably not for much longer), and if I don't care if people use my ideas, as long as they don't try to stop me using theirs (which I may or may not have arrived at independently) then I've got no reason to be interested in the patent system, have I?
Let's say I come up with something very clever, and I don't bother to patent it. Someone in the US downloads my implementation of my idea, writes it up as a patent, and applies for it, and then starts enforcing it, possibly even pursuing people that are using my implementation. Is that a realistic scenario? What should I then do to rectify the situation? Attempt to overturn the patent, by showing prior art? Am I right in thinking that overturning a bad patent costs about $1.5M these days? Don't you think that scenario highlights some fundamental problems with the US patent system?
But -- think of the possibility of a closed-source company liking an Open Source idea so much that they put resources into implementing the patented idea. The OPL would require them to then release the source.
You're thinking like a Patent Lawyer, while I'm thinking like a Computer Scientist. I don't think anyone should have the right to stop me using the best solution to a problem that I can think of, even if they thought of it first, because I consider algorithms to be equivalent to mathematical formulae, and as such discoveries, not inventions. Equally, I don't care if I thought of something first, I don't feel I have the right to stop other people using that idea. I certainly don't want to think of something only to discover I'm not allowed to use it because some government decided to issue a patent on it. I REALLY don't want to have to cross-check every idea I have and then have to choose between licensing the patent or producing a (possibly technically inferior) alternative solution, just because someone else thinks like me, but was quicker at getting to the patent office.
The current law in Europe states something like "programs for computers are not patenable as such", which the folks at the EPO cheerfully ignore and go ahead and issue software patents all the same (BTW, you couldn't explain what "as such" actually means there could you?). What nobody has succeeded in explaining to me is why we should contemplate allowing Software Patents at all. Obviously some (but not all) Lawyers like the idea, but the patent system is supposed to be for the benefit of society, or the industry, or something, but we Eurpean computer scientists have been cheerfully innovating without them, so why burden us with them now?
OK, I shouldn't have dived straight in with the inflammatory language, sorry about that. I'm afraid that software patents have always upset me quite deeply, so I tend to lose my rag about them.
:-)
I don't see that patenting Open Source ideas will encourage the discovery of a single algorithm that would not otherwise have been thought of, so all that would occur is that some of the limited monetary resources of the Free Software community would be diverted to propping up the structure of the patent office.
You say that $350 is a small price to pay. How many patentable ideas do you think are contained in the body of Free Software code Debian GNU/Linux (which only integrates a fraction of the software available) currently runs to about 5GB of compressed source code --- there are a lot of ideas in there, and that's only the tip of the iceberg.
Assuming that someone decided to go through that and patent all the as yet unpatented ideas, what purpose would that serve?
How about ideas that were thought up by people in teritories that do not recognise sotware patents?
As a non-lawyer, I'd assume that that fact that all that source has been published would mean that it would act as prior art, but I have a feeling that the USPTO allows after the event registration (feel free to enlighten me).
If it is prior art, haw can one ever register a Free Software idea, given that the publication is inherent in the development process.
If it is not prior art, what is to stop someone else from claiming to have originated these ideas, and patenting them themselves?
Assuming that it turns out that the an was patentable, and what you suggest was done, in what sense would it be more useful that simply publishing the code under the GPL (apart perhaps from the fact of preempting someone else's patenting of the same idea)?
I can see that it would allow one to prevent others from using that idea, but that is pretty much entirely against the principles of Free Software (if someone else wants to use an idea from one of my programs, and they go to the effort of reimplementing it, they're welcome to it).
I can also see that it is likely to be much more difficult to win a case relying on patent law, rather than copyright law, which means that the likelihood is that the party with the deeper pockets has an unfair advantage, which is not likely to be the Free Software developer.
Was that better? (I've calmed down, now that they've given me one of my dried frog tablets
As to your point about the patent system being there to stay, it seems that there are likely to be parts of the world where that will not be the case for some time, and in places like South Africa and India they've been moving in the opposite direction (at least as far as pharmaceuticals are concerned). Reform of the USPTO has even been rumoured, and parts of Europe seem less than keen to get involved in this form of silliness.
Have you ever seen or taken the RHCE tests?
;-) I'm an RHCE & an RHCX (i.e. I get to invigilate RHCE exams).
/etc/init.d, rpm vs dpg, um, what else?)
Yes I have (to my shame
The RedHat test is a pretty decent test, unlike the one you describe. There are one or two bits that I found anoying (as a Debian user) because they appear to have decided that rather than fixing some things that I'd consider to be usability bugs, they'll just add the work-around proceedure into the test, but that only accounts for a couple of percent.
I'd be unhappy about letting someone that failed the RHCE be a sysadmin on anything I cared about.
Most people that pass it, and also get a year or so's experience around the same time, can probably be trusted to have a clue. They also would have little trouble administrating a Debian system, or any other GNU/Linux system, once they got over their confidence problem about it not being RedHat.
After all, I got 99.6% on the RHCE exam, and the first time I had installed RedHat was 4 days earlier at the start of the fast-track course. It's really not that different from Debian (/etc/rc.d/nit.d vs
Patents on software are a moronic idea.
As a lawyer, how would you like to have to check each tactic you were planing to use in defending one of your clients, before actually using it, in order to check that it had not been patented by another lawyer?
That what the patent industry is trying to do to us. They (you?) pretend they're are doing us a favour (chanting "Innovation", "Protection of Property" etc), but in fact you are burdening us with the extra workload of (if anyone could be bothered) having to check every line of code against a patent database, or in the absence of that, getting sued for thinking of an idea after (of sometimes several years before) someone else.
Not only that, but the patents are worded to ensure that they provide almost no information whatsoever to someone interested in the technique they describe, so the claimed goal of driving forward the state of the art is total nonsense (can you cite a single instance of a Computer Scientist referring to patents in order to learn a novel technique? I doubt it).
Software patents are a government authorised tax on the software industry to make monopolistic corporations and patent lawyers rich. They have no positive effect on the state of the art in the field of computing whatsoever.
Unfortunately the patent lawyers are in charge of the patent offices, and those arms of government that are supposed to regulate them, so we're likely to end up as thoroughly shafted in Europe as is the current situation in the USA.
Having said all that, patents on other, material inventions seem totally fine to me, so I'm not saying patents or patent lawyers are evil per se, just the ones that try to take my (software) toolbox away, when I made my toolbox myself.
There will be no restrictions on the use of TE by the Linux open source community
Before a colo company, or anyone else uses the technology commercially, it will be necessary to negotiate a license with Secure Computing
Those two statements are incompatible.
If there are no restrictions for the open source community then that includes there being no restriction on subsequently using the software for commercial use.
I make my living out of selling support on Free Software, but I'm a Debian Developer, so is my usage restricted (commercial use) or unrestricted (member of the community)?
If they decide to try to enforce this patent, they will reap the whirlwind.
Reading section 7 of the GPL:
[emphasis mine]
As I see it, if they give me a copy of their patched code under the GPL, as a member of the community, and I pass it on to a commercial entity that they were trying to talk into paying for the patent, they lose, becasue they are only allowed to distribute Linux as long as they abide by the GPL.
I'm pretty certain that that means that if they attempt to restrict some people's (i.e. commerce's) right to use their code with their patent, they lose the right to distribute Linux (and any other GPLed code they've patched, with or without their code), permanently, until they repent and are then forgiven by all the relevant copyright holders.
Seeking that forgiveness may take a little time (say, forever) during which period they my find their precious Intellectual Property will be worth a little less than they had hoped, given that it will have become a patch looking for a kernel to be applied to.
Because I'm always installing Linux for clients, RedHat is always specified, so I have no choice,...
;-), so you do have a choice.
;-).
;-)
I too regularly install GNU/Linux for clients, and more often than not they specify RedHat (occasionaly SuSE), but I've not installed anything other than Debian for years (except during RHCE/RHCX courses
The trick is to ask them why they specified RedHat. Most of them will cheerfully admit that they said that because it's the only distro.they've heard of so they were saying "RedHat Linux" in the same way they might say "Microsoft DOS", never realising there was a "DR-DOS" (once upon a time
It doesn't normally take too much effort to convince them that they are paying me for my expertise, so if I recommend a particular flavour they might as well listen.
The only time I'd take such a request seriously would be if they were already a RedHat shop, and had a lot of in-house RedHat expertise.
The last time that was claimed the "expert" turned out to be clueless, and the existing RedHat systems so broken (9GB swap, 800MB root, permissions all +rwx) that I ended up having to reinstall them anyway --- they're a Debian shop now
I've since decided that any RedHat shop that decides to hire me in, is probably not full of experts (otherwise they'd do it themselves), so take a lot of convincing that going with the flow is the wise thing to do.
If you have a good technical reason not to install RedHat, and you can justify it, give it a try.
The worst that happens is they say no (and you get to look smug if their decission bites them).
The best that might happen is that they decide to respect your opinion, which bodes well for the future business relationship, and means you get to work on the system you feel best fits the problem, which avoids stress and frustration.
OK, I can see that the gawk thing pissed you off, but it's actually pretty simple to either make your package depend on gawk (if you package it) or put a note in the README saying "This package depends on gawk" if that's the case.
You were saying that Debian should have provided a symlink from awk to gawk though, which is just plain wrong I'm afraid. If the user chose not to install gawk, then you don't get gawk. End of story.
Were you really using gawk specific features? or would awk have done the job for you? If the latter, you should have been invoking it as awk.
Having said all that, I'd imagine that it's pretty easy to miss that subtlety, in which case it would be pretty bloody frustrating, so I do understand where you're coming from. Just a shame nobody bothered to point this out to you earlier really.
Cheers, Phil.
That's what 'de facto' means, or didn't you get the memo? ;-) De facto standards come from the most popular Linux distributions.
/usr/bin/gawk that points at /usr/bin/awk, which may in fact be a link to /usr/bin/mawk, then you're wrong.
/usr/bin/gawk because you are relying on features unique to gawk, then you would be disappointed to find that you were actually running mawk, true?
/usr/bin/awk.
:-)
[aside: Debian is one of the most popular distros --- does that mean that Debian is the de facto standard?]
So, let's take something that you presumably would have considered to be a de facto standard at the time. Linuxconf was the standard configuration tool on RedHat for a while. So, do RedHat get to change that "standard" just because they originated it, or were they non-standard then, or are they non-standard now?
I'm not criticising RedHat specifically, since the other commercial distros do similar things, but
if they change their defaults over time, then is the de facto standard limited to the logo on the startup screen?
Debian on the other hand does not have the organisational structure to allow it to chop & change rapidly, so any change ends up having to take account of all the people that will choose not to follow the change, resulting in a more robust distribution.
Chaos brings forth order.
Consistently difficult to install. Consistently Balkanized. Great. No wonder Debian is so popular. It is the Archie Bunker of Linux distributions: you can always rely on it to behave predictably.
Well, thanks for that ringing endorsement. I didn't ask you to use it did I?
Perhaps you should see the dozens of e-mails I received from Debian users reporting the "'which gawk' returned an error" message from my software. Perhaps you should tell those users that gawk was indeed there when each one specifically told me it wasn't. It wasn't there when I installed Debian in late 2001
Ah, sorry, you're complaining that gawk wasn't present when awk was? I read you to mean the reverse.
If you want gawk on Debian you can run:
apt-get install gawk
but you have the freedom not to, and there are other alternative implementations (i.e. mawk) so there's no need to do so.
If you are saying that someone should provide a link
If you are invoking
If on the other hand you are not using gawk specific features, you should be calling
Either way the problem you describe is not a fault in Debian, it is either a bug in your code, because you were saying gawk when you meant awk, or it's a failure in your package dependencies (if you packaged it) or in you readme, which should have indicated that your program depended on gawk.
I notice you didn't bother tackling me on cramfs.
You were saying that Debian was non-standard --- I don't think that what a particular version of a particular other distribution is doing counts as standard, so I didn't bother considering it. I don't like initrd/cramfs anyway --- it's just another thing to go wrong at boot time, so I don't use it on any of my machines. We use it on the install floppies/CDs to make things fit, but I don't see that that's something to get overly stressed about, and why anyone should care, I don't know.
If you're so excited about this de facto standard of yours, what do you think of using file system labels in fstab, instead of devices? The majority of RedHat systems that are running don't use that scheme, so are RedHat breaking the standard by introducing change, or are all historical RedHat installs now non-standard? How about Grub vs. Lilo? Do you want me to go on?
P.S. Read your Gun control post. --- glad we can agree on something, anyway
So you're saying that Debian doesn't follow "standards" (presumably meaning "whatever RedHat decided to do this week"), and that you don't like moving targets?
;-)
Debian can hardly be described as a moving target , given it's 18-month release cycle.
If you want to measure consitency of behaviour over time, I think you'll find that Debian would win hands down.
If you want to measure some sort of "least surprise quotient" when a random *nix user comes across a distro for the first time, I think Debian would win again.
As for the awk link, mawk has been providing such a link since Mar 1997, and gawk since before Dec 1995, so I don't know when you last looked, but perhaps you should look again.
Another way of putting that might be:
Great way to make sure that the market for cracked DVDs expands massively.
isn't the rainbow actually working on the dropplet scale, rather than the molecular scale.
In other words, is it not the fact that each raindrop is a sphere with a refractive index different from that of air, and that each one projects a light cone where the colour is determined by the angle to the original incoming light.
Also, you might want to mention that there is a second rainbow, resulting from the light internaly reflecting in each raindrop, with the colours in reverse order to the main rainbow, at a wider angle than the main one. (there's probably a third, resulting from a double internal bounce, but I don't think it's visible to the human eye in daylight, so you'll never see it)
Coincidence is not all that surprising.
In fact, when you analyse most of these things, it's often surprising how few coincidences happen.
Once you calculate in the number of songs you think of in a day, and the number of opportunities you have to hear songs, and the fact that you will tend to think of songs that are in some way prompted by some external influence, and the people that choose the songs to play on the radio are likely to have some of those inputs in common with you, you might come to the conclusion that it's not that surprising after all.
Also, we tend to forget all the times when the coincidence didn't occur, because that is simply not memorable, so over time we are left with a biased series of memorable coincidences.
It is not evident (to me, anyway) that theories about evolution and creation are really even scientific theories, because they're not directly testable, though they are based on scientific understanding of underlying physical processes which are separately testable.
I presume that you're saying that evolution is not testable, on the basis that it's all already happened, and we're just looking at the evidence left behind.
That would be fair enough, if it were not for the fact that we are still witnessing evolution in action, and so it can be observed.
There is the example of moths during the Industrial revolution, where the soot produced by the factories killed the lichen on the trees, and stained the bark. The moths had been white prior to this happening, which was good camoflage when sitting on the light coloured bark --- once the trees changed colour, the lighter moths were easy targets for the birds, and so the moths very quickly evolved to be black --- same species, different colour.
The factories have since cleaned up their act (or shut down) and the trees have lichen again, and the moths have evolved back to being white.