My main argument against that is that if they want to make GNOME easy-to-use, their current users aren't the ones they should be listening to. Sad though it sounds.
There's an academically interesting (I'm assured:)) Java system similar to this called Bdbj. I'm not sure if it's useful in a real context, but I assume it is to some degree.
When I said its "architecture", I was referring to the fact that it's a microkernel system. I was just trying to explain this in a quick and easy way to something.
I'm sure some people would be happier if I posted a 3000-word essay on kernel types in general, but I reckon more people are happier with the simpler idea, in context. The reason I mentioned its "architecture", though, was simply because I didn't know enough about it in practice to say further. But I do feel it's fair to describe a microkernel architecture as "more modern" than a monolithic kernel.
Nope. I know the difference between microkernels and their monolithic counterparts, and I feel that "more modern" is a fair comparison to make, in architectural terms. There are better examples of the microkernel model now, though, obviously.
Oh, I'm well aware of that particular debate. Both sides do have good points, but I'm not sure that I think Torvalds's point is as strong as Tannenbaum's -- although to be fair, Tannenbaum can be a bit of an arse throughout.
It's an interesting debate, in any case. I'm a big fan of Microkernels myself, but it is certainly a debatable subject.
Note that I'm not the person you were responding to, but Mach's architecture is certainly more modern than that of Linux -- although I've no idea how it stacks up in real life terms.
Cheers, I saw the other post and I think you were saying something similar to what I was trying to convey. XML was my example of a config language because if worse comes to worse, it's user-editable and it's more expressive than flat text files (theoretically - obviously strictly it is a text file), to the degree that there's little credence in the argument that you need something that "can't be done" in XML, so there's no excuse for people not using a compatible format. But yes, ideally one would want an editing front-end on top of any data representation.
I don't agree with the last phrase in that, but you make a good point. Creeping featurism is what's killing a lot of systems, particularly in terms of usability.
I just feel relative terms to "rating" seem redundant when rating has not yet occurred. I've felt cheated when modded "underrated" with a score of 1 before, and it seems horrid.
I'll change my sig, and hope for mercy in any case. Cheers for input.
Yeah, I should've actually mentioned that, the report reads like (and is) and ad for their software. It's been reported enough places to be given at least some credence, although whether that's due to the sensationalist figures I wouldn't like to comment.
Sometimes things just get big, and there's not a lot we can do about it. "Keep It Simple" is a good phrase to develop by, but in the real world it ain't always possible.
I think the report most frequently cited is CacheLogic's one (news article). They give a figure of 53% of P2P traffic, and their graphs show BitTorrent overwhelming many other forms of traffic, compared to the wider internet. I can't find the actual "one third" figure, but I did see 35% in a couple of places while looking for this.
Interesting! I'd never heard of YAML before. Looks worth a look-in, anyway. Is it as expressive as XML (out of the box)? Looking into it when I have time in any case, cheers:)
It just seems a pity that a lot of people knock it, simply because it breaks old ways -- sometimes, that just has to happen. Thanks for listening, anyway:)
I just feel that such a system would remedy these problems, in the long-term. Once you have an XML-based system, XML views can be used to abstract the configuration of a specific package, and various other features can help you out.
XML can be constrained in many ways to ensure and preserve its validity -- although it has its caveats, its proper use can prevent more human errors than it could ever introduce (unless you have someone with angle bracket dyslexia, I guess:)).
If the structure is semantic enough, it shouldn't matter -- you don't need to learn a million syntaxes for a million systems, you just need learn one. The XML store file would obviously have to be editable, but that shouldn't be a large obstacle. If you've learned the structure of the data, the close links of this to the XML document(s) should aid its readability no end, particularly if you've got a passing familiarity with XML to begin with. I'm not convinced that a specific, but constant, syntax tailored to each program is an improvement over a single, generalised, logically-structured format. But saying that generated XML is not portable seems false to me, in any case. It's as portable as any XML. And if it's logical, the links between the front-end and the data can be inferred by the user.
Interesting point, though. Cheers. Backwards compatibility, in particular, is a sticking point. But I do believe that we should be moving forward to a generalised system, while (for now) retaining the knowledge of older systems, if only to (as in your example) maintain systems where such a system has not been deployed.
Are you actually claiming that XML isn't portable? There are myriad tools for administering XML files as-is, an abstract tool just makes it a degree easier.
Text files need to be given a structure, by definition XML is more expressive than a plain label:value structure, and it's standard. If you want a standard for all systems, you should go for the most expressive data representation available, to make sure you cover all bases.
What's your justification for hating it, beyond pure "hacker joy" of editing configuration files? Sad though it might seem, topics of data representation and manipulation really interest me, I'd be interested to hear a full justification for such an assertion.
Not trying to troll or annoy, I just like finding these things out! Also, if the other poster made your point, you might want to read my answer to him to get a fuller understanding of my point of view on the whole issue. If you haven't already adamantly made up your mind, that is!
Slashdot Mirror
Are the users that are talking to them right now really the ones they want to adapt to, though?
My main argument against that is that if they want to make GNOME easy-to-use, their current users aren't the ones they should be listening to. Sad though it sounds.
There's an academically interesting (I'm assured :)) Java system similar to this called Bdbj. I'm not sure if it's useful in a real context, but I assume it is to some degree.
The newest vulnerability on that list is from 2003.
When I said its "architecture", I was referring to the fact that it's a microkernel system. I was just trying to explain this in a quick and easy way to something.
I'm sure some people would be happier if I posted a 3000-word essay on kernel types in general, but I reckon more people are happier with the simpler idea, in context. The reason I mentioned its "architecture", though, was simply because I didn't know enough about it in practice to say further. But I do feel it's fair to describe a microkernel architecture as "more modern" than a monolithic kernel.
Nope. I know the difference between microkernels and their monolithic counterparts, and I feel that "more modern" is a fair comparison to make, in architectural terms. There are better examples of the microkernel model now, though, obviously.
Oh, I'm well aware of that particular debate. Both sides do have good points, but I'm not sure that I think Torvalds's point is as strong as Tannenbaum's -- although to be fair, Tannenbaum can be a bit of an arse throughout.
It's an interesting debate, in any case. I'm a big fan of Microkernels myself, but it is certainly a debatable subject.
Note that I'm not the person you were responding to, but Mach's architecture is certainly more modern than that of Linux -- although I've no idea how it stacks up in real life terms.
Heh, I'll bear it in mind if I ever get the chance to develop such a system! Cheers for input though, this sort of stuff really does interest me.
Cheers, I saw the other post and I think you were saying something similar to what I was trying to convey. XML was my example of a config language because if worse comes to worse, it's user-editable and it's more expressive than flat text files (theoretically - obviously strictly it is a text file), to the degree that there's little credence in the argument that you need something that "can't be done" in XML, so there's no excuse for people not using a compatible format. But yes, ideally one would want an editing front-end on top of any data representation.
I don't agree with the last phrase in that, but you make a good point. Creeping featurism is what's killing a lot of systems, particularly in terms of usability.
I just feel relative terms to "rating" seem redundant when rating has not yet occurred. I've felt cheated when modded "underrated" with a score of 1 before, and it seems horrid.
I'll change my sig, and hope for mercy in any case. Cheers for input.
Yeah, I should've actually mentioned that, the report reads like (and is) and ad for their software. It's been reported enough places to be given at least some credence, although whether that's due to the sensationalist figures I wouldn't like to comment.
Like Linux?
Sometimes things just get big, and there's not a lot we can do about it. "Keep It Simple" is a good phrase to develop by, but in the real world it ain't always possible.
I think the report most frequently cited is CacheLogic's one (news article). They give a figure of 53% of P2P traffic, and their graphs show BitTorrent overwhelming many other forms of traffic, compared to the wider internet. I can't find the actual "one third" figure, but I did see 35% in a couple of places while looking for this.
Interesting! I'd never heard of YAML before. Looks worth a look-in, anyway. Is it as expressive as XML (out of the box)? Looking into it when I have time in any case, cheers :)
Thanks a lot :)
:)
It just seems a pity that a lot of people knock it, simply because it breaks old ways -- sometimes, that just has to happen. Thanks for listening, anyway
More "not ported" than "not portable", but your point stands, yes. I think I alluded to that in the last post though.
I just feel that such a system would remedy these problems, in the long-term. Once you have an XML-based system, XML views can be used to abstract the configuration of a specific package, and various other features can help you out.
:)).
XML can be constrained in many ways to ensure and preserve its validity -- although it has its caveats, its proper use can prevent more human errors than it could ever introduce (unless you have someone with angle bracket dyslexia, I guess
If the structure is semantic enough, it shouldn't matter -- you don't need to learn a million syntaxes for a million systems, you just need learn one. The XML store file would obviously have to be editable, but that shouldn't be a large obstacle. If you've learned the structure of the data, the close links of this to the XML document(s) should aid its readability no end, particularly if you've got a passing familiarity with XML to begin with. I'm not convinced that a specific, but constant, syntax tailored to each program is an improvement over a single, generalised, logically-structured format. But saying that generated XML is not portable seems false to me, in any case. It's as portable as any XML. And if it's logical, the links between the front-end and the data can be inferred by the user.
Interesting point, though. Cheers. Backwards compatibility, in particular, is a sticking point. But I do believe that we should be moving forward to a generalised system, while (for now) retaining the knowledge of older systems, if only to (as in your example) maintain systems where such a system has not been deployed.
Are you actually claiming that XML isn't portable? There are myriad tools for administering XML files as-is, an abstract tool just makes it a degree easier.
Yeah, there's been a little bit of discussion about this in the sibling posts to yours, and their children. It looks like a nice thing, really.
Text files need to be given a structure, by definition XML is more expressive than a plain label:value structure, and it's standard. If you want a standard for all systems, you should go for the most expressive data representation available, to make sure you cover all bases.
What's your justification for hating it, beyond pure "hacker joy" of editing configuration files? Sad though it might seem, topics of data representation and manipulation really interest me, I'd be interested to hear a full justification for such an assertion.
Not trying to troll or annoy, I just like finding these things out! Also, if the other poster made your point, you might want to read my answer to him to get a fuller understanding of my point of view on the whole issue. If you haven't already adamantly made up your mind, that is!