Hi, I think we're generally in agreement - we both want access to better tools, and I don't think there's even that much difference between us about how this can be achieved.
I think it's important that people who spend a lot of time doing something that improves all of our lives gets some kind of reward for that -whether it's free pizza, a pat on the back or enough money to buy themselves a brand new powerboat. People that wish to be marters to the FSF cause are most welcome to, but I think they deserve more than that. If a company can afford to support their effort - by paying for a bit of the OS they're using, for example, then that is surely a good thing. It seems that your main point is the route taken by the money - how does it get from me/my employer to John Doe the FSF genius?
Personally I don't care that much as long as:
It doesn't stifle the production of good tools to make you and me happier.
It gets to the developers, who decide whether to fund their pizza lust, or their future software developing habit. My guess is that most FSF contributers will concentrate on the latter, but it's up to them. The worst thing that can happen is that someone somewhere does bugger all work and ends with a packet by exploiting other people's efforts. Most Distributions recognise this, I think, because they realise how important it is not to alienate the community that they rely on.
I realise this is straying away from the general Debian issue, but I'm interested in your views about whether it would be possible to build a profit-sharing cooperative to generate code in a way that rewards developers with a little more than a pat on the back,fame, and a 'thanks!'.
With regards to the statistics, I agree with you that the original poll was extremely loaded. I also agree with your analogy... At least the form wasn't laid out so that it was really easy to click on the wrong checkbox;-).
However, I think my point was the opposite of how you interpreted it - the fact is that 32% said they were willing to pay for something if it was good enough (given the nature of the questions, as you pointed out). This is quite an incentive for a commercial distro to work out how to make 'something good enough'. I think a more serious floor in my argument is that it assumes people always put their money where their mouth is.
Another point you hinted at was the way you perceive the attitude of non-Debian-aware users. The view that an MS solution is the only affordable one for the desktop is starting to be challenged in communities other than the Linux/geek chatrooms, and if things keep moving the way they are, the openeness to new ideas can only increase. Fingers crossed - lets just hope the associated cash finds a good home...
If I polled the CD buying community and asked how many would be prepared to buy the next Britney Spears Album and 32% said 'yes', I'd think she was doing pretty well. I think I might even buy stocks in her record company.
If 32% of people, used to getting something for free, say they are prepared to spend cash on a commercial variant, then doesn't that sounds like good news for the companies involved?
Whether that's good for Open Source, Unix etc.. is a different question, which, I think, might have been discussed sometime before on Slashdot.:-)
As someone who spends an unhealthy amount of time infront of my computer, anybody who can help make that experience more enjoyable for me is going to be my friend - I might even ask my company to pay them for it....
If it was a hot air car, I could run it off my boss - finally, turning him into a useful resource. More generally, We could power entire public transportation systems simply by holding regular meetings to discuss great new e-commerce ideas with venture capitalists.
I was thinking exactly the same thing. I made a version of the Nipkow disk (wooden thing with holes in acting like the lenses in Logi Bird's) when I was at school. It worked if you had a good sense of imagination. It also got me beaten up for being a geek - funny how times change isn't it... Anyway, if you start to think seriously about building one of these things, keep us posted...
The other alternative is to do it all on the server using XML to specify your content and XSL to do the transformation. It exists, it works, and it doesn't rely on hardware manufacturers implementing new standards properly.
I can't see that, in the near future, anybody will want the same style sheet for a desk top browser and a mobile phone - and I would expect the content displayed on a phone to be very much more restricted. Rather than send a lot of data that's going to get ignored, there's a good argument for only serving the data that's required. A producer->processor->formatter architecture allows for this. A servlet generates dynamic content, which is then passed to a processor that mungs it about a bit - inserts dates, side bars, menus and the like, before a formatter turns it into, for example, HTML, WAP, VoxML, or a binary format like PDF. Apache Cocoon is one architecture that does this, now.
I don't understand these video phone things - fdon't they just send a close up of your cheek bone to the person on the other end. And, anyway they havent got an eye on the side of their cheek so they could'nt see the picture anyway. Or am I missing something?
The one that looks like a make-up compact is fun though...Very Lady Penelope.
Agreed - but it's not just a case of tweaking an existing algorithm - something pretty major needs to happen in terms of intellectual leaps. This suggests that design time is going to be far greater than runtime.
Great. Nobody knows how a protein folds, what's important, what other bits of cellular machinary are involved and so on. All this means is they can calculate a answer which is probably wrong a hell of a lot faster than they used to be able to. Garbage in, garbage out.
So how fast can you get data on and off this thing? How big is the cache per processor? What is the bandwidth and latency? The bottleneck in most parallel tasks is inter-processor communication. Beowulf (drat - I vowed I'd never say it) is cheap because compromises have been made w.r.t. bandwidth and latency between processors - that's great, but it makes it a poor choice for some tasks. Presumeably, this thing will have other compromises, and, IMHO, until we know what they are, it's not really possible to make any kind of reasoned judgement about what it can be used for...
Looks interesting - a good, standard-ish Beowulf distribution with knobs on would be most appealing. But, my applications require large amounts of database-type searches against huge files that change rarely. I want to be able to store that data on the local drives of my cluster rather than shunt it around the machine. I don't think that is very easy to do at the moment... It would be nice if there were accepted ways to do that. Does anyone have any advice BTW?
I totally agree - that's not my point. Once you've paid for your wire and your electricity bill, the cost of sending data is free. Sure, charge per minute, or byte to stop people forcing you to buy a fatter cable, but I'm under the impression that that's going to be less and less of an issue in the future.
Sorry I wasn't trolling, although I realise I could have been a little more explicit in my original post...
>Am I missing something, or am I being ripped off?
>I seem to remember that things cost more to send internationally.
I don't know how much international post there was from the UK when they did that calculation... I think I could have expressed myself if I'd said 'flat rate throughout the UK'.... I guess charging more to read Slashdot down an expensive transatlantic link would be analagous to charging more for international mail.
As far as the comment about wires is concerned, I agree that different types of network infrastructure cost different amounts and so on, but that isn't my point.
My point is that it costs nothing to actually send a byte down a wire, the cost is in putting the infrastructure in place and keeping it going (or am I missing something?). So calculating a 'cost per byte' for a metering system is a bit odd, because it's totally dependent on the number of bytes being sent.
Errrmmmm.... I don't understand what you mean. Agreed, capitalism doesn't force people to sell things at a particular price, instead, it lets people charge what they can get. Alongside that are laws that try to maintain enough competition to allow competitors to try to make a similar thing, but cheaper.
If you consider people as a valuable resource (more valuable than things, IMHO) then the whole 'cost' thing gets much harder to work out. But, IMHO, a free market is much better at getting the price of things right than any other system.
The cost of letter-delivery used to be calculated according to the distance the letter was going to go. One of the first things that Babagge did (way before he designed the difference engine) was point out that it cost more to do the calculation than it did to deliver the letter. Hence the flat rate stamp was born.
I would have thought that the same sort of thing is probably true for Internet access - especially since sending data down a wire is just as expensive as not sending data down the same wire, once the wire has been laid.
When you follow their download link you end up getting their hideous installer that connects to their site and pulls things down for you (if you're lucky) and then tries to put them in the wrong place).
You know the one:
'Thank you for downloading Netscape 6, you are held in a queue and will be attended to shortly....your download is important to us... please hold...if you have a numeric kepad press * now....etc...'
It still doesn't play The William Tell Overture in four-part square-wave harmony though, so it's not all bad...
Fair enough we can set up a farm to do something, that's the easiest bit, but how many problems can be solved with coarse grained parallelism, such as seti@home? I'm not convinced that the real problem is lack of brawn, sometimes it's more about brains.
For instance, the human genome project, which is often sited as a potential application of the GRID, suffers more from lack of knowledge, not not number crunching. We only need to to assemble the bits of the genome once, and when its done, it's a few gig of data lying around on our hard drive. The problem is more that we don't fully understand how the body gets from a DNA sequence to a properly folded protein, let alone, what the resulting lump of atoms actually does. More crunch won't help - deeper insights will.
Secondly, what generally happens when you do a lot of processing of discrete bits of data, is that you end up with another database holding the results. Then, you have to sift through that database.... Crunch is all very well, but unless you know what to do with the results, it's a waste of time.
Slashdot Mirror
Title says it all really: how easy would it be?
I think it's important that people who spend a lot of time doing something that improves all of our lives gets some kind of reward for that -whether it's free pizza, a pat on the back or enough money to buy themselves a brand new powerboat. People that wish to be marters to the FSF cause are most welcome to, but I think they deserve more than that. If a company can afford to support their effort - by paying for a bit of the OS they're using, for example, then that is surely a good thing. It seems that your main point is the route taken by the money - how does it get from me/my employer to John Doe the FSF genius?
Personally I don't care that much as long as:
I realise this is straying away from the general Debian issue, but I'm interested in your views about whether it would be possible to build a profit-sharing cooperative to generate code in a way that rewards developers with a little more than a pat on the back
With regards to the statistics, I agree with you that the original poll was extremely loaded. I also agree with your analogy... At least the form wasn't laid out so that it was really easy to click on the wrong checkbox
However, I think my point was the opposite of how you interpreted it - the fact is that 32% said they were willing to pay for something if it was good enough (given the nature of the questions, as you pointed out). This is quite an incentive for a commercial distro to work out how to make 'something good enough'. I think a more serious floor in my argument is that it assumes people always put their money where their mouth is.
Another point you hinted at was the way you perceive the attitude of non-Debian-aware users. The view that an MS solution is the only affordable one for the desktop is starting to be challenged in communities other than the Linux/geek chatrooms, and if things keep moving the way they are, the openeness to new ideas can only increase. Fingers crossed - lets just hope the associated cash finds a good home...
If 32% of people, used to getting something for free, say they are prepared to spend cash on a commercial variant, then doesn't that sounds like good news for the companies involved?
Whether that's good for Open Source, Unix etc.. is a different question, which, I think, might have been discussed sometime before on Slashdot. :-)
As someone who spends an unhealthy amount of time infront of my computer, anybody who can help make that experience more enjoyable for me is going to be my friend - I might even ask my company to pay them for it....
If it was a hot air car, I could run it off my boss - finally, turning him into a useful resource. More generally, We could power entire public transportation systems simply by holding regular meetings to discuss great new e-commerce ideas with venture capitalists.
I was thinking exactly the same thing. I made a version of the Nipkow disk (wooden thing with holes in acting like the lenses in Logi Bird's) when I was at school. It worked if you had a good sense of imagination. It also got me beaten up for being a geek - funny how times change isn't it... Anyway, if you start to think seriously about building one of these things, keep us posted...
I can't see that, in the near future, anybody will want the same style sheet for a desk top browser and a mobile phone - and I would expect the content displayed on a phone to be very much more restricted. Rather than send a lot of data that's going to get ignored, there's a good argument for only serving the data that's required. A producer->processor->formatter architecture allows for this. A servlet generates dynamic content, which is then passed to a processor that mungs it about a bit - inserts dates, side bars, menus and the like, before a formatter turns it into, for example, HTML, WAP, VoxML, or a binary format like PDF. Apache Cocoon is one architecture that does this, now.
The one that looks like a make-up compact is fun though...Very Lady Penelope.
Agreed - but it's not just a case of tweaking an existing algorithm - something pretty major needs to happen in terms of intellectual leaps. This suggests that design time is going to be far greater than runtime.
Great. Nobody knows how a protein folds, what's important, what other bits of cellular machinary are involved and so on. All this means is they can calculate a answer which is probably wrong a hell of a lot faster than they used to be able to. Garbage in, garbage out.
So how fast can you get data on and off this thing? How big is the cache per processor? What is the bandwidth and latency? The bottleneck in most parallel tasks is inter-processor communication. Beowulf (drat - I vowed I'd never say it) is cheap because compromises have been made w.r.t. bandwidth and latency between processors - that's great, but it makes it a poor choice for some tasks. Presumeably, this thing will have other compromises, and, IMHO, until we know what they are, it's not really possible to make any kind of reasoned judgement about what it can be used for...
Looks interesting - a good, standard-ish Beowulf distribution with knobs on would be most appealing. But, my applications require large amounts of database-type searches against huge files that change rarely. I want to be able to store that data on the local drives of my cluster rather than shunt it around the machine. I don't think that is very easy to do at the moment... It would be nice if there were accepted ways to do that. Does anyone have any advice BTW?
When it's full, everyone else has to suffer.
I totally agree - that's not my point. Once you've paid for your wire and your electricity bill, the cost of sending data is free. Sure, charge per minute, or byte to stop people forcing you to buy a fatter cable, but I'm under the impression that that's going to be less and less of an issue in the future.
>Am I missing something, or am I being ripped off?
>I seem to remember that things cost more to send internationally.
I don't know how much international post there was from the UK when they did that calculation... I think I could have expressed myself if I'd said 'flat rate throughout the UK'.
As far as the comment about wires is concerned, I agree that different types of network infrastructure cost different amounts and so on, but that isn't my point.
My point is that it costs nothing to actually send a byte down a wire, the cost is in putting the infrastructure in place and keeping it going (or am I missing something?). So calculating a 'cost per byte' for a metering system is a bit odd, because it's totally dependent on the number of bytes being sent.
If you consider people as a valuable resource (more valuable than things, IMHO) then the whole 'cost' thing gets much harder to work out. But, IMHO, a free market is much better at getting the price of things right than any other system.
Yes -exactly...
:-)
The cost of letter-delivery used to be calculated according to the distance the letter was going to go. One of the first things that Babagge did (way before he designed the difference engine) was point out that it cost more to do the calculation than it did to deliver the letter. Hence the flat rate stamp was born.
I would have thought that the same sort of thing is probably true for Internet access - especially since sending data down a wire is just as expensive as not sending data down the same wire, once the wire has been laid.
When you follow their download link you end up getting their hideous installer that connects to their site and pulls things down for you (if you're lucky) and then tries to put them in the wrong place).
You know the one:
'Thank you for downloading Netscape 6, you are held in a queue and will be attended to shortly....your download is important to us... please hold...if you have a numeric kepad press * now....etc...'
It still doesn't play The William Tell Overture in four-part square-wave harmony though, so it's not all bad...
For instance, the human genome project, which is often sited as a potential application of the GRID, suffers more from lack of knowledge, not not number crunching. We only need to to assemble the bits of the genome once, and when its done, it's a few gig of data lying around on our hard drive. The problem is more that we don't fully understand how the body gets from a DNA sequence to a properly folded protein, let alone, what the resulting lump of atoms actually does. More crunch won't help - deeper insights will.
Secondly, what generally happens when you do a lot of processing of discrete bits of data, is that you end up with another database holding the results. Then, you have to sift through that database.... Crunch is all very well, but unless you know what to do with the results, it's a waste of time.
..adds a whole new meaning to the phrase e-commerce...
Does the microsoft bug-tracking database count as interactive fiction?
Is this what happens when you put PERLs before Swine?
The olympic sailing, laser class resulted in the following commentary last night:
Ainslie is holding out in first place, but he still has Scheidt sticking to his stern.
Do these filters also look for phonetic spelling?
Shame they couldn't wait for KDE2.0, or at least put the newest beta version on it.
It talks about:
the brave new unwired world
Shouldn't that be
the brave new underwired world?
Does it contain undo information?