...what they did with Emagic. Emagic Logic, lovely music sequencing program, worked on Windows and Macs. Apple buy them up, first thing they do, "sorry guys, its going Mac only".
Now, if they do that with Adobe software, what do you think will happen?
Yes, academia is about sharing with out fellow human beings, we do this through the process of peer review. However, wikipedia is not peer review. The relevant definition of 'peer' is: "a person who is equal to another in abilities, qualifications, age, background, and social status." [www.dictionary.com]
Whereas wikipedia is: "the free encyclopedia that anyone can edit"
QED. I don't think you have to send this post off to two referees to figure that out.
I love the little tagline on facebook: "join a network to see people who live, study, or work around you". There's me thinking you could just walk outside your front door, or take a stroll around your offices or college to do that...
I don't recommend LaTeX because it's open source, I recommend it because it is the right tool for the job we do - math, math, and more math. If I recommend someone hit a nail with a hammer and they do it with a screwdriver and cut their finger, why should I have to mop the blood up?
The problem is that a short-sighted decision will come back to bite you. In this case, the user is now causing inter-operability problems with most of our department who are LaTeX users because of LaTeX's superior quality. As their bibliography builds they'll need EndNote licenses. If they want to work with other people they're going to need some sort of collaborative authoring solution that deals with Word and all its change log stuff. With LaTeX you get BiBTeX for free and can do your collaboration with SVN.
I've argued this with people in the past, and they just go, 'yeah, but I just want to click "X" and then "squared"'. In the future I shall say "fine, you do that, you're Doing It Wrong, and if you want help I will tell you how to Do It Right, and that will be LaTeX".
Because the idiots are winning, to quote Dan Ashcroft. We geeks may have spent the last 15 years telling people how Windows and proprietary software sucks so badly, yet still they buy it.
This week someone in my department had a problem reading and editing equations in a Word DOC file in OpenOffice. And then he tried genuine Microsoft Word and still had problems. Turned out the equations were done using 'Mathtype', some extra add-on for Word. "Doing it wrong!", I cried, "why is anyone writing papers chock-full of equations in Word anyway? Use LaTeX like everyone else in the department. It's free, it produces nicer papers, it is just beautiful."
So they asked for my advice, and it was 'use LaTeX'. What they did: buy more MathType licenses.
I'm thinking of starting a policy of not helping anyone who asks for tech support in our department and then does something else. I'm speccing up two labs with about 50 seats this week. I'm recommending 50 thin clients and 5 fat servers just to make it manageable. If the suit overrules this and says 'no, no, no, just put 50 desktop boxes in' then I'm not supporting it - to the extent of quitting the project. The idiots are not winning this one (although the servers will of course be running Win TS 2003 as well as Linux).
I'm not sure how we can stop the idiots from winning - MS will up their muscle in the fight, since they fatten up the idiots in order to feed off them. Or is it more like giving them free heroin and getting them hooked and then charging for the next fix. It's idiocy whatever way you cut it.
End of Sunday night before Monday Work Rant.
[Dan Ashcroft was a character in Nathan Barley...]
Geek to suit: "Hey look, Microsoft are now *really* getting scared by open source stuff! They want to throw *real money* at it!"
Also, people might now start investing in open source projects in the hope of getting a slice of that MS cash a few years down the line. This looks like a Good Thing.
Instead of killing them be a bit more creative. Get your phone out and go 'Michael? Hi! Mike! How's Mrs Hayden?, Do me a favour? Pull up the passenger manifest for AA96 and make the guy in 5E disappear will you? Great! Golf tomorrow? Sure! Bye!"
It'll be one of those low-numbered rows since only people in First and Business class will be able to afford to yak away for a whole long flight. Since you'll be in that class too, don't kill them yourself, get one of your staff to do it.
If you really are in Sardine class and have someone talk all the time he's either very rich, in which case make him change his will and then kill him, or he's not really talking to anyone and just wants you to think he's rich and popular. I don't have a cellphone, I think they don't really work, people just clamp them to their heads and pretend they have friends.
If you're interested, I think I've managed to get the relevant galaxy data out of SDSS. For one of his cutoffs he's not explicit as to whether its a lower or upper bound, so I've just got all the data and I'll figure out which way it swings later. I'm guessing the cutoff is obvious if you're an astronomer. I'm just a failed astrophysicist:)
No its not - his test statistic, , is the mean of the 48 points of his ellipticity histogram (see fig 4). If there's a significant effect then the ellipticity histogram will show a peak _somewhere_, and hence push the mean up.
I can probably think of better test statistics, since there's so much correlation in the 48 points...
His "random computer generated results" test is what we statisticians call a 'Monte-Carlo' test. Its perfectly valid, given the assumptions he is working under.
Suppose you throw 10 possibly biased dice and score 50 in total (where the average score would be 30).
You then get 10 definitely fair dice and throw them 100 times, counting the total each time. If these trials only score 50 or more once, then the chances of your possibly biased dice being fair are 1 in 99. That's pretty much what he's done.
With dice its possible to compute the probability exactly without doing the trials, since the behaviour of uniform probabilities (ie even chance of scoring 1 to 6) are well known and easy to compute. But if you have a situation of elliptical galaxies and their apparent projection on a sphere viewed from the earth then I suspect the computations may be harder...
Its not just game devs. Animators try real hard to make their animations have that 'film' look. I think Nick Park did motion blur in some of his early stuff - possibly the early Wallace and Grommit shorts. Since then I think he's taken every trick in the cinematographer's book, even borrowing from Spielberg with the crash-zoom shot where you zoom out whilst tracking the camera up to the subject, keeping the same size and making the background do crazy things. I bet there's some lifts from Citizen Kane in his work too. Marvellous.
Its when they ask for metrics on the number of metrics you can supply them with that you should really start worrying.
Stuff like MTBF (mean time between failures) or %uptime are possible things to keep the bean counters happy. You can probably get the data straight from monitoring systems like 'mon' or nagios. On a longer timescale, do a regular report (annual, six monthly, whatever) on number of new desktop systems installed, number of OS upgrades done, that kind of thing. Bean counters like reports.
If you run a web server, google analytics can provide you with a huge amount of numbers to pass on to management...
I have no idea what they do with all this data anyway.
Its "open" in the sense of being the opposite of "closed". The source code is there, you can read it.
"Free" means you are free to do stuff to it (exactly what stuff you are free to do is license-dependent).
Your examples are pretty bogus: getting software to work with more than one compiler (or compiler system) is not trivial, even with so-called ANSI compliance. Try grepping source files for '#ifdef GCC' or '#ifdef CYGWIN' to see. And your apache example? Sure most HTTP servers will spit out static pages, but they might not support CGI, apache's virtual hosts, security, logging, mod-perl, mod-python, mod-ruby, mod-fortran (does that exist?) and so on. Any moderately advanced user of Apache is locked-in as tight as a gnat's chuff.
There are arguments for a truly open and free software stack from applications to the BIOS in certain fields, (for example in science to enable reproducible research), but that doesn't mean that all components in a stack with non-free and closed source parts are tainted. They are free. They are open. Whatever their requirements.
ABBA (whats the Unicode for a backwards 'B'?) were just first in alphabetical order, because popular beat combo Aaron's Aardvarks hadn't formed yet. And still haven't.
Slashdot Mirror
...what they did with Emagic. Emagic Logic, lovely music sequencing program, worked on Windows and Macs. Apple buy them up, first thing they do, "sorry guys, its going Mac only".
Now, if they do that with Adobe software, what do you think will happen?
Yes, academia is about sharing with out fellow human beings, we do this through the process of peer review. However, wikipedia is not peer review. The relevant definition of 'peer' is: "a person who is equal to another in abilities, qualifications, age, background, and social status." [www.dictionary.com]
Whereas wikipedia is: "the free encyclopedia that anyone can edit"
QED. I don't think you have to send this post off to two referees to figure that out.
I love the little tagline on facebook: "join a network to see people who live, study, or work around you". There's me thinking you could just walk outside your front door, or take a stroll around your offices or college to do that...
I don't recommend LaTeX because it's open source, I recommend it because it is the right tool for the job we do - math, math, and more math. If I recommend someone hit a nail with a hammer and they do it with a screwdriver and cut their finger, why should I have to mop the blood up?
The problem is that a short-sighted decision will come back to bite you. In this case, the user is now causing inter-operability problems with most of our department who are LaTeX users because of LaTeX's superior quality. As their bibliography builds they'll need EndNote licenses. If they want to work with other people they're going to need some sort of collaborative authoring solution that deals with Word and all its change log stuff. With LaTeX you get BiBTeX for free and can do your collaboration with SVN.
I've argued this with people in the past, and they just go, 'yeah, but I just want to click "X" and then "squared"'. In the future I shall say "fine, you do that, you're Doing It Wrong, and if you want help I will tell you how to Do It Right, and that will be LaTeX".
Because the idiots are winning, to quote Dan Ashcroft. We geeks may have spent the last 15 years telling people how Windows and proprietary software sucks so badly, yet still they buy it.
This week someone in my department had a problem reading and editing equations in a Word DOC file in OpenOffice. And then he tried genuine Microsoft Word and still had problems. Turned out the equations were done using 'Mathtype', some extra add-on for Word. "Doing it wrong!", I cried, "why is anyone writing papers chock-full of equations in Word anyway? Use LaTeX like everyone else in the department. It's free, it produces nicer papers, it is just beautiful."
So they asked for my advice, and it was 'use LaTeX'. What they did: buy more MathType licenses.
I'm thinking of starting a policy of not helping anyone who asks for tech support in our department and then does something else. I'm speccing up two labs with about 50 seats this week. I'm recommending 50 thin clients and 5 fat servers just to make it manageable. If the suit overrules this and says 'no, no, no, just put 50 desktop boxes in' then I'm not supporting it - to the extent of quitting the project. The idiots are not winning this one (although the servers will of course be running Win TS 2003 as well as Linux).
I'm not sure how we can stop the idiots from winning - MS will up their muscle in the fight, since they fatten up the idiots in order to feed off them. Or is it more like giving them free heroin and getting them hooked and then charging for the next fix. It's idiocy whatever way you cut it.
End of Sunday night before Monday Work Rant.
[Dan Ashcroft was a character in Nathan Barley...]
Was it Turing or Bill Gates who said that an infinitely long tape ought to be enough for anyone?
"Constantia can replace ... Helvetica".
Ah, I think not. Nobody will ever make a film about Constantia - http://www.helveticafilm.com/
Maybe one will be made about Comic Sans, but it will be a horror story.
Geek to suit: "Hey look, Microsoft are now *really* getting scared by open source stuff! They want to throw *real money* at it!"
Also, people might now start investing in open source projects in the hope of getting a slice of that MS cash a few years down the line. This looks like a Good Thing.
Instead of killing them be a bit more creative. Get your phone out and go 'Michael? Hi! Mike! How's Mrs Hayden?, Do me a favour? Pull up the passenger manifest for AA96 and make the guy in 5E disappear will you? Great! Golf tomorrow? Sure! Bye!"
It'll be one of those low-numbered rows since only people in First and Business class will be able to afford to yak away for a whole long flight. Since you'll be in that class too, don't kill them yourself, get one of your staff to do it.
If you really are in Sardine class and have someone talk all the time he's either very rich, in which case make him change his will and then kill him, or he's not really talking to anyone and just wants you to think he's rich and popular. I don't have a cellphone, I think they don't really work, people just clamp them to their heads and pretend they have friends.
Or even the second-best upgrade they ever released. Marketing eh, dontcha love it?
Next slashdot headline:
Ownz0ring teh Wireless Cam, Its LUser and teh Netw0rk LOL! ROLFMAYO!
I read the headline and thought it meant it would save millions of people... Better imaging of flood areas, hurricane tracking, or something.
Imagine my disappointment when I discovered it meant dollars...
I'm guessing you've never seen an academic's desk before...
whoops yeah, well spotted :)
If you're interested, I think I've managed to get the relevant galaxy data out of SDSS. For one of his cutoffs he's not explicit as to whether its a lower or upper bound, so I've just got all the data and I'll figure out which way it swings later. I'm guessing the cutoff is obvious if you're an astronomer. I'm just a failed astrophysicist :)
No its not - his test statistic, , is the mean of the 48 points of his ellipticity histogram (see fig 4). If there's a significant effect then the ellipticity histogram will show a peak _somewhere_, and hence push the mean up.
I can probably think of better test statistics, since there's so much correlation in the 48 points...
His "random computer generated results" test is what we statisticians call a 'Monte-Carlo' test. Its perfectly valid, given the assumptions he is working under.
Suppose you throw 10 possibly biased dice and score 50 in total (where the average score would be 30).
You then get 10 definitely fair dice and throw them 100 times, counting the total each time. If these trials only score 50 or more once, then the chances of your possibly biased dice being fair are 1 in 99. That's pretty much what he's done.
With dice its possible to compute the probability exactly without doing the trials, since the behaviour of uniform probabilities (ie even chance of scoring 1 to 6) are well known and easy to compute. But if you have a situation of elliptical galaxies and their apparent projection on a sphere viewed from the earth then I suspect the computations may be harder...
I'm trying to think of a valid reason why A could be right. Maybe air is denser in rooms with CD players...
Its not just game devs. Animators try real hard to make their animations have that 'film' look. I think Nick Park did motion blur in some of his early stuff - possibly the early Wallace and Grommit shorts. Since then I think he's taken every trick in the cinematographer's book, even borrowing from Spielberg with the crash-zoom shot where you zoom out whilst tracking the camera up to the subject, keeping the same size and making the background do crazy things. I bet there's some lifts from Citizen Kane in his work too. Marvellous.
True, but if your list of usernames leaks out it saves remote attackers having to try non-existent usernames in a dictionary attack...
Corollary: dont use passwords vulnerable to dictionary attacks...
Its when they ask for metrics on the number of metrics you can supply them with that you should really start worrying.
Stuff like MTBF (mean time between failures) or %uptime are possible things to keep the bean counters happy. You can probably get the data straight from monitoring systems like 'mon' or nagios. On a longer timescale, do a regular report (annual, six monthly, whatever) on number of new desktop systems installed, number of OS upgrades done, that kind of thing. Bean counters like reports.
If you run a web server, google analytics can provide you with a huge amount of numbers to pass on to management...
I have no idea what they do with all this data anyway.
Its "open" in the sense of being the opposite of "closed". The source code is there, you can read it.
"Free" means you are free to do stuff to it (exactly what stuff you are free to do is license-dependent).
Your examples are pretty bogus: getting software to work with more than one compiler (or compiler system) is not trivial, even with so-called ANSI compliance. Try grepping source files for '#ifdef GCC' or '#ifdef CYGWIN' to see. And your apache example? Sure most HTTP servers will spit out static pages, but they might not support CGI, apache's virtual hosts, security, logging, mod-perl, mod-python, mod-ruby, mod-fortran (does that exist?) and so on. Any moderately advanced user of Apache is locked-in as tight as a gnat's chuff.
There are arguments for a truly open and free software stack from applications to the BIOS in certain fields, (for example in science to enable reproducible research), but that doesn't mean that all components in a stack with non-free and closed source parts are tainted. They are free. They are open. Whatever their requirements.
ABBA (whats the Unicode for a backwards 'B'?) were just first in alphabetical order, because popular beat combo Aaron's Aardvarks hadn't formed yet. And still haven't.
The chances of anything coming from Mars, are a million to one they said.
I get all my science from 70's prog rock concept albums based on HG Wells novels...