Or failing that, have a requirement that every person that wants to own guns has to go through training and an examination on the correct use, storage and maintenance of the weapon. All important points in the safe responsible use of a gun.
I love the little pointing nub/joystick thingies on some laptops. Someone really needs to come up with a catchy name. What I would really love is a stand alone usb keyboard with one of those for a desktop. Does anybody make one?
Anyway they are great, as you don't have to move your hands off the keyboard to move the mouse and you can use either hand and share the load abit between hands. For some reason I tend to use my left hand the most instead of my dominant right hand which is weird, especially as I use my right hand to use a normal mouse.
Or building houses, or fixing plumbing, repairing the cars and many other trades they have to be performed here. Manufacturing workers maybe in decline, but a guy that can pour a good concrete slab will still be in demand as long as buildings are still being built.
Well the question is if you would have actually done more study if you had not been working, and if so then yes your grades would have improved. The real answer is that people with good time management skills and ability to handle high stress situations will do well even having to juggle work and college, and they will do well later on in life to. The well off student that doesn't have to support themselves, doesn't have to have those skills (and consequently will have more problems in their future workplaces).
On a more political bent, the rise in the cost of a college degree is out pacing inflation and the hourly pay of the minimum wage jobs that most college students are able to get. So there is an increasing financial pressure on students, which really means that they either have to work more, come out of college with a larger debts, study part time and take longer to get the degree, or study at cheaper and less presitigious institutions.
Because it is the only way to actually test an individual about their knowledge without a lot of intimate contact with them. Now take home essays, projects and test are never all an individuals work but the combined effort of a whole study group.
The only other way is to have really small classes, and get to know the student well enough to understand what they know, and give a mark from that knowledge. This works in primary school, but it even starts to break down in high school let alone university. So in the end it is an arbitrary mark to quantify what the student knows against what they should know, spewed out in 2-3 hours or so. Not perfect but a practical solution to a complex problem, evaluating the knowledge of people.
In the end problem solving is the loser here as this type of testing favours people who rote learn and are practiced at operating under artificially short time frames, and not people who understanding the problem fully and solving it from the ground up and finding better ways of doing it.
Better security is to have two things, a keycard type thing and a password. Then it requires some person to a steal the keycard, (which risks detection), and either try some form of attack to crack the pasword (and this should be prevented by proper security implementation) or obtain the password by other means (some form of spying or getting the cooperation of the authorised individual).
I pretty much had the same solution written in a comment in this thread, except for how to actually calculate the time difference between the server and clients. Your solution could work, though I'm sure there are other ways. So my idea was the clients and regularly synchronisation of their clocks with the server. The server pushes sound files (mp3 or wav or whatever) along with the exact time to commence playback, thus synchronising the sound. Playing a playlist is easy as all it requires is for the server to calculate the correct times to start the playback of the next file, but a little bit of finesse will be required to handle things like the user changing songs quickly, skipping sections of a song or glitches in the network etc.
At the client level, actually getting accurate initiation of playback shouldn't be a big problem if you using a low latency or realtime operating system (opensource solution is to use realtime and low latency patched versions of the linux kernel). The JACK and ALSA projects actually already have some good tools and a large body of knowledge in dealing with similar low latency sound issues.
The streaming to various little sound clients using iTunes, or some hacking together of some opensource solution is pretty easy enough to setup. The problem that is really interesting, and which many people have said is difficult already is the synchronisation of two digital streams so two different sound sources in two different rooms are in sync. This is a difficult task.
One way I can see that it could work digitally (this is probably a case were an analog switching solution is probably more robust and easy to setup) is to have some combination sound server and clock synchronisation server. A number I've pulled out of my rear-end but have some memory hearing some where, that sound sources need to be within ~10 ms of each other to be in sync to a listener. My idea would be have dedicated sound clients and a dedicated sound server, the sound clients would be continually updated with the servers correct time from the central server (how networking delays would be dealt with here I haven't figured out, it would have to be within 5-10ms probably, but a quick google seemed to show a fair bit of literature on the topic). When a song is streamed to two clients there is also an instruction about the exact time the stream should start playing. Of course there would be some nasty headaches getting this system to be reasonable robust, dealing with networking delays and recovering from glitches. Assuming no gaps in the playback, the two streams should remain in sync, but if one stream skips a bit, it would require some hack to recover or wait till the next song to regain sync. Also the realtime hooks needed to get the accurate commencement of playback on the sound clients pretty much rules out getting this to work on a desktop computer running an OS without good real-time capabilities.
Well if you used a flag to kindle a fire say because you couldn't find anything else, well that would be illegal. But if you us the flag as a symbol of the US (which is its intent), and you are so disgusted or angry at the US or its government that you want to express that by burning the flag, that is protected by the first amendment as free speach. And the constitution trumps any other law.
Zinc is pretty harmless, and is actually necessary for life. It has important roles in the active sites of many enzymes that are vital. So it is very safe and this seems to be a bit of overzeolous reporting.
Cadmium and lead are dangerous as they do not have a biochemical function, and until resently exposure to high levels of them was unusual, so there are no biochemical or fast geochemical mechanisms to deal with these in the environment.
Arsenic is in between, as their are some rare organism that utilise arsenic for some purposes. It is generally not as bioaccumalative as Cadmium or lead, but it is still very dangerous in high enough concentrations.
What do you mean by unstable? And unstable in vivo, or in vitro? Embryonic stem cells are quite stable in culture and thus are reasonable easy to culture, and many techniques already exist from IVF (which pretty much means they are stable in vitro).
Immune rejection in vivo will be the usual cause of any "unstable" stem cells, and is a problem for both umbilical and embryonic stem cells as the animal/human and the cell line are usually not genetically identical.
Re:"New stem cell harvesting was outlawed in the U
on
US Stem Cells Contaminated
·
· Score: 2, Insightful
To be consistent then you have to be against invitro fertilization. As the techniques necessitates a certain amount of waste in the production of fertilized eggs which have to be destroyed anyway, and their is really no way around this (the irony, a way for invertile couples to have a baby requires the destruction of fertilized embryos). Protest and try and ban invitro-fertilization if you believe that life begins with fertilization and thus destroying fertilized eggs is immoral, but if you choose not to do not split hairs protesting stem cell research.
Umbilical cord stem cells are not a replacement for embryonic stem cells, though they are still useful. Also to say that one form works while the other doesn't is trying to second guess the research, as extrodinarily enough the stem cell research is designed to find out how exactly what the stem cells are capable of and what they are useful for, so we don't know what each type of stem cell will work for a particular situation and that is one of the reasons for doing research.
I was really trying to make the point that this thing could be packed in a crate and shipped without to much risk of it suffering damage, even if it accidently got dropped of the ramp while they are loading the truck or some other stupid event. A laptop is more likely to be fragile. Also it could be the musicians personal computer, and so may not be solely dedicated to the task of an instrument, which risks the thing having a driver conflict or virus or what ever preventing it working on any particular random night. Not pretty if you have an audience. I have been to enough laptop live PA's of various electronic artists to see the occasional embarassing situation of a laptop refusing to cooperate, or a program crashing in the middle of a performance. This happens rarely as the guys doing this are good at administering their computers, and the more experienced the better they are at it, but it does still happen. Obviously you can have a dedicated system for performances, but then the question of why not buy an off the shelf solution like this keyboard?
You are right that this makes sure all the controls are easily accessible from the keyboard.
As for boot up, I think it could be made to be pretty short being essentially an embedded device, and not a general purpose computer, and it should just work all night flawlessly or it will get a reputation for being a crap piece of equipment.
"Aesthetic appeal is top priority". Aesthetic appeal for it's own sake shouldn't be your top priority, your top priority is to convey the information you want to convey in the clearest and most concise way, as this is the purpose of giving a presentation. The only time it isn't your top priority is if you are trying to be deceitful and hide and obscure important information for some purpose like conning someone, or you are in advertising (is their actually a difference). It is pretty clear cut distinction. Sure you can put on some nice borders and themes if you want but in the end that doesn't matter one bit to the impact of your presentation, and can only detract from it if done badly. I have seen to many people choose bad themes combined with poor text choices (unreadable fonts, dark text on dark backgrounds etc) that simply detract from what would have otherwise been a good presentation to think that themes in powerpoint/keynote etc. is actually a necessary or even a good feature. It just adds in another avenue for someone not experienced in giving presentations to make another bad mistakes that make it even harder for the already suffering audience to understand.
Personally I use powerpoint as people used to use slide projectors. Not for talk outlines but using it as a convenient way to present visual information like graphs and pictures. Having simple solid white or at least a solid colour as a background helps the visuals pop out more readily instead of sinking into a complicated background.
Apple really pissed me off taking down the 10.2 supporting beta x11 server. I had it installed on one 10.2 mac and trashed the installer, then I little while later I wanted to put it on another 10.2 mac and all of a sudden I couldn't download it. It was a very "soup nazi" type moment, "no x11 server for you!". It also seemed pretty damn arbitrary since it obviously works on 10.2 quite well in the beta version.
>It's clunky, slow, disorganized, and makes really >unattractive slides. It cannot even compete with >Powerpoint, let alone with Keynote. hmmm well I could maybe say yes, it is clunky and slow, but unattractive slides? I strip out all colourful backgrounds and text, so this doesn't effect me. Clear white background with a black easy to read text. Colour should be used in the text and in pictures to highlight important points, differences and correlations. Using anything else but the simplest base colour scheme robs you of much of the effect of using colour in this way. Simply put start off simple and it gives you more options to create impact where you want to impress upon people important points.
Now if you don't actually have much to talk about, sure hide that fact with colourful borders and backgrounds and all the other tricks that people employ to give a talk seem more than it is by using window dressing.
Because their roadie proof. Single box, no interconnections from a keyboard to a computer to setup. Ruggidised compenents. Oh and a stable OS that is tuned for exactly the hardware of the keyboard, with only the necessary software etc. so the possibilities of embarassing crashes in the middle of a performance are very low.
Basically every night it can be plugged in by a thick as a brick roadie, and it will work and play flawlessly the whole night. Then it can be packed up again by the same roadie and shipped off to the next venue.
Wordprocessors and spreadsheets are complicated applications, and at some point trying to simplify things fails and you have to increase the complexity of the UI to provide functionality. Wordprocessors have always had the toolbars with a collection of buttons providing various functions and access to other toolbars along the top of the screen. And iWork follows this model to some extent to.
It is either have those buttons in a toolbar somewhere for easy access to common functions or waste time opening the top level menus, or learn obscure shortcut keys.
Look at some other complicated applications, say photoshop. This has a huge number of complicated toolbars. At some point you have to learn what each icon in the toolbar means. And their is no way to get around it.
I have never understood the logic that says having more than one way of doing something is bad UI. It would seem to me to be benificial to finding functions quickly if their is more than one place to find them.
Having multiple tabs in preferences. This kind of design, keeps all preferences in one preferences window, and clearly distinguishes between them by the name on the tab. The other common way is to have Icons in a folder, like the OSX system preferences which has a downside of having to return to the main window to access another preference pane, or something like a firefoz/mozilla style tree with nested preference panels under top level tabs which is probably the best, but can be frustrating at times when you are trying to track down the exact right preference panel if a preference doesn't fit neatly into one of the top level tab definitions.
Well they are kind of the same thing actually. NeoOffice was supposed to be a testing ground to work out how to port OOo to Aqua. Then they were then going to intergrate those changes back into the OOo.
They have decided that maybe that isn't a practical way to do it after all, and instead maintaining a separate fork "NeoOffice" is a better way.
The big problem eventually with this is that NeoOffice is forked off the OOo 1.x code base, and so will lack new features coming out for newer versions of OOo. How hard these features will be to backport, or alternatively how hard it is to fork off OOo again and do another port to aqua? I don't know but it isn't trivial.
OK so you want to do it cheap, and you seem quite handy. So here is an idea. buy a case with minimal air intakes (one to two if your can), and pipe that intake outside, with large diameter PVC plumbing. The fans blowing out of the case you put good filters on, and let them blow into the shed. That takes care of the guts. Wireless keyboards and mice? The truth is most external connections are pretty resiliant, but a little duct or insulation tape around the connected plugs can go along way to protect it even further. That leaves the remaining open parts of the system, the monitor vent holes (of CRT monitors) and the open buttons of the keyboard and mouse. For the mouse and keyboard you can buy quality sealed versions, or very cheap disposable unsealed versions, and you don't lose much from testing a cheap keyboard and mouse to see how long they last before commiting to a more expensive sealed keyboard and mouse. So that leaves the hard part, the monitor. LCD's can be bought totally sealed at a premium cost, or you can buy specially bought expensive cases for CRT's. Either that or you you try and put a cheaper monitor under glass in a container that you have made. Ideally get cooling air from the outside like the main box and vent it into the shed.
Protection against Microsoft pulling MS Office off Macs would actually be betterl served by throwing some developers and money at the OSX port of Openoffice at getting it properly intergrated (supposed to be easier to do in the upcoming v2.0 which has had much work done on it to make it more portable). It would actually be pretty cheap and provide the best non-MS MSOffice compatible Officesuite. I doubt that this new iwork will be able to handle Office documents anywhere near as well as OpenOffice currently does.
Slashdot Mirror
Or failing that, have a requirement that every person that wants to own guns has to go through training and an examination on the correct use, storage and maintenance of the weapon. All important points in the safe responsible use of a gun.
I love the little pointing nub/joystick thingies on some laptops. Someone really needs to come up with a catchy name. What I would really love is a stand alone usb keyboard with one of those for a desktop. Does anybody make one?
Anyway they are great, as you don't have to move your hands off the keyboard to move the mouse and you can use either hand and share the load abit between hands. For some reason I tend to use my left hand the most instead of my dominant right hand which is weird, especially as I use my right hand to use a normal mouse.
Or building houses, or fixing plumbing, repairing the cars and many other trades they have to be performed here. Manufacturing workers maybe in decline, but a guy that can pour a good concrete slab will still be in demand as long as buildings are still being built.
Well the question is if you would have actually done more study if you had not been working, and if so then yes your grades would have improved. The real answer is that people with good time management skills and ability to handle high stress situations will do well even having to juggle work and college, and they will do well later on in life to. The well off student that doesn't have to support themselves, doesn't have to have those skills (and consequently will have more problems in their future workplaces).
On a more political bent, the rise in the cost of a college degree is out pacing inflation and the hourly pay of the minimum wage jobs that most college students are able to get. So there is an increasing financial pressure on students, which really means that they either have to work more, come out of college with a larger debts, study part time and take longer to get the degree, or study at cheaper and less presitigious institutions.
Because it is the only way to actually test an individual about their knowledge without a lot of intimate contact with them. Now take home essays, projects and test are never all an individuals work but the combined effort of a whole study group.
The only other way is to have really small classes, and get to know the student well enough to understand what they know, and give a mark from that knowledge. This works in primary school, but it even starts to break down in high school let alone university. So in the end it is an arbitrary mark to quantify what the student knows against what they should know, spewed out in 2-3 hours or so. Not perfect but a practical solution to a complex problem, evaluating the knowledge of people.
In the end problem solving is the loser here as this type of testing favours people who rote learn and are practiced at operating under artificially short time frames, and not people who understanding the problem fully and solving it from the ground up and finding better ways of doing it.
Boxes is a perfectly adequate plural form. Also on principle I dislike "boxen" because of its widespread use in leet speak.
Better security is to have two things, a keycard type thing and a password. Then it requires some person to a steal the keycard, (which risks detection), and either try some form of attack to crack the pasword (and this should be prevented by proper security implementation) or obtain the password by other means (some form of spying or getting the cooperation of the authorised individual).
I pretty much had the same solution written in a comment in this thread, except for how to actually calculate the time difference between the server and clients. Your solution could work, though I'm sure there are other ways. So my idea was the clients and regularly synchronisation of their clocks with the server. The server pushes sound files (mp3 or wav or whatever) along with the exact time to commence playback, thus synchronising the sound. Playing a playlist is easy as all it requires is for the server to calculate the correct times to start the playback of the next file, but a little bit of finesse will be required to handle things like the user changing songs quickly, skipping sections of a song or glitches in the network etc.
At the client level, actually getting accurate initiation of playback shouldn't be a big problem if you using a low latency or realtime operating system (opensource solution is to use realtime and low latency patched versions of the linux kernel). The JACK and ALSA projects actually already have some good tools and a large body of knowledge in dealing with similar low latency sound issues.
The streaming to various little sound clients using iTunes, or some hacking together of some opensource solution is pretty easy enough to setup. The problem that is really interesting, and which many people have said is difficult already is the synchronisation of two digital streams so two different sound sources in two different rooms are in sync. This is a difficult task.
One way I can see that it could work digitally (this is probably a case were an analog switching solution is probably more robust and easy to setup) is to have some combination sound server and clock synchronisation server. A number I've pulled out of my rear-end but have some memory hearing some where, that sound sources need to be within ~10 ms of each other to be in sync to a listener. My idea would be have dedicated sound clients and a dedicated sound server, the sound clients would be continually updated with the servers correct time from the central server (how networking delays would be dealt with here I haven't figured out, it would have to be within 5-10ms probably, but a quick google seemed to show a fair bit of literature on the topic). When a song is streamed to two clients there is also an instruction about the exact time the stream should start playing. Of course there would be some nasty headaches getting this system to be reasonable robust, dealing with networking delays and recovering from glitches. Assuming no gaps in the playback, the two streams should remain in sync, but if one stream skips a bit, it would require some hack to recover or wait till the next song to regain sync. Also the realtime hooks needed to get the accurate commencement of playback on the sound clients pretty much rules out getting this to work on a desktop computer running an OS without good real-time capabilities.
Well if you used a flag to kindle a fire say because you couldn't find anything else, well that would be illegal. But if you us the flag as a symbol of the US (which is its intent), and you are so disgusted or angry at the US or its government that you want to express that by burning the flag, that is protected by the first amendment as free speach. And the constitution trumps any other law.
You forgot reason 3) To get something out of it, eg. a college education, or other training like joining the airforce to learn how to be a pilot.
Zinc is pretty harmless, and is actually necessary for life. It has important roles in the active sites of many enzymes that are vital. So it is very safe and this seems to be a bit of overzeolous reporting.
Cadmium and lead are dangerous as they do not have a biochemical function, and until resently exposure to high levels of them was unusual, so there are no biochemical or fast geochemical mechanisms to deal with these in the environment.
Arsenic is in between, as their are some rare organism that utilise arsenic for some purposes. It is generally not as bioaccumalative as Cadmium or lead, but it is still very dangerous in high enough concentrations.
What do you mean by unstable? And unstable in vivo, or in vitro? Embryonic stem cells are quite stable in culture and thus are reasonable easy to culture, and many techniques already exist from IVF (which pretty much means they are stable in vitro).
Immune rejection in vivo will be the usual cause of any "unstable" stem cells, and is a problem for both umbilical and embryonic stem cells as the animal/human and the cell line are usually not genetically identical.
To be consistent then you have to be against invitro fertilization. As the techniques necessitates a certain amount of waste in the production of fertilized eggs which have to be destroyed anyway, and their is really no way around this (the irony, a way for invertile couples to have a baby requires the destruction of fertilized embryos). Protest and try and ban invitro-fertilization if you believe that life begins with fertilization and thus destroying fertilized eggs is immoral, but if you choose not to do not split hairs protesting stem cell research.
Umbilical cord stem cells are not a replacement for embryonic stem cells, though they are still useful. Also to say that one form works while the other doesn't is trying to second guess the research, as extrodinarily enough the stem cell research is designed to find out how exactly what the stem cells are capable of and what they are useful for, so we don't know what each type of stem cell will work for a particular situation and that is one of the reasons for doing research.
Which is why people should distrust any presentation that is light on detail and full of flash.
I was really trying to make the point that this thing could be packed in a crate and shipped without to much risk of it suffering damage, even if it accidently got dropped of the ramp while they are loading the truck or some other stupid event. A laptop is more likely to be fragile. Also it could be the musicians personal computer, and so may not be solely dedicated to the task of an instrument, which risks the thing having a driver conflict or virus or what ever preventing it working on any particular random night. Not pretty if you have an audience. I have been to enough laptop live PA's of various electronic artists to see the occasional embarassing situation of a laptop refusing to cooperate, or a program crashing in the middle of a performance. This happens rarely as the guys doing this are good at administering their computers, and the more experienced the better they are at it, but it does still happen. Obviously you can have a dedicated system for performances, but then the question of why not buy an off the shelf solution like this keyboard?
You are right that this makes sure all the controls are easily accessible from the keyboard.
As for boot up, I think it could be made to be pretty short being essentially an embedded device, and not a general purpose computer, and it should just work all night flawlessly or it will get a reputation for being a crap piece of equipment.
"Aesthetic appeal is top priority".
Aesthetic appeal for it's own sake shouldn't be your top priority, your top priority is to convey the information you want to convey in the clearest and most concise way, as this is the purpose of giving a presentation. The only time it isn't your top priority is if you are trying to be deceitful and hide and obscure important information for some purpose like conning someone, or you are in advertising (is their actually a difference). It is pretty clear cut distinction. Sure you can put on some nice borders and themes if you want but in the end that doesn't matter one bit to the impact of your presentation, and can only detract from it if done badly. I have seen to many people choose bad themes combined with poor text choices (unreadable fonts, dark text on dark backgrounds etc) that simply detract from what would have otherwise been a good presentation to think that themes in powerpoint/keynote etc. is actually a necessary or even a good feature. It just adds in another avenue for someone not experienced in giving presentations to make another bad mistakes that make it even harder for the already suffering audience to understand.
Personally I use powerpoint as people used to use slide projectors. Not for talk outlines but using it as a convenient way to present visual information like graphs and pictures. Having simple solid white or at least a solid colour as a background helps the visuals pop out more readily instead of sinking into a complicated background.
Apple really pissed me off taking down the 10.2 supporting beta x11 server. I had it installed on one 10.2 mac and trashed the installer, then I little while later I wanted to put it on another 10.2 mac and all of a sudden I couldn't download it. It was a very "soup nazi" type moment, "no x11 server for you!". It also seemed pretty damn arbitrary since it obviously works on 10.2 quite well in the beta version.
>It's clunky, slow, disorganized, and makes really >unattractive slides. It cannot even compete with >Powerpoint, let alone with Keynote.
hmmm well I could maybe say yes, it is clunky and slow, but unattractive slides? I strip out all colourful backgrounds and text, so this doesn't effect me. Clear white background with a black easy to read text. Colour should be used in the text and in pictures to highlight important points, differences and correlations. Using anything else but the simplest base colour scheme robs you of much of the effect of using colour in this way. Simply put start off simple and it gives you more options to create impact where you want to impress upon people important points.
Now if you don't actually have much to talk about, sure hide that fact with colourful borders and backgrounds and all the other tricks that people employ to give a talk seem more than it is by using window dressing.
Because their roadie proof. Single box, no interconnections from a keyboard to a computer to setup. Ruggidised compenents. Oh and a stable OS that is tuned for exactly the hardware of the keyboard, with only the necessary software etc. so the possibilities of embarassing crashes in the middle of a performance are very low.
Basically every night it can be plugged in by a thick as a brick roadie, and it will work and play flawlessly the whole night. Then it can be packed up again by the same roadie and shipped off to the next venue.
Wordprocessors and spreadsheets are complicated applications, and at some point trying to simplify things fails and you have to increase the complexity of the UI to provide functionality. Wordprocessors have always had the toolbars with a collection of buttons providing various functions and access to other toolbars along the top of the screen. And iWork follows this model to some extent to.
It is either have those buttons in a toolbar somewhere for easy access to common functions or waste time opening the top level menus, or learn obscure shortcut keys.
Look at some other complicated applications, say photoshop. This has a huge number of complicated toolbars. At some point you have to learn what each icon in the toolbar means. And their is no way to get around it.
I have never understood the logic that says having more than one way of doing something is bad UI. It would seem to me to be benificial to finding functions quickly if their is more than one place to find them.
Having multiple tabs in preferences. This kind of design, keeps all preferences in one preferences window, and clearly distinguishes between them by the name on the tab. The other common way is to have Icons in a folder, like the OSX system preferences which has a downside of having to return to the main window to access another preference pane, or something like a firefoz/mozilla style tree with nested preference panels under top level tabs which is probably the best, but can be frustrating at times when you are trying to track down the exact right preference panel if a preference doesn't fit neatly into one of the top level tab definitions.
Well they are kind of the same thing actually. NeoOffice was supposed to be a testing ground to work out how to port OOo to Aqua. Then they were then going to intergrate those changes back into the OOo.
They have decided that maybe that isn't a practical way to do it after all, and instead maintaining a separate fork "NeoOffice" is a better way.
The big problem eventually with this is that NeoOffice is forked off the OOo 1.x code base, and so will lack new features coming out for newer versions of OOo. How hard these features will be to backport, or alternatively how hard it is to fork off OOo again and do another port to aqua? I don't know but it isn't trivial.
"Imagine a Beowulf cluster of those...."
That would be a brewery.
OK so you want to do it cheap, and you seem quite handy. So here is an idea. buy a case with minimal air intakes (one to two if your can), and pipe that intake outside, with large diameter PVC plumbing. The fans blowing out of the case you put good filters on, and let them blow into the shed. That takes care of the guts. Wireless keyboards and mice? The truth is most external connections are pretty resiliant, but a little duct or insulation tape around the connected plugs can go along way to protect it even further. That leaves the remaining open parts of the system, the monitor vent holes (of CRT monitors) and the open buttons of the keyboard and mouse. For the mouse and keyboard you can buy quality sealed versions, or very cheap disposable unsealed versions, and you don't lose much from testing a cheap keyboard and mouse to see how long they last before commiting to a more expensive sealed keyboard and mouse. So that leaves the hard part, the monitor. LCD's can be bought totally sealed at a premium cost, or you can buy specially bought expensive cases for CRT's. Either that or you you try and put a cheaper monitor under glass in a container that you have made. Ideally get cooling air from the outside like the main box and vent it into the shed.
Protection against Microsoft pulling MS Office off Macs would actually be betterl served by throwing some developers and money at the OSX port of Openoffice at getting it properly intergrated (supposed to be easier to do in the upcoming v2.0 which has had much work done on it to make it more portable). It would actually be pretty cheap and provide the best non-MS MSOffice compatible Officesuite. I doubt that this new iwork will be able to handle Office documents anywhere near as well as OpenOffice currently does.