In Google case, Id say they just seal off the container and be done with it. If there is a fire, they bring in a new (40') box.
But anyway. A rack mount HP UPS I installed in the past year has a stand-off that you can hook into the "Big Red Button System". I'm guessing such hookups are either standard on rack mount units, or at least it wouldnt be hard to find models with that feature.
As a rule, for ever Linux based desktop that I've run (work, home, over 15 years.. maybe 50 unique installs/upgrades that lasted), the most most consistently flaky glue has been Java working in any browser.
Historically, I stumble across, and need, Java applets pretty infrequently. As a result, nearly 100% of the time I do stumble across, and want to run, a Java applet, it involves somewhere between 5 and 30 minutes of fucking around to make it work. Netscape, Mozilla, now Firefox automatic plugin find/install has worked exactly 0 times for a JVM. And, except for the last year or so, it was always a regular fight at java.sun.com to figure out exactly WTF I needed. I consider myself an educated person, but my paths usually led me to download a 200MB blob that included a IDE, but not a plugin for my always up to date browser. And does installing the JRE actually install the pulgin, when I do manage to get the right thing? No. Creating symlinks is apparently beyond Suns ability. Or maybe its a Java security thing; its quite willing to take up 2GB of memory, but creating a symlink is a violation of security. Fuckers.
In the last 18 months, I spent 9 of them as a Java developer, on a team of 13. I've yet to hear an adequate explanation as to what the difference between JSE, JEE, Java FX, J2ME, JDK, in less then 500 words. Sure, I know it now. Well, maybe not. WTF is JavaFX, and why, just now, when I click on "download JEE", it asks me to download "Glassfish"? THAT ISN'T WHAT I FUCKING ASKED FOR.
It amazes me that people are willing to put up with this shit at all. For a language that isn't even that good.
Anyway. If Sun has finally managed to ship something that actually manages to install a plugin without me digging around on the command line, I'm amazed. That it installs an extension as well is a small price to play.
Another specific example was the goal of ARPANET.. To allow remote access to centralized (scarce, and expertly managed) computer resources.
On an unrelated matter, (this was pre-GPLv3 discussion) I remember reading a mailing list thread where RMS seemed suppressed at the existence of gaming consoles (paraphrasing: why would anyone want a computer they cant program?)
A major debate these days is firmware blobs, if they count as source, if there existence in the kernel, in distros, etc. Personally, I don't see this as being any "worse" then actual firmware... But firmware has been around for... For ever. At a speech RMS gave locally a month ago, I kinda called him on this, and he more or less dismissed the issue as saying that firmware wasn't software, but loading a blob from disk clearly was.
From a purely pragmatic point of view, 10 years ago things like SCSI HBAs were not upgradeable in the field (if at all); users could not change the code running on the board at all. Today, I can load bug-fixed firmware from the driver; I still can't make local changes to the code. Pragmatically I'm better off today. Freedom wise, I'm no worse off.
I guess my point is that these things: SaaS, Tivos or Playstations, loadable firmware seem to be issues that the community is struggling with currently. But they are all pretty predictable.
Also, how is Hurd going? "Zero know bugs" release policy for Emacs. GCC being basically forked (and un-forked).
I'll take RMS over a "series of tubes" politician, but I think it is pretty clear that the leadership is out of touch with modern technology.
Yes there was SaaS when the GPL was written. Significantly before it. Or at least attempts at it.
The Multics project has the very specific purpose of making computing resources a service, just like phones and electricity. CompuServe was one specific example, setup in 1969, to provide time-sharing services to external companies.
Arguably, pre-1975, SaaS was the default, not the exception. Computers were rented, perhaps on your premises, and tended to by vendor techs.
Anyway, this mantra of "we didn't know about it" really makes [GPL supporters] look stupid. I mean really, are you trying to tell me that RMS had never heard of Multics?
Reviewed by a board, and then maybe, they get a prize after their name is drawn? I see that as total bullshit, treating your employees as children. And thinking you are not, which makes it worse.
Why do I suspect that when you say "prize" you don't mean a million dollars in 20 $50k installments over 20 years? Maybe something like a $10 gift card to Starbucks? Am I close?
How about this: give cash - or stock, not options, stock - to people who's ideas are implemented? Straight up: you have a good idea not within your direct job responsibility, and we implement it, you get cash.
I'm sorry if the connection wasn't clear enough for you.
Having a large prison population is evidence of problem with the big picture. Having a large prison population is evidence that ones government is attacking symptoms, not causes, of crime.
If someone breaks the law, they should be punished for it. Send criminals to jail.
Some crimes fall into a category where, I think, pretty much whatever government does, they will still happen. Serial killers, for example.
Other crimes are different. Street level drug dealing is a prime example. There are people addicted to drugs, there a new people wanting to try drugs. There is a demand for drugs. Arresting a drug dealer today, and putting him in jail does nothing to reduce the demand for his product. Someone else will quickly take his spot on the corner, as it were.
Street level dealers, at least, are generally from a poor, uneducated background. They deal drugs because (apparently) it looks like a better gig then anything else.
If _all_ that government does is arrest drug dealers, then nothing has been accomplished - except making lots of jobs for prison guards. There are still drug dealers on the street.
1 drug dealer is a criminal problem. 100 drug dealers is a social problem.
In this specific case, an alternative approach would focus on reducing demand for drugs. Treat drug addiction as a health care problem. (well, in the US that would mean having good health care, in general, but I digress). Further, make drug dealing a less appealing source of income, by encouraging better alternatives. Better schools means people getting jobs, and having better options for jobs - which means fewer people willing to be dealers.
Re:Because you don't need more cycles in biz
on
Less Is Moore
·
· Score: 1
I, for one, care very little about battery life. I mean, I just don't care, as long as its more then 5 minutes. I'm either just moving to a place where there is also power (a meeting room, say), where a shutdown would just suck, or changing buildings, where a shutdown is acceptable. (of course, if I trusted suspend or hibernate, or whatever, I could use those).
Yes, yes, there are scenarios where some people actually care about battery life beyond a few minutes... But I think that a huge chunk of people just don't care.
4 hours.. That is a lot? A scenario where I might want to use a battery is if I'm dragging my laptop on course, or over to a friends house, and want to leave the power adapter behind. Courses last longer then 4 hours, and hanging out with friends could as well. So I still need to plug in. Your 4 hours is, oh, 3:55 more minutes then I need.
Unless battery life gets to the point where the "default" mode would be to go without power (as with a cell phone, or PDA), then fuck it. It doesn't matter. I will need to plug in the laptop, so I will need to bring the power supply. And if I have the power supply with me, its trivial to use.
In short, if it cant do 10+ hours, then it might as well not be able to do more then 5 minutes.
"Built" means what, exactly? Components are built both in Italy and the UK; the airframe is assembled in Yeovil. Maybe the electronics suit is bolted on somewhere in the USA, but the bulk of it is European.
Well, that may be fair, and may be the reality. IBM does continue to write paychecks to a bunch of Eclipse people;-)
OTOH, Eclipse, indeed, due to its nature as a "platform", attracts lots of external developers. A bunch of the plugins have become core features (well, features shipping by default in the IDE). Mylyn come quickly to mind.
The real question would be what would happen if you submitted a patch against a bug in Eclipse vs OOo? I've never tried to do either (I can hardly read C++, I might be able to do some trivial Java stuff)...
But my understanding is that Eclipse would be easier. First off, its my understanding, that most people give up so much as trying to get a OOo build environment set up. Sure, shitty technology isn't shitty community relations.... Maintaining that shitty technology for 8 years because it "works for us" is shitty community relations. Consider Firefox ("bloat/leak" debate notwithstanding), while they might very well close things as "valid, won't fix" or, more likely, "valid, but doesn't conform to our narrow, focused, vision"... But they have gone to great lengths to allow for extensions to be built. And as with Eclipse, if/when these are proven as extensions, they may be picked up by the core project.
OpenSolaris code is (or was until recently) released as infrequent snapshots of their firewalled-off-the-world SCM system. Go-OO, well, Go-OO exists.
Sun is just going through the motions. Well, one motion: give away the code. And yeah, that is amazing. And more then most big companies do. But giving away code isn't the same thing as being involved in the community.
I'd think that IBM would be the largest, if you are counting something as meaningless as LOC. Eclipse is huge, maybe at least as big as OpenOffice.
Sun is very anti-community. MySQL was arguably anti-community as well.
The JCP: Who, at the end of the day decides what happens? Sun employees. OO.o: Who, at the end of the day decides what happens? Sun employees. MySQL: Who, at the end of the day decides what happens? A year ago, MySQL employees, now Sun employees.
Compare that with:
Kernel: Who, at the end of the day decides what happens? Technical leadership; the community Mozilla: Who, at the end of the day decides what happens? Mozilla employees, sure, but not Google or Netscape; an organization formed specifically for that codebase, arms length from anyone else Eclipse: Who, at the end of the day decides what happens? I'm not as sure:), but an organization formed specifically for that codebase, arms length from anyone else
Sure, specific Sun employees may be empowered to contribute to external projects. And sure, they throw a metric crap-load of very useful code over the wall, but what they have regularly proven themselves really bad at is actually developing any kind of net new community.
Audited internally or externally? By the IRS, or just an outside firm?
For for-profit, publicly traded companies, the "interested parties" are shareholders. They can easily see when there interests are being met: share prices go up and/or dividends are paid. Fancy financial reports are nice, but all that really matters is a check. Customers only care if they get reasonable value for the price. At the end of the day, as a shareholder or customer, I don't care if the officers of the company spend company money on fancy boats... provided that I'm getting good dividends and/or cheap products (relative to their competition).
With non-for-profits, its less clear if they are "successful". And then your not-for-profit might also be a charity. Either way, be it people buying stuff from them (or using their services), or giving donations, those people choosing that org over another might do so because they are "good". They cant easily check on dividends to see that, so an independent audit might be how they validate their decisions. As someone making donations, or buying product, from a non-profit, I care (more) about the officers doing shit like renting boats on the companies dime.
I'll disagree on this point. Windows - the NT Kernel, at least - is very "secure", by design. Excluding things like Multics, or SELinux with MAC, NT has about the most comprehensive security model available.
Its insecure in its implementation.
- The security resolution process didn't agree with the very limited early documentation; developers gave up and told everyone to "run as administrator". - Even for something simple, given the poor early documentation, developers gave up and told everyone to "run as administrator". - Lots (most?) NT and/or XP developers came from a DOS/9x background where security <em>was</em> non-existent, so they gave up, and developers gave up and told everyone to "run as administrator" - These things are also true of many internal MS products, encouraging a community of developers who give up and tell everyone to "run as administrator" - Users were similarly trained with DOS, 9x that security doesn't matter, and thus even with the option of doing the right thing, they gave up and run everything as administrator (and/or click "OK" to everything).
So. The Windows security implementation is shitty. And the windows security culture is shitty. But the windows security design is pretty good,
Within the context of a publicly traded company there there is exactly one point of ethics or morality that matters: make the shareholders money.
Bitching and complaining about companies doing something that isn't nice is pointless. Its like tripping, breaking your ankle, and then complaining that gravity didn't shut of. Gravity pulls down; deal with it. As a person, buy sensible shoes. As a government, have building and labor codes that require railings and fall-arresting systems.
If you want companies to behave, as an individual, shop elsewhere. As a government, legislate that behavior as being required (and make the penalties sufficiently large as to actually be a financial intensive). Companies treat their employees "poorly"? Increase them minimum wage. Allow them to unionize. Companies pollute the environment? Have inspectors, and fine them, severely.
If think you are responding to the (unasked question), "why not just keep 1 shuttle around?" I diddnt think that was on the table. But maybe it should be. And to respond to you response: or you could just restrict missions to those where a crippled shuttle _could_ get to the ISS. Then you can leisurely launch Soyuz's. and all is good.
There is a difference between & and &&. & is a bitwise operation and && is a logical one. Which are different. Given that PHP has an AND (which has a different precedence then &&, I grant), it would not be much a stretch to drop both & and && in favor of a (new) bit_and() function and just using AND with parenthesizes. bit_and() could also work across platforms with different word lengths, too (or throw an exception).
It was dangerous in the '60s because you had zero chance of rescue if you get in trouble. If humanity is going to be building telescopes on the moon, that means moon bases. And moon tow trucks. Or at least other rovers around.
Fiberoptics would be unaffected by solar radiation.
Getting the cable even 6" under the surface would be pretty protective of most threats, I think.
The alternate would be what.. A radio relay satellite? What are the relative odds of something hitting a satellite and a cable 5cm wide x 5000km long? And the relative cost of launching a new satellite and some lunar fiber splicing when something does hit it?
Slashdot Mirror
A true manual tranny has somewhere between 1 and 0 sticks... and usually underneath a skirt.
Destination Unknown, indeed.
Concrete pylons supported by air don't support very much, very long.
In Google case, Id say they just seal off the container and be done with it. If there is a fire, they bring in a new (40') box.
But anyway. A rack mount HP UPS I installed in the past year has a stand-off that you can hook into the "Big Red Button System". I'm guessing such hookups are either standard on rack mount units, or at least it wouldnt be hard to find models with that feature.
As a rule, for ever Linux based desktop that I've run (work, home, over 15 years.. maybe 50 unique installs/upgrades that lasted), the most most consistently flaky glue has been Java working in any browser.
Historically, I stumble across, and need, Java applets pretty infrequently. As a result, nearly 100% of the time I do stumble across, and want to run, a Java applet, it involves somewhere between 5 and 30 minutes of fucking around to make it work. Netscape, Mozilla, now Firefox automatic plugin find/install has worked exactly 0 times for a JVM. And, except for the last year or so, it was always a regular fight at java.sun.com to figure out exactly WTF I needed. I consider myself an educated person, but my paths usually led me to download a 200MB blob that included a IDE, but not a plugin for my always up to date browser. And does installing the JRE actually install the pulgin, when I do manage to get the right thing? No. Creating symlinks is apparently beyond Suns ability. Or maybe its a Java security thing; its quite willing to take up 2GB of memory, but creating a symlink is a violation of security. Fuckers.
In the last 18 months, I spent 9 of them as a Java developer, on a team of 13. I've yet to hear an adequate explanation as to what the difference between JSE, JEE, Java FX, J2ME, JDK, in less then 500 words. Sure, I know it now. Well, maybe not. WTF is JavaFX, and why, just now, when I click on "download JEE", it asks me to download "Glassfish"? THAT ISN'T WHAT I FUCKING ASKED FOR.
It amazes me that people are willing to put up with this shit at all. For a language that isn't even that good.
Anyway. If Sun has finally managed to ship something that actually manages to install a plugin without me digging around on the command line, I'm amazed. That it installs an extension as well is a small price to play.
I've been saying this basically for ever.
Another specific example was the goal of ARPANET.. To allow remote access to centralized (scarce, and expertly managed) computer resources.
On an unrelated matter, (this was pre-GPLv3 discussion) I remember reading a mailing list thread where RMS seemed suppressed at the existence of gaming consoles (paraphrasing: why would anyone want a computer they cant program?)
A major debate these days is firmware blobs, if they count as source, if there existence in the kernel, in distros, etc. Personally, I don't see this as being any "worse" then actual firmware... But firmware has been around for... For ever. At a speech RMS gave locally a month ago, I kinda called him on this, and he more or less dismissed the issue as saying that firmware wasn't software, but loading a blob from disk clearly was.
From a purely pragmatic point of view, 10 years ago things like SCSI HBAs were not upgradeable in the field (if at all); users could not change the code running on the board at all. Today, I can load bug-fixed firmware from the driver; I still can't make local changes to the code. Pragmatically I'm better off today. Freedom wise, I'm no worse off.
I guess my point is that these things: SaaS, Tivos or Playstations, loadable firmware seem to be issues that the community is struggling with currently. But they are all pretty predictable.
Also, how is Hurd going? "Zero know bugs" release policy for Emacs. GCC being basically forked (and un-forked).
I'll take RMS over a "series of tubes" politician, but I think it is pretty clear that the leadership is out of touch with modern technology.
Yes there was SaaS when the GPL was written. Significantly before it. Or at least attempts at it.
The Multics project has the very specific purpose of making computing resources a service, just like phones and electricity. CompuServe was one specific example, setup in 1969, to provide time-sharing services to external companies.
Arguably, pre-1975, SaaS was the default, not the exception. Computers were rented, perhaps on your premises, and tended to by vendor techs.
Anyway, this mantra of "we didn't know about it" really makes [GPL supporters] look stupid. I mean really, are you trying to tell me that RMS had never heard of Multics?
Reviewed by a board, and then maybe, they get a prize after their name is drawn? I see that as total bullshit, treating your employees as children. And thinking you are not, which makes it worse.
Why do I suspect that when you say "prize" you don't mean a million dollars in 20 $50k installments over 20 years? Maybe something like a $10 gift card to Starbucks? Am I close?
How about this: give cash - or stock, not options, stock - to people who's ideas are implemented? Straight up: you have a good idea not within your direct job responsibility, and we implement it, you get cash.
I'm sorry if the connection wasn't clear enough for you.
Having a large prison population is evidence of problem with the big picture. Having a large prison population is evidence that ones government is attacking symptoms, not causes, of crime.
If someone breaks the law, they should be punished for it. Send criminals to jail.
Some crimes fall into a category where, I think, pretty much whatever government does, they will still happen. Serial killers, for example.
Other crimes are different. Street level drug dealing is a prime example. There are people addicted to drugs, there a new people wanting to try drugs. There is a demand for drugs. Arresting a drug dealer today, and putting him in jail does nothing to reduce the demand for his product. Someone else will quickly take his spot on the corner, as it were.
Street level dealers, at least, are generally from a poor, uneducated background. They deal drugs because (apparently) it looks like a better gig then anything else.
If _all_ that government does is arrest drug dealers, then nothing has been accomplished - except making lots of jobs for prison guards. There are still drug dealers on the street.
1 drug dealer is a criminal problem. 100 drug dealers is a social problem.
In this specific case, an alternative approach would focus on reducing demand for drugs. Treat drug addiction as a health care problem. (well, in the US that would mean having good health care, in general, but I digress). Further, make drug dealing a less appealing source of income, by encouraging better alternatives. Better schools means people getting jobs, and having better options for jobs - which means fewer people willing to be dealers.
I agree, shit head laptops are not an option.
I, for one, care very little about battery life. I mean, I just don't care, as long as its more then 5 minutes. I'm either just moving to a place where there is also power (a meeting room, say), where a shutdown would just suck, or changing buildings, where a shutdown is acceptable. (of course, if I trusted suspend or hibernate, or whatever, I could use those).
Yes, yes, there are scenarios where some people actually care about battery life beyond a few minutes... But I think that a huge chunk of people just don't care.
4 hours.. That is a lot? A scenario where I might want to use a battery is if I'm dragging my laptop on course, or over to a friends house, and want to leave the power adapter behind. Courses last longer then 4 hours, and hanging out with friends could as well. So I still need to plug in. Your 4 hours is, oh, 3:55 more minutes then I need.
Unless battery life gets to the point where the "default" mode would be to go without power (as with a cell phone, or PDA), then fuck it. It doesn't matter. I will need to plug in the laptop, so I will need to bring the power supply. And if I have the power supply with me, its trivial to use.
In short, if it cant do 10+ hours, then it might as well not be able to do more then 5 minutes.
"Built" means what, exactly? Components are built both in Italy and the UK; the airframe is assembled in Yeovil. Maybe the electronics suit is bolted on somewhere in the USA, but the bulk of it is European.
All the more reason to write your software in a portable way. Even if you only care about one platform today, that platform will change over time.
Well, that may be fair, and may be the reality. IBM does continue to write paychecks to a bunch of Eclipse people ;-)
OTOH, Eclipse, indeed, due to its nature as a "platform", attracts lots of external developers. A bunch of the plugins have become core features (well, features shipping by default in the IDE). Mylyn come quickly to mind.
The real question would be what would happen if you submitted a patch against a bug in Eclipse vs OOo? I've never tried to do either (I can hardly read C++, I might be able to do some trivial Java stuff)...
But my understanding is that Eclipse would be easier. First off, its my understanding, that most people give up so much as trying to get a OOo build environment set up. Sure, shitty technology isn't shitty community relations.... Maintaining that shitty technology for 8 years because it "works for us" is shitty community relations. Consider Firefox ("bloat/leak" debate notwithstanding), while they might very well close things as "valid, won't fix" or, more likely, "valid, but doesn't conform to our narrow, focused, vision"... But they have gone to great lengths to allow for extensions to be built. And as with Eclipse, if/when these are proven as extensions, they may be picked up by the core project.
OpenSolaris code is (or was until recently) released as infrequent snapshots of their firewalled-off-the-world SCM system. Go-OO, well, Go-OO exists.
Sun is just going through the motions. Well, one motion: give away the code. And yeah, that is amazing. And more then most big companies do. But giving away code isn't the same thing as being involved in the community.
I'd think that IBM would be the largest, if you are counting something as meaningless as LOC. Eclipse is huge, maybe at least as big as OpenOffice.
:), but an organization formed specifically for that codebase, arms length from anyone else
Sun is very anti-community. MySQL was arguably anti-community as well.
The JCP: Who, at the end of the day decides what happens? Sun employees.
OO.o: Who, at the end of the day decides what happens? Sun employees.
MySQL: Who, at the end of the day decides what happens? A year ago, MySQL employees, now Sun employees.
Compare that with:
Kernel: Who, at the end of the day decides what happens? Technical leadership; the community
Mozilla: Who, at the end of the day decides what happens? Mozilla employees, sure, but not Google or Netscape; an organization formed specifically for that codebase, arms length from anyone else
Eclipse: Who, at the end of the day decides what happens? I'm not as sure
Sure, specific Sun employees may be empowered to contribute to external projects. And sure, they throw a metric crap-load of very useful code over the wall, but what they have regularly proven themselves really bad at is actually developing any kind of net new community.
Or, as an alternative, you could actually make technology that doesn't suck and that does what users want and expect.
Audited internally or externally? By the IRS, or just an outside firm?
For for-profit, publicly traded companies, the "interested parties" are shareholders. They can easily see when there interests are being met: share prices go up and/or dividends are paid. Fancy financial reports are nice, but all that really matters is a check. Customers only care if they get reasonable value for the price. At the end of the day, as a shareholder or customer, I don't care if the officers of the company spend company money on fancy boats... provided that I'm getting good dividends and/or cheap products (relative to their competition).
With non-for-profits, its less clear if they are "successful". And then your not-for-profit might also be a charity. Either way, be it people buying stuff from them (or using their services), or giving donations, those people choosing that org over another might do so because they are "good". They cant easily check on dividends to see that, so an independent audit might be how they validate their decisions. As someone making donations, or buying product, from a non-profit, I care (more) about the officers doing shit like renting boats on the companies dime.
I'll disagree on this point. Windows - the NT Kernel, at least - is very "secure", by design. Excluding things like Multics, or SELinux with MAC, NT has about the most comprehensive security model available.
Its insecure in its implementation.
- The security resolution process didn't agree with the very limited early documentation; developers gave up and told everyone to "run as administrator".
- Even for something simple, given the poor early documentation, developers gave up and told everyone to "run as administrator".
- Lots (most?) NT and/or XP developers came from a DOS/9x background where security <em>was</em> non-existent, so they gave up, and developers gave up and told everyone to "run as administrator"
- These things are also true of many internal MS products, encouraging a community of developers who give up and tell everyone to "run as administrator"
- Users were similarly trained with DOS, 9x that security doesn't matter, and thus even with the option of doing the right thing, they gave up and run everything as administrator (and/or click "OK" to everything).
So. The Windows security implementation is shitty. And the windows security culture is shitty. But the windows security design is pretty good,
Within the context of a publicly traded company there there is exactly one point of ethics or morality that matters: make the shareholders money.
Bitching and complaining about companies doing something that isn't nice is pointless. Its like tripping, breaking your ankle, and then complaining that gravity didn't shut of. Gravity pulls down; deal with it. As a person, buy sensible shoes. As a government, have building and labor codes that require railings and fall-arresting systems.
If you want companies to behave, as an individual, shop elsewhere. As a government, legislate that behavior as being required (and make the penalties sufficiently large as to actually be a financial intensive). Companies treat their employees "poorly"? Increase them minimum wage. Allow them to unionize. Companies pollute the environment? Have inspectors, and fine them, severely.
If think you are responding to the (unasked question), "why not just keep 1 shuttle around?" I diddnt think that was on the table. But maybe it should be. And to respond to you response: or you could just restrict missions to those where a crippled shuttle _could_ get to the ISS. Then you can leisurely launch Soyuz's. and all is good.
who are you calling a dollar dollar carrot?
There is a difference between & and &&. & is a bitwise operation and && is a logical one. Which are different. Given that PHP has an AND (which has a different precedence then &&, I grant), it would not be much a stretch to drop both & and && in favor of a (new) bit_and() function and just using AND with parenthesizes. bit_and() could also work across platforms with different word lengths, too (or throw an exception).
There were exactly two differences between 3.0 and 3.1:
3.1 had support for 386 protected mode. And one of the two included games were different.
It was dangerous in the '60s because you had zero chance of rescue if you get in trouble. If humanity is going to be building telescopes on the moon, that means moon bases. And moon tow trucks. Or at least other rovers around.
Fiberoptics would be unaffected by solar radiation.
Getting the cable even 6" under the surface would be pretty protective of most threats, I think.
The alternate would be what.. A radio relay satellite? What are the relative odds of something hitting a satellite and a cable 5cm wide x 5000km long? And the relative cost of launching a new satellite and some lunar fiber splicing when something does hit it?