Linux's chroot would be PITA. It would kill download. Capabilities might help, but Java permissions are far better (you can give an extension no permission for sockets or disk or allow just one directory or...). Yes, I know, capabilities can almost the same level of control (but not quite).
The problem here anyway has nothing to do with language choice
I disagree strongly. Although there have been (and will be) holes in Java runtime, it still has the goal of restricting programs from attacks. Something C++ really cannot do.
Binary files (as long as they are not large - YMMV) are no problem for any VCS.
Rename/move is better handled in e.g. Mercurial than in SVN - the version history follows. But you need to do it in the DVCS (e.g. "hq mv A B" instead of "mv A B").
Photo album where the photos do not change (almost at all) might not be the best for DVCS, they are designed for text files where merge makes sense and file size (and therefore history size) is not huge. After all, they do waste disk (you need the repository and the working copy).
I agree large files may pose a problem, especially if one of the machines is "memory limited".
Adding files is very much automatic (GUI frontends handle that, e.g. qct or tortoisehq for Mercurial).
My point was, if s/he is already using SVN, then Mercurial (or Git, or...) would be a drop in replacement which more than solves all problems BETTER than Unison and SVN combined.
They are quite easy (to learn), they handle merges apparently much better, they give version history, rollback, branches, etc. After all, the OP was using SVN.
How does it handle merges? That is, situations where the data has been modified in both (or several) machines.
You see, there are distributed version control systems (Mercurial,...) which are designed to handle that kind of situations. And they handle them extremely well (at least Mercurial does and I have no reason to believe the others do not).
Better yet, use Mercurial (or any other distributed version control system). It is designed for this kind of use!
I have three clones (of a project), one in desktop, one in laptop and one in an USB stick. Just "pull" (or "push") and "commit". Sometimes a "merge". With network connectivity you do not need the USB stick (in my case it is just a bit easier).
Should any of those clones/repositories die it would not be a big problem.
Note: this is not a backup solution, this is a solution to sync the machines!
Firefox on Ubuntu (8.04) is horribly slow. This page (Slashdot) has occasionally almost half a second delays while scrolling down. Or while I am editing here, "end" key takes sometimes 200-300ms.
One terabyte HD costs next to nothing, why on earth should I not rip my CDs to FLAC?
Are you sure none of my music in no circumstances I listen does exhibit hearable loss when encoded to XXXbit/s MP3?
I originally encoded my CD's to 128kbit/s OGG (that was beta 1.0 of the encoder), I had to re-rip everything because one song in one place was really horrible. Sure, perhaps final 1.0 with 190kbit/s or whatnot that particular point might have been OK, but why bother?
I'm a Finn. I've been waiting that for more than I've been living (small exaggeration is needed in this case), which makes half a century.
Let me tell you a story. I once went to Soviet Union and got out of Russia. The people in there, when I told the historic event I heard from radio, said "nothing is going to change".
After reference to nazi the discussion will go so bad that no sensible argument is ever going to appear.
(If you keep on going to call some enviro-nazis, aren't the other group those-who-did-not-deny-raping-and-killing-a-girl-in-1990-antienvironmentalists?)
I think government should force companies to address the needs of disabled as long as it is reasonable. For example there is no point allowing programs to use colours which are impossible for colour blind to distinguish thus making the use of the program unnecessarily difficult.
Whether in this case the modifications needed are substantial or not is different matter to which I have no opinion.
Linux's chroot would be PITA. It would kill download. Capabilities might help, but Java permissions are far better (you can give an extension no permission for sockets or disk or allow just one directory or ...). Yes, I know, capabilities can almost the same level of control (but not quite).
(I have no clue about Vista, sorry)
Either but *NOT* both? If so, Unison might be perfect for you.
It isn't for me.
Actually, garbage collection does protect against reuse of "freed" pointer attack (after "free" you cannot access the memory).
This is requirement for safe code, though far from being sufficient.
The problem here anyway has nothing to do with language choice
I disagree strongly. Although there have been (and will be) holes in Java runtime, it still has the goal of restricting programs from attacks. Something C++ really cannot do.
Do you happen work for Microsoft, by chance?
Just curious, nothing personal.
There really needs to be Java (or other "managed" language based) based browser (like Lobo). Unfortunately Lobo is not (yet?) ready for prime time.
Binary files (as long as they are not large - YMMV) are no problem for any VCS.
Rename/move is better handled in e.g. Mercurial than in SVN - the version history follows. But you need to do it in the DVCS (e.g. "hq mv A B" instead of "mv A B").
Photo album where the photos do not change (almost at all) might not be the best for DVCS, they are designed for text files where merge makes sense and file size (and therefore history size) is not huge. After all, they do waste disk (you need the repository and the working copy).
I agree large files may pose a problem, especially if one of the machines is "memory limited".
Adding files is very much automatic (GUI frontends handle that, e.g. qct or tortoisehq for Mercurial).
My point was, if s/he is already using SVN, then Mercurial (or Git, or ...) would be a drop in replacement which more than solves all problems BETTER than Unison and SVN combined.
So why bother? Why not use distributed VCS?
They are quite easy (to learn), they handle merges apparently much better, they give version history, rollback, branches, etc. After all, the OP was using SVN.
How does it handle merges? That is, situations where the data has been modified in both (or several) machines.
You see, there are distributed version control systems (Mercurial, ...) which are designed to handle that kind of situations. And they handle them extremely well (at least Mercurial does and I have no reason to believe the others do not).
Better yet, use Mercurial (or any other distributed version control system). It is designed for this kind of use!
I have three clones (of a project), one in desktop, one in laptop and one in an USB stick. Just "pull" (or "push") and "commit". Sometimes a "merge". With network connectivity you do not need the USB stick (in my case it is just a bit easier).
Should any of those clones/repositories die it would not be a big problem.
Note: this is not a backup solution, this is a solution to sync the machines!
A mirror on an ICBM needs to last maybe one blast, so it does need replacement nor cooling.
95% reflectivity seems to achievable relatively easily. A megawatt * 5% = less than it has to endure when going down, I'd assume.
I would try it.
Firefox on Ubuntu (8.04) is horribly slow. This page (Slashdot) has occasionally almost half a second delays while scrolling down. Or while I am editing here, "end" key takes sometimes 200-300ms.
Machine I am running is 5050e with 4G memory ...
The laser *ITSELF* has two mirrors.
Why doesn't it burn?
So what?
One terabyte HD costs next to nothing, why on earth should I not rip my CDs to FLAC?
Are you sure none of my music in no circumstances I listen does exhibit hearable loss when encoded to XXXbit/s MP3?
I originally encoded my CD's to 128kbit/s OGG (that was beta 1.0 of the encoder), I had to re-rip everything because one song in one place was really horrible. Sure, perhaps final 1.0 with 190kbit/s or whatnot that particular point might have been OK, but why bother?
I see unbelievable buzzword soup.
Starting to heal???
I'm a Finn. I've been waiting that for more than I've been living (small exaggeration is needed in this case), which makes half a century.
Let me tell you a story. I once went to Soviet Union and got out of Russia. The people in there, when I told the historic event I heard from radio, said "nothing is going to change".
And, boy, were they right!
Then you cannot watch the movie until you have
1. watched all the advertisements
2. passed level one of the game.
Have a nice, romantic evening!
No. Distributing virus information is illegal in Finland (where "virus" is "program or part of it which causes harm to computers or data networks").
Sorry for offtopic ...
How about, er, a microkernel?
It loses less than 6% ...
Godwin was so right.
After reference to nazi the discussion will go so bad that no sensible argument is ever going to appear.
(If you keep on going to call some enviro-nazis, aren't the other group those-who-did-not-deny-raping-and-killing-a-girl-in-1990-antienvironmentalists?)
Did the OP claim "getting something for money increases incentive to produce it" or the opposite?
Think about it. Think what is bullshit and what is not.
Excuse me?
My point was entirely that GP changed from logic[1] to bullshit[2] from paragraph to other, just because both use same SW.
[1] getting something for free decreases market value
[2] getting something for money increases incentive to produce it
Eh?
Both are illegal activities which are now relatively safe and easy due to p2p.
The fishy part is whether you did notice the error in your logic or not.
I care. Visually impaired != blind.
I think government should force companies to address the needs of disabled as long as it is reasonable. For example there is no point allowing programs to use colours which are impossible for colour blind to distinguish thus making the use of the program unnecessarily difficult.
Whether in this case the modifications needed are substantial or not is different matter to which I have no opinion.