They try to make it happen. But admittedly it doesn't reliably work, yet.
Most of the autodetection issues are really about proper scripting and driver availability. When you bought the cheap USB winprinter then cups has no chance to pick it up automatically. But if you bought a printer that is supported by linux (or should I say supports linux?) there is no reason why it shouldn't be plug&play.
If recent distros still don't get this right (I haven't checked so I can't tell) then someone failed to pull his head outta his...
I can live with all that, if only they would decide on one thing and stick with it./etc is great, I can expect to find the config- and init stuff there. Some distros put it into subdirectories or make it a symlink-festival but usally blind-shots like vim/etc/ht^T^H^Hapa^T/htt^T are quite successful.
But the whole/usr/{share,local},/opt (glad my debian distro seems not to have that one) organization is a mess.
All distros use a slightly different layout (esp. when X11 comes into play) and I can hardly ever predict where my stuff goes when I type make install.
Usually it's/usr/local/{bin,man,lib} but sometimes it differs.
I'd really like if the autoconf tools (configure and friends) would get a small extension that just displays (or better yet, saves to a central location) the locations of all files that were just installed. Just so that I don't have to dive into a Makefile (or 90 pages of./configure --help) to figure it out.
The distro-specific mess is addressed by package managers. But the real mess starts when you mix it with hand-compiled apps.
What else do you need to control other than read, write, and execute? If you have to ask, you don't get it.
ACLs are a two-edged sword. Usually (unless you're the NSA) you will not want to deal with the immense complexity they introduce to your setup. Old fashioned rwx has proven to be a good balance between complexity and granularity.
Creating different groups with different priveleges is simple in Linux and will give you much finer grained control than the standard 2 or 3 groups Windows gives you. No, it won't. Not only that, it *can't*. Unix's file permissions and groups are a *subset* of NT's ACLs. In other words, anything Unix file permissions can do, NT's ACLs can do - but not vice versa.
When you have time take another look at make xconfig (or make menuconfig). ACLs are available for you. Setting them up is a bit of effort but if you really need ACLs you'd better know what you're doing in first place anyways.
Rule of thumb for computer UI-design: If you have an idea how to make it good, do it. If you have no clue, copy MacOS (maybe take a short look at NeXT, Be or amiga for inspiration).
Why someone would want to copy the windows UI is beyond me. I mean the "start"-menu and taskbar approach is common sense now and there really is not much more worth copying in the win UI.
Linux will be fine when all important apps are available for linux.
No reason to run windows apps. Or mac apps. Or some old atari apps. Or stuff that was written for OS/2.
The cool thing about this whole open source idea is that Linux can get its own applications. The whole palette. They just pop up once and then they keep improving over the years and never go away.
A few of them are already better than their counterparts in the windows world today. Give it some more years and we might have gimp with a usable GUI (byebye photoshop), average hardware that can launch openoffice in under 30secs (byebye ms office), gnucash will have learned the whole business shebang (byebye quicken and friends), kde will be a flashy blinking monstrosity that can look like the windows longhorn UI, mac os XI and palmOS at the same time (via 3d shutter glasses) (byebye wintendo-GU*oops just entered my homebanking pin into icq*I), konqueror will read your mind, figure out what you want to do next, ignore it and do the right thing.
We really need no stinkin' win32 binaries or emulate a legacy GUI-framework. We are in the progress of reinventing the wheel once more. When that's done we'll just use and improve it and not care about software that we have no control over anymore.
For the nitpickers: Yes, reinventing a wheel involves building many competing prototypes until the optimal design (the one that survives the process) was found.
Well, I haven't looked at konqueror specifically to find out how it behaves when doing FTP.
Usually webbrowsers will open a new connection every time you click on a FTP link. It will receive the directory list or file that you requested and close the conn. Next click, next connection. And so on.
I have not yet seen a browser that was smart enough to keep the FTP connection open until it hits a timeout or the user clicks elsewhere. All browsers I tried happily went through the hammering-game with all the problems involved, like temporary site-bans for "too fast clicking" and overall high latency.
Umm, but to get to the point, I don't know how konq behaves or what you could tweak to make it better. Yes, disabling previews/content introspection would be wise. But konq should really be smart enough to do that by default when talking FTP. If it doesn't: Ouch! FTP ain't webdav nor samba...
If you do a lot of FTP shuffling I'd recommend using a real client for your own comfort (much faster/reliable).
But in the end it doesn't matter. If you prefer konq, use konq! FTP is a dying protocol anyways and exactly for these reasons...
It also beats shit out of the ftp sites you're using. FTP is a broken protocol with even more broken implementations. The worst of these implementations usually are browsers.
I have one question: WHY? (no offense!) I'm really curious what the motivation behind these projects is.
Are you in it mainly for the learning expirience or what is it that keeps you working on this?
I mean you guys spend a lot of time duplicating a lot of work that others already went through. Your website looks nice (didn't read it all, yet but will do some more reading) and the screenshots look good, too (actually quite impressive considering it's all built from scratch).
But still. Personally I'd prefer to see you talented guys working on improving the more "mainstream" GUI components (X, gtk/kde whatever) or probably inventing completely new goodness to the linux desktop.
Looking at what you came up with (syllable) tells me you could do a lot of good in that area.
I guess you've been asked this a trillion times, but I didn't find it in your FAQ, so I'm asking again...
Would you want it to try and mess with your config files? I prefer to do that part myself instead of having a script guessing what belongs where. (bound to fail)
Well, it deals with compiling and debugging just well for its size. Unless you're talking Mozilla...
Ofcourse it would be nice if the device had a 3.0GHZ P4 instead of the 0.4GHZ XScale CPU. Also I'd prefer 1GB ram over the 64MB that's included. But the fact alone that you can stick a small, yet full featured (real keyboard!) linux box into your pocket (baggy) makes it attractive, doesn't it? I really hardly ever do any coding-work when on the road. And if I wanted to I think the main limiting factor would still be the form factor (small screen) long before RAM/CPU become an issue...
You must see it to appreciate howsmall it really is...
I know I sound like an ad. Disclaimer: not affiliated with anyone
People who prefer MM/DD/YYYY are also the sort of people that prefer the Imperial System to the Metric System. IME, they also tend not to be programmers (for reasons that should be obvious). It's demonstrably inferior to both systems (both for digital manipulation and for manual parsing) and this is why it restricted to common use only in North America. Like the Imperial System, it's an anachronism best left in the past where it belongs (though, also like the Imperial System, it appears that people will continue to use it for quite some time, no matter how obviously inferior it is to avalible alternatives).
Umm.. Well, you are right and I agree with you on all points. But still I see MM/DD/YYYY every day. Look, only a few posts down this guy used it.
If you want to engage the scriptkiddy-community for your noble goal you'll have to point them to the playground.
I imagine a central website providing (primarily) a list of ip addresses and good reasoning (at least 2000 lines) + verification for every single of them.
Every ip should be backed up by a copy of at least 100 distinct complaints from different sources.
But ofcourse this conflicts with both law and common sense so don't even think about it. I did and trust me, it's a bad idea.
Slashdot Mirror
HAHA! :-)
Exactly what I thought when I read his post!
They try to make it happen.
But admittedly it doesn't reliably work, yet.
Most of the autodetection issues are really about proper scripting and driver availability.
When you bought the cheap USB winprinter then cups has no chance to pick it up automatically.
But if you bought a printer that is supported by linux (or should I say supports linux?) there is no reason why it shouldn't be plug&play.
If recent distros still don't get this right (I haven't checked so I can't tell) then someone failed to pull his head outta his...
I can live with all that, if only they would decide on one thing and stick with it. /etc is great, I can expect to find the config- and init stuff there. Some distros put it into subdirectories or make it a symlink-festival but usally blind-shots like vim /etc/ht^T^H^Hapa^T/htt^T are quite successful.
/usr/{share,local}, /opt (glad my debian distro seems not to have that one) organization is a mess.
/usr/local/{bin,man,lib} but sometimes it differs.
./configure --help) to figure it out.
But the whole
All distros use a slightly different layout (esp. when X11 comes into play) and I can hardly ever predict where my stuff goes when I type make install.
Usually it's
I'd really like if the autoconf tools (configure and friends) would get a small extension that just displays (or better yet, saves to a central location) the locations of all files that were just installed. Just so that I don't have to dive into a Makefile (or 90 pages of
The distro-specific mess is addressed by package managers. But the real mess starts when you mix it with hand-compiled apps.
You'd probably like ion.
It's worth a shot, I used it for about 6 months but just recently switched back to a more "traditional" wm.
What else do you need to control other than read, write, and execute?
If you have to ask, you don't get it.
ACLs are a two-edged sword.
Usually (unless you're the NSA) you will not want to deal with the immense complexity they introduce to your setup.
Old fashioned rwx has proven to be a good balance between complexity and granularity.
Creating different groups with different priveleges is simple in Linux and will give you much finer grained control than the standard 2 or 3 groups Windows gives you.
No, it won't. Not only that, it *can't*. Unix's file permissions and groups are a *subset* of NT's ACLs. In other words, anything Unix file permissions can do, NT's ACLs can do - but not vice versa.
When you have time take another look at make xconfig (or make menuconfig).
ACLs are available for you. Setting them up is a bit of effort but if you really need ACLs you'd better know what you're doing in first place anyways.
Thank god.
Rule of thumb for computer UI-design:
If you have an idea how to make it good, do it.
If you have no clue, copy MacOS (maybe take a short look at NeXT, Be or amiga for inspiration).
Why someone would want to copy the windows UI is beyond me. I mean the "start"-menu and taskbar approach is common sense now and there really is not much more worth copying in the win UI.
The lack of printer drivers?
Linux will be fine when all important apps are available for linux.
No reason to run windows apps. Or mac apps. Or some old atari apps. Or stuff that was written for OS/2.
The cool thing about this whole open source idea is that Linux can get its own applications. The whole palette. They just pop up once and then they keep improving over the years and never go away.
A few of them are already better than their counterparts in the windows world today.
Give it some more years and we might have gimp with a usable GUI (byebye photoshop), average hardware that can launch openoffice in under 30secs (byebye ms office), gnucash will have learned the whole business shebang (byebye quicken and friends), kde will be a flashy blinking monstrosity that can look like the windows longhorn UI, mac os XI and palmOS at the same time (via 3d shutter glasses) (byebye wintendo-GU*oops just entered my homebanking pin into icq*I), konqueror will read your mind, figure out what you want to do next, ignore it and do the right thing.
We really need no stinkin' win32 binaries or emulate a legacy GUI-framework. We are in the progress of reinventing the wheel once more. When that's done we'll just use and improve it and not care about software that we have no control over anymore.
For the nitpickers: Yes, reinventing a wheel involves building many competing prototypes until the optimal design (the one that survives the process) was found.
Well, I haven't looked at konqueror specifically to find out how it behaves when doing FTP.
Usually webbrowsers will open a new connection every time you click on a FTP link. It will receive the directory list or file that you requested and close the conn. Next click, next connection. And so on.
I have not yet seen a browser that was smart enough to keep the FTP connection open until it hits a timeout or the user clicks elsewhere. All browsers I tried happily went through the hammering-game with all the problems involved, like temporary site-bans for "too fast clicking" and overall high latency.
Umm, but to get to the point, I don't know how konq behaves or what you could tweak to make it better. Yes, disabling previews/content introspection would be wise. But konq should really be smart enough to do that by default when talking FTP.
If it doesn't: Ouch!
FTP ain't webdav nor samba...
If you do a lot of FTP shuffling I'd recommend using a real client for your own comfort (much faster/reliable).
But in the end it doesn't matter. If you prefer konq, use konq! FTP is a dying protocol anyways and exactly for these reasons...
Or over local maxima.
To the big white phone.
Argh.
For a moment I almost forgot why flash had such a bad reputation. Thanks for the reminder TDR.
(macromedia should sue you)
It also beats shit out of the ftp sites you're using. FTP is a broken protocol with even more broken implementations. The worst of these implementations usually are browsers.
I have one question: WHY? (no offense!)
I'm really curious what the motivation behind these projects is.
Are you in it mainly for the learning expirience or what is it that keeps you working on this?
I mean you guys spend a lot of time duplicating a lot of work that others already went through.
Your website looks nice (didn't read it all, yet but will do some more reading) and the screenshots look good, too (actually quite impressive considering it's all built from scratch).
But still. Personally I'd prefer to see you talented guys working on improving the more "mainstream" GUI components (X, gtk/kde whatever)
or probably inventing completely new goodness to the linux desktop.
Looking at what you came up with (syllable) tells me you could do a lot of good in that area.
I guess you've been asked this a trillion times, but I didn't find it in your FAQ, so I'm asking again...
Would you want it to try and mess with your config files? I prefer to do that part myself instead of having a script guessing what belongs where. (bound to fail)
OH MY GOD!
They suspended goatse.cx!!!
Run for your tinfoil hats, the world is coming to an end..!
Well, it deals with compiling and debugging just well for its size. Unless you're talking Mozilla...
Ofcourse it would be nice if the device had a 3.0GHZ P4 instead of the 0.4GHZ XScale CPU.
Also I'd prefer 1GB ram over the 64MB that's included. But the fact alone that you can stick a small,
yet full featured (real keyboard!) linux box into your pocket (baggy) makes it attractive, doesn't it?
I really hardly ever do any coding-work when on the road. And if I wanted to I think the main limiting factor would still be the form factor (small screen) long before RAM/CPU become an issue...
You must see it to appreciate how small it really is...
I know I sound like an ad.
Disclaimer: not affiliated with anyone
Check this out.
Good enough for posting to slashdot from the backseat of a car (over GSM)!
Shameless plug.
People who prefer MM/DD/YYYY are also the sort of people that prefer the Imperial System to the Metric System. IME, they also tend not to be programmers (for reasons that should be obvious). It's demonstrably inferior to both systems (both for digital manipulation and for manual parsing) and this is why it restricted to common use only in North America. Like the Imperial System, it's an anachronism best left in the past where it belongs (though, also like the Imperial System, it appears that people will continue to use it for quite some time, no matter how obviously inferior it is to avalible alternatives).
Umm..
Well, you are right and I agree with you on all points. But still I see MM/DD/YYYY every day.
Look, only a few posts down this guy used it.
The insane amount of dollars that eBay runs into the pockets of Meg Whitman and friends is a reputation issue of its own to me.
That's why I never sell stuff on eBay.
I refuse to pay a Meg Whitman tax.
That's already fixed via DMCA.
They're going to print an expiration date on every piece and pass a law that makes it illegal to use their intellectal blockery after expiration.
*shrug*
i don't get it, can't you just paste the damn thing?
If you want to engage the scriptkiddy-community for your noble goal you'll have to point them to the playground.
I imagine a central website providing (primarily) a list of ip addresses and good reasoning (at least 2000 lines) + verification for every single of them.
Every ip should be backed up by a copy of at least 100 distinct complaints from different sources.
But ofcourse this conflicts with both law and common sense so don't even think about it.
I did and trust me, it's a bad idea.
It's called grass root marketing.
...broadband rents YOU!