Good points but I guess it all depends on whether the stuff you're doing makes you a criminal in the eyes of the gov., whether you think a competitor has some control over a portion of the gov., whether IBM wants to buy you
It's not even a question of doing something illegal or fearing some nutcase government conspiracy--individuals within the government can use their priviledged access against you. Witness yesterday's revelation that a pair of FBI agents may have been (they haven't stood trial yet) using FBI data to manipulate the stock markets. Or the current case involving an agent accused of feeding information to the mob.
As soon as individuals have the information, it's only as secure as they're willing to make it. And that was my point. As I said, if you're protecting $100 secrets and you think it costs $10,000 to corrupt the agents who could get at your data, that's probably fine (unless, of course, your secretary also has a copy of the key and can be bribed for $20....) But you need to understand those tradeoffs.
However using open source does mean you can be sure it really is well written, unless someone throws a grid computer at it...
_if_ it's widespread enough and mature enough that the community has had time and inclination to thoroughly examine it. I don't think you want to trust an algorithm or a complex system just because you've examined it yourself and think it's secure, even if you're Ron Rivest or Bruce Schneier. But components like PGP, ssh, and SSL have been examined pretty well and you can expect to have a good idea of how secure they're considered by the security community; the most popular open-source implementations have been examined closely as well, (even they still turn up minor problems from time to time).
Of course, whatever you choose you have to implement it correctly.
You have the right idea, but your technology is about a decade out of date.
You can get smart cards/crypto cards or dongles that keep the keys on the card. It's *never* revealed, and all crypto is done on the card or dongle.
Until someone reads the key back off the card, which has been done with cheap ($50) hardware for "tamper-proof, unreadable" smart cards.
The real key is the human factors: don't let anyone get their hands on the key who you don't trust to keep it reasonably secure, and make sure that people understand how important it is.
Also, remember that the current setup is only of marginal security; don't spend millions of dollars to make the digital signatures bulletproof when someone can either forge a paper signature, fast-talk the employees into authorizing a purchase/signature/whatever, or otherwise circumvent the crypto. The free tools (gpg/OpenPGP or openssl based S/MIME solutions), if properly implemented, are more than sufficient to force would-be imposters back to their old tried and true ways.
In other words, worry more about how you implement the system (eg: key management) and how you train the users than about the particular cryptosystems and techie gadgets you choose.
_If_ you trust Lotus/IBM, who have been known to put backdoors into their export versions to give the NSA substantial portions of the encryption keys. And their code hasn't been vetted by the security community. And your actual level of security is highly dependent on your key management scheme.
Not to say it's a bad idea, but "Notes is very secure" is a pretty glib answer.
Start->Settings->Control Panel->Tweak UI->Mouse Tab->Check "Activation Follows Mouse (X-Mouse)". Semi-supported by microsoft, should work with 99.9% of all windows and apps. Hardly a "hack".
It's worthless with most apps. As soon as you click in them to get keyboard focus in the right place, the window is raised. There are some apps that it's good enough with. Definitely a hack for a system that for whatever unknown reason wasn't designed to differentiate well between "raise a window" and "give a window focus".
Click and drag the top of the taskbar down to the bottom. It disappears. Or turn on auto-hide."Clippy"
He's dead. Why kick a dead horse?Windows window management has come of age. Looks like you dumped it back in '95 and never looked back
Actually, I dumped it in '94. But my current job has me using XP, so I've been full-time in front of that for 6 months now. And previous jobs sometimes had me full-time in NT 4.0 for periods of months (I've used TweakUI for several years now).
Its window management sucks. It still has this legacy thought that the user is only running one, or at most a few, apps at once (witness the amount of screen real estate you'd need to have the taskbar effectively manage 20-30 apps). And that only one of those apps will be used at a time (witness the impossibility of looking at one thing and typing into another). And that the apps should override the user wrt window management (witness pop-unders, which are impossible in my window manager under X because only the user or window manager can lower a window, not to mention that even without those user overrides window.blur() will loose focus but not lower a window in my window manager...) Gagaga.
The window management in Windows is better than anything I've seen in Linux.
I'll agree that translucence and themability are fluff. I might be able to envision an actual use for translucence if I thought about it long enough, but it'd be a real corner case.
But...
Things that suck in Windows window management:
Click to focus + focus autoraise. The latter is the biggest problem. I can't tell you how often I want to be typing into the window _behind_ another window, so I can see the contents of another window while I type. And click to focus is just annoying, why put another step in the way of my work? There are some hacks to get focus-follows-mouse, but a lot of apps don't work well with it.
The task bar. This thing just blows, it's the first thing I turn off in Gnome/KDE. At most I want a couple of launch buttons and a clock, but I _don't_ need the entire bottom half of the screen real estate taken up by icons of every running app, and the thing is only usable if I have at most 8-10 windows open. Usually I have 4-5 times that. Give me alt-tab, windowshade, window groups (and raise/iconify/etc working on entire groups), virtual desktops, and restricted alt-tabs (meta-tab limited to xterm, control-tab limited to mozilla, etc) over that any day. In other words, real tools for managing the windows (which is what I want out of a window manager). Sawfish lets me do that. The groups, especially, are a godsend. Launch an editor, debugger, and GUI designer all in one group, then operate on that group as a whole when I need to. Which leads to...
...MDI or whatever it's called when the IDE/Word/whatever opens a bunch of subwindows inside its own window instead of just opening them as real windows. God this sucks. I already have a window manager, I don't want every application to _also_ have a window manager. Of course, if your IDE takes the approach of putting everything into one window rather than seperate windows which can be grouped together then you need something like this. Ugh.
Clippy. Yeah, he's not related to window management but even now that he's dead he deserves to be kicked around.
Despite what you see on TV and read in the papers, the economic meltdown continues - the media is masking the severity of this downturn. If you want to see for yourself how bad things really are, take a trip through the US and note the closures of many businesses, the 'downsizing', and actually talk to people.
That's not a good way to get a clear picture of the economy. Employment is a lagging factor in economic recessions: things are bad for months before layoffs begin, and things are good for months before layoffs stop and hiring begins.
So if you go around looking at downsizing and talking to people who have lost their jobs, you'll get a picture of what the economy was like several months earlier.
It's also worth noting that unemployment in the US is currently at 6%. Not nearly as good as it was 2 years ago, but worth remembering that until then unemployment hadn't been below 5% since 1970 and the Fed was worried that an unemployment rate below 6% would spark inflation similar to that seen in the early 70s, decimating the economy. The current "horrible unemployment rate" was thought of just 7 years ago as "unsustainably high levels of employment that risk pushing the economy into a depression"--obviously alarmist, but the point is that the job situation is still better than it was in 1995, or in the economic boom of the 1980s.
Sure, I know a lot of people who lost their jobs--myself included. The company (a dot-com) I worked at went from 450 to 50 employees. But I don't know anyone who lost their job and didn't find another one at equal or higher pay within 3 months--and that's in the (comparitively) hard-hit software industry. Indeed, without 2 quarters of economic contraction or substantial unemployment, many people don't think we're in a recession at all. It's certainly a much milder one than we've had in my lifetime.
Some people call this "liberal" or "liberal arts" education. It's funny because the US is the only country I know of that follows "the great European tradition of liberal education." I don't know of too many foreign universities with a "Great Books" program. French and German engineers don't have to take political philosophy courses in college. In fact, they probably won't take any polisci in the last couple years of high school.
Now, you try to get into a math grad school in the US with your BS and good grades, and you'll have a good chance; but why? Because schools can only take so many Hungarian and Romanian mathematicians who've actually been doing mathematics since high school.
Not really true in my experience. I went to public school (high school) in the U.S. and certainly had my share of mathematics--we finished AP Calculus (AB and BC) and some linear algebra in junior year and did differential equations and more linear algebra senior year. Of course, not everyone opts for the accelerated track which leads to...
The real difference between school in Europe vs. stateside is that in Europe there's no stigma associated with going to a trade school instead of university. Here in the US, trade schools are viewed as an option for the uneducated and so a lot of people who don't have any need or desire for a university education wind up in university anyway, which means that universities have a high percentage of courses covering things that really ought to be covered in high school. But the top US students, those who would go on to university in a European system, tend to have just as good an education. In my experience, anyway. In other words, the problem isn't that US high schools are that bad, it's that too many people are going to college who might be better served with European-style trade educations.
Or to put it delicately, most colleges have two tracks; one is those who are in college because it's the thing to do socially, and one is those who actually want/need a college degree. But the second group is about the same size (slightly larger per capita) as the European university group and in my estimation equally well-prepared.
The reason you stand a good shot at getting into grad school is that the tremendous number of college graduates in the US mostly don't apply to grad school. Only the top few do--there isn't the same social pressure to get a postgraduate degree. Indeed, about the same number of US students apply to grad school each year as the totals for France, Germany, and the UK, which works out to 2/3 as many per capita and less than 1/3 as many per college graduate. Those that apply are among the best prepared and the best qualified and matriculation rates from doctorate programs are actually higher for the Americans than the Europeans (even when grouped by university, though I can only find those numbers for U.S. doctoral programs and not for European or masters programs).
Handful of facts from the NSF's last study on higher education: U.S. university attendance has skyrocketed over the last 50 years, from 2.5 million enrolled to over 14 million enrolled. 35 out of every 100 college age individuals in the US are enrolled in college, as compared to 24 per 100 in Germany and 13 per 100 in France. The UK, Canada, and Australia are the major exceptions (UK sends 35/100).
I very much agree with you about spending time with family, but speaking from experience, freelance is not the way to go. I and most of my colleagues work freelance/contract and over the past few years have spent almost every major holiday with last-minute crunches destroying our family time. And with freelance, as opposed to a "job" job, you need to do the time whenever it's required, or you don't work anymore.
The worst part is that no matter how hard you try to be available to go out with the kids on Hallowe'en, you'll get a life-or-death call from a client that afternoon which you have to take seriously.
That depends heavily on the sort of freelance work you do. When I do consulting, I _never_ allow any sort of on-call option. I definitely never let them have my cell phone or home phone number. I also do absolutely zero system admin or web development work, I stick strictly to software development (whether new products, new features, project rescue for failing projects, platform ports, cleanup, optimization, whatever...) with written requirements and I make sure that the contract backs me when it comes to saying "changing requirements==changing time estimates".
This is not to say I'm inflexible, changing requirements is a good thing as you work with the client to understand their needs (and often what they really want bears little resemblance to what they say they want, and they'll usually be very pleased if you can actually nail down something that helps them). Just that as you get a clearer view of what needs to be done, you also get a clearer view of how long it'll take.
I'd say 90% of non-web software development work (as opposed to admin work) has absolutely no reason to call the developer in late nights or weekends. You might need to pull some weekends if you're way off in your time estimate, just to meet deadline. But do that a couple times and you start to get good at providing better estimates. And it's usually way more interesting than web dev.
Good call on the single-cluster angle, but the folks at Alesis [alesis.com] already beat you to it.
It's not an issue of getting beat to it, it's a really old idea. Heck, ReiserFS was hardly the first and they've done it for over 5 years. ext2 hasn't, but there is some movement in that direction.
Every key name/value pair--being a file--will take up one cluster, regardless of how much data it actually contains
Not if your filesystem has decent support for small files, like e.g. ReiserFS or some of the recent ext2 devel branches. This is one thing ReiserFS definitely got right (though I'm less sure about the direction they're taking for ReiserFSv4).
As I said, it's a workable approach for grafting extensible metadata onto an existing FS, but it's not how you would design an ideal from-scratch FS.
It's definitely how I would design an ideal from-scratch FS. Metadata in the FS is a really horrible hack that goes against all the namespace unification efforts that OS designers have been making in the last 10-15 years.
Bear in mind that one of the core driving philosophies that has allowed Unix to survive and evolve is "everything's a file", where a file is a stream of bytes. Attaching metadata directly to a file gets to be a nightmare when you start copying files over the network, dealing with editors that create new files rather than modify in place, backing things up, etc.
NeXT had a great solution to this: instead of just having a file "mysong.mp3", have a directory "mysong" that includes files like data.mp3, title.txt, etc. The file browser presented this to the user in a simple UI (e.g. the directory would have the "mp3" icon in the file browser, double-clicking it would play the song in the mp3 player, displaying the title.txt, etc). But traditional file transfer and backup tools worked just fine, and you could use standard tools (your regular editor, copy program, etc) instead of crufty special purpose ones (regedit, or binhex, or whatever) to work with the metadata.
Think of it this way: you don't want to just have 2-3 defined metadata fields, you would like to be able to add arbitrary new ones in a flexible way. And you want the contents of each to be flexible, without size limits or such. So you want a way of associating names with values--wait, we have that, it's called a filesystem. And you want a way of grouping a bunch of name-value pairs together--wait, we have that, it's called a directory. You'll either end up reimplementing a filesystem, badly, as your metadata, but need special tools to view/edit the metadata or move it around (a la resource forks or the registry), or you'll end up going with a NeXT-like solution and use the filesystem itself (which deals quite nicely with things like this).
GPCL = GTK Pattern Creation Lab GTK = Gimp Toolkit (original expansion) GIMP = GNU Image Manipulation Program GNU = GNU's not Unix
Re:GNOME vs KDE for the newbie
on
GNOME 2.0 Beta
·
· Score: 5, Funny
you also lose the one thing that makes gnome slower than tar
Actually, tar is pretty fast--it's bzip2 that makes it seem slow. Try gzip or lzop instead, or don't compress if you are storing compressed files--though maybe cpio is somewhat faster than tar.
I didn't notice anything about online backups, point-in-time recovery, or standby database
Online backups: yes, for quite some time
point-in-time recovery: postgres uses WAL undo/redo logging, but I'm not sure what the state of rollback tools is at the moment.
Standby database: Assuming you mean Master/Slave replication, this is one of the major features planned for 7.3; 7.x has added a lot of the infrastructure needed for replication, and by 7.4 they hope to have multimaster replication (ie a fully distributed database).
Worth keeping in mind that you can form a corporation for less than $300, which still excludes the poor but it's not just the super-rich that can exploit corporate law.
Qt 3.0 and later supports XRender for alpha transparency, etc. Requires XFree86 4.1 or later. KDE 3.0 supports this (in CVS), using it for transparent menus and other stuff. Not released yet, and added 29 December (which was after the first KDE 3.0 Beta release, I think). The transparency is hardware accelerated on supported platforms (including lots of them, e.g. nvidia, ATI rage128, Matrox, etc)
Speed isnt as important as you think when you have a fast computer, frame buffer is slower? SLOWER? Its frame buffer, if you have nvidia and a pentium 4 is it slower?
For anything where speed really matters (e.g. video playback or other Xv stuff, 3D) yes, fb is definitely slower. It's probably slower for "normal" 2D stuff as well, that's true of most 2D cards but I don't have an nvidia so I can't bench it.
This is the worst kind of offtopic nitpick, but I have to point out that Vi does run on a dumb terminal.
Y'know, I originally had written "you can't even run vi on a true dump terminal except in open mode", but I figured people would quibble over what a dumb terminal was and so changed to the glass tty aside. Should've left the open mode comment in too, I guess.:-)
Sumner
Re:MS Windows vs. X, same hardware
on
Xfree86 4.2.0 Out
·
· Score: 2
While I agree that the toolkits suck, that email is rather irrelevent. If I interpret the email correctly, backing is referring to double-buffering.
No, it's referring to duplicate work being done by the toolkit. e.g. X sends notices saying "the lower-left needs a redraw", "the mid-left needs a redraw", and "the top-left needs a redraw. A properly written app would coalesce those into a single call to redraw the left side (ie clear the event queue of all exposure events before repainting), rather than repaint several times. gtk 2 takes some steps toward this.
Even as it is, cards with good driver support don't show the flicker problem. But well-written toolkit code would do the coalesce and work fine even on your old Cirrus Logic without flicker (or the new whiz-bang card with crappy unaccelerated drivers).
and hell, Photon is even more flexible than X. It even runs the graphics driver as a seperate process
How does this make it more flexible than X? X keeps the driver as a loadable object file, and puts the high-level GUI code (from Xlib on up) into a seperate address space from the X server. If you like, you can run a framebuffer video driver and fb X server, putting the low-level hardware support in a seperate address space. None of these really impacts the flexibility in any great way, though, the X extension mechanism is way more crucial to flexibility than which address space various components happen to run in.
QNX does rock, though. Between it and L4, I'm almost tempted to forget the travesty that is Mach. How's the swapper doing?:-O
Check the docu, you need to have it compiled into your X server and might need an Option line in XF86Config-4. If it's on, you'll get:
(==) R128(0): Silken mouse enabled
or similar in the XFree86 output.
If you're running a kernel with good latency (e.g. 2.4.17 + Andrew Morton's LL patches, I use 2.4.17-jl11) you'll no longer see any mouse cursor dragging/skipping/etc problems. If the kernel has bad latency, there's nothing X can do about it.
Sumner
Re:MS Windows vs. X, same hardware
on
Xfree86 4.2.0 Out
·
· Score: 5, Interesting
I think part of the problem is the fact that there never seems to have been any coherent work done on this. The windowing system oriented people who work on X say "the toolkit authors fault". The toolkit authors would say "it's your drivers or the limitations of X Windows"
Nope, read the thread I quoted and you'll see that gtk developer Owen Taylor agrees and that gtk 2.0 includes some of the optimization mentioned. The toolkit and X11 authors do work together on these things, and the toolkit authors have had a huge amount of input into the design of the XRender extension and the DRI infrastructure.
While I do appreciate the flexibility of X Windows, I honestly DON'T think the windowing system and toolkits should be these totally orthogonal projects, and the toolkits just "draw as they see fit" on a canvas that they expect the windowing system to render dumbly. This is the X model, inherited from the dumb terminal days.
Actually that's not the X model (BTW, X wouldn't run on a dumb terminal--even vi wouldn't run on a true dumb terminal (ie glass tty)). The X model is to provide high-level graphics primitives to the application, which then submits them to the server which can turn them into whichever low-level calls are most efficient on the hardware in question. Not only that, but the library used to submit those request can (and does) batch them together so that the application writer can have a simpler model and still get efficient code--for instance, multiple XDrawLine calls are batched by XLib into a single XDrawLines call that's sent on to the server, saving on round-trips and in some cases saving on bus traffic to the video card by eliminating redundant traffic. Or servicing those high-level requests in whatever manner is most efficient for the hardware in use.
Highly efficient graphics can be done this way. Witness SGI, who were for years the undisputed leaders in the graphics field. They used X11.
But think of X as being more of a device-driver with a unified API, the GUI is to be built on top of that. It's a highly reasonable and well considered model that is ideal for building the high-performance GUIS of the future on. Far better than e.g. a framebuffer, which is already obsolete (doesn't handle many 2D features like overlays & alpha blending, doesn't do 3D acceleration, doesn't allow for hardware security a la SGI, doesn't handle hardware video decoding, etc) and is low-level enough that you can't have the driver do intelligent optimizations without rewriting the apps. And designed with the foresight to be extremely flexible.
Sumner
Re:MS Windows vs. X, same hardware
on
Xfree86 4.2.0 Out
·
· Score: 5, Informative
There are VERY obvious performance differences between any version of Windows and as new of version of X as you want. X Windows programs flicker like mad when moving or resizing, objects aren't responsive
If you had read the thread I mentioned in the article you replied to, you'd see the anser to this one:
> > Not to be too non-technical...
> > > > If the protocol overhead is so small, why can't my 1200 mips (600mhzPIII)
> > machine resize windows without widgets streaking? My 486 could do
> > this fine running MS Windows. Is this because many widget toolkits (GTK,
> > QT) use XPutImage? There must be some way to speed things up.
blame your widget set. basically (sorry owen and co on the list) gtk a
(and i presume qt) dont render optimially at all. the do a semi-decent
job.. EXCEPt for opaque resizing, and when redrawing is more than a few
lines and boxes... this is a toolkit issue and imho the current set of
toolkits (motif, qt, gtk etc.) do a god-awful job of this kind of
stuff. right now i have silky smooth "opaque resize" stuff working here
with enlightenment 17 - but i do the rendering completely differently
to gtk/qt - its all a canvas and thus the rendering happens in a
"backing" so updates are smooth. on todays hardware this is the best
way to do it and have almots no artifacts ANd retain speed.
> "Streaking"? Are these opaque resizes? Alot of apps aren't doing
> event compression. They repaint the whole damn window every time they
> get an event. They could have at least checked that there weren't
> more events in the queue and got rid of them instead of handling each
> one in turn.
true. its a very bad thing that there are a LOT of apps that behave like
this... a LOT. some of the most commonly used are guilty of this
(netscape for one....)
the mouse frame rate is low
If you enable Silken Mouse in XFree86 4.0 and later, this should be fixed. Certainly an implementation issue and not an architectural issue (i.e. not a reason to throw out X and start over)
applications all have inconsistent look and feel, keyboard support is lacking...
These aren't X11 problems but GUI problems, GUI standardization is certainly a huge issue. But, gtk-2.0's accessibility enhancements include excellent keyboard support and some steps toward simplifying and unifying look&feel. KDE is moving in that direction as well. Obviously you need to use a single unified UI on your desktop, but having two decent ones available to choose from is not a bad thing (not to say that either is decent yet, but they're both heading there rapidly).
Sumner
Re:Moving away from X
on
Xfree86 4.2.0 Out
·
· Score: 5, Interesting
X extensions shouldn't be thought of as just being tacked on, they're a good and efficient way of doing things. The whole point was that the rendering engine would be replaced via an extension, this was anticipated and designed for.
In fact, when X was originally developed Jim Gettys et al considered putting _all_ graphics rendering in an extension (leaving just the core windowing w/o rendering in the core). They fully expected the original rendering model to be replaced fairly soon, but that's taken a long time. XRender hopes to do that and probably will largely supplant the old rendering primitives for new apps in a few years. Maybe sooner for gtk/Qt/other whizzbang bleeding-edge stuff.
We toyed with leaving graphics entirely to an extension, but the argument that a window system without any graphics would be useless won out pretty fast:-).
We never thought the existing rendering would last as long as it did: we expected significant extensions would have occurred long since.
--Jim Gettys, 2001
We can't get rid of the core X11 primitives because they are a part of the X11 specification and all apps use it and it isn't going to go away any time soon. Once render is complete and stabilized we can just encourge people to not use the core primitives. Eventually we can care less about making them fast and concentrate on making them unobtrusive.
--Keith Packard, 2001
From the thread Proposal for server-side Anti-Aliased fonts
Sumner
Re:Moving away from X
on
Xfree86 4.2.0 Out
·
· Score: 5, Informative
The problem with X11 is, in part, the separation of client/server; this causes extra latency and a heap of context switches
The context switches aren't a significant overhead. They weren't even a significant overhead in 1986 when Sun first started spreading FUD about this (at the time, Sun was trying to push NeWs over X11). See e.g. Jim Gettys' posts in the "rendering model in X" thread in the Xrender mailing list archives
It's not all sunshine, he's willing to own up to places where X needs improvement (exposure lists are a big one, througput for e.g. texture mapping is another), but it's way better than a lot of people claim. And Xrender and DRI address the vast majority of the problem cases very effectively.
Sumner
Re:Thats nice, but whats with kernel.org?
on
Xfree86 4.2.0 Out
·
· Score: 1
Most services stopped responding, and the administrative port wasn't turned on when they installed the new hardware. H.P Anvin posted to lkml about this. They're working on getting it back up and running.
Including the patches, which are not enabled unless you choose the option while configuring the kernel, wouldn't mess up kernel debugging, would it?
But if a user has preempt enabled and submits a bug report, that bug report becomes very difficult for kernel developers to decipher. Not to mention that hard-to-reproduce bugs may show up under preempt, then you have to build a non-preempt kernel and try for a potentially long time to reproduce the error.
I am not a kernel hacker, and I was not commenting on the code; I feel that the idea of leveraging the SMP locks is elegant.
It would be, if the locking requirements for SMP were sufficient to deal with the locking requirements for preempt. But they're not, these are really two different beasts--see my other message under this topic for a laundry list of reasons and linux-kernel topic to read.
Could you explain what "live locks" are, please? I'm not familiar with the term.
Slashdot Mirror
Good points but I guess it all depends on whether the stuff you're doing makes you a criminal in the eyes of the gov., whether you think a competitor has some control over a portion of the gov., whether IBM wants to buy you
It's not even a question of doing something illegal or fearing some nutcase government conspiracy--individuals within the government can use their priviledged access against you. Witness yesterday's revelation that a pair of FBI agents may have been (they haven't stood trial yet) using FBI data to manipulate the stock markets. Or the current case involving an agent accused of feeding information to the mob.
As soon as individuals have the information, it's only as secure as they're willing to make it. And that was my point. As I said, if you're protecting $100 secrets and you think it costs $10,000 to corrupt the agents who could get at your data, that's probably fine (unless, of course, your secretary also has a copy of the key and can be bribed for $20....) But you need to understand those tradeoffs.
However using open source does mean you can be sure it really is well written, unless someone throws a grid computer at it...
_if_ it's widespread enough and mature enough that the community has had time and inclination to thoroughly examine it. I don't think you want to trust an algorithm or a complex system just because you've examined it yourself and think it's secure, even if you're Ron Rivest or Bruce Schneier. But components like PGP, ssh, and SSL have been examined pretty well and you can expect to have a good idea of how secure they're considered by the security community; the most popular open-source implementations have been examined closely as well, (even they still turn up minor problems from time to time).
Of course, whatever you choose you have to implement it correctly.
Sumner
You have the right idea, but your technology is about a decade out of date.
You can get smart cards/crypto cards or dongles that keep the keys on the card. It's *never* revealed, and all crypto is done on the card or dongle.
Until someone reads the key back off the card, which has been done with cheap ($50) hardware for "tamper-proof, unreadable" smart cards.
The real key is the human factors: don't let anyone get their hands on the key who you don't trust to keep it reasonably secure, and make sure that people understand how important it is.
Also, remember that the current setup is only of marginal security; don't spend millions of dollars to make the digital signatures bulletproof when someone can either forge a paper signature, fast-talk the employees into authorizing a purchase/signature/whatever, or otherwise circumvent the crypto. The free tools (gpg/OpenPGP or openssl based S/MIME solutions), if properly implemented, are more than sufficient to force would-be imposters back to their old tried and true ways.
In other words, worry more about how you implement the system (eg: key management) and how you train the users than about the particular cryptosystems and techie gadgets you choose.
Sumner
It's expensive, but very secure
_If_ you trust Lotus/IBM, who have been known to put backdoors into their export versions to give the NSA substantial portions of the encryption keys. And their code hasn't been vetted by the security community. And your actual level of security is highly dependent on your key management scheme.
Not to say it's a bad idea, but "Notes is very secure" is a pretty glib answer.
Sumner
Start->Settings->Control Panel->Tweak UI->Mouse Tab->Check "Activation Follows Mouse (X-Mouse)". Semi-supported by microsoft, should work with 99.9% of all windows and apps. Hardly a "hack".
It's worthless with most apps. As soon as you click in them to get keyboard focus in the right place, the window is raised. There are some apps that it's good enough with. Definitely a hack for a system that for whatever unknown reason wasn't designed to differentiate well between "raise a window" and "give a window focus".
Click and drag the top of the taskbar down to the bottom. It disappears. Or turn on auto-hide."Clippy"
He's dead. Why kick a dead horse?Windows window management has come of age. Looks like you dumped it back in '95 and never looked back
Actually, I dumped it in '94. But my current job has me using XP, so I've been full-time in front of that for 6 months now. And previous jobs sometimes had me full-time in NT 4.0 for periods of months (I've used TweakUI for several years now).
Its window management sucks. It still has this legacy thought that the user is only running one, or at most a few, apps at once (witness the amount of screen real estate you'd need to have the taskbar effectively manage 20-30 apps). And that only one of those apps will be used at a time (witness the impossibility of looking at one thing and typing into another). And that the apps should override the user wrt window management (witness pop-unders, which are impossible in my window manager under X because only the user or window manager can lower a window, not to mention that even without those user overrides window.blur() will loose focus but not lower a window in my window manager...) Gagaga.
Sumner
I'll agree that translucence and themability are fluff. I might be able to envision an actual use for translucence if I thought about it long enough, but it'd be a real corner case.
But...
Things that suck in Windows window management:
Sumner
Despite what you see on TV and read in the papers, the economic meltdown continues - the media is masking the severity of this downturn. If you want to see for yourself how bad things really are, take a trip through the US and note the closures of many businesses, the 'downsizing', and actually talk to people.
That's not a good way to get a clear picture of the economy. Employment is a lagging factor in economic recessions: things are bad for months before layoffs begin, and things are good for months before layoffs stop and hiring begins.
So if you go around looking at downsizing and talking to people who have lost their jobs, you'll get a picture of what the economy was like several months earlier.
It's also worth noting that unemployment in the US is currently at 6%. Not nearly as good as it was 2 years ago, but worth remembering that until then unemployment hadn't been below 5% since 1970 and the Fed was worried that an unemployment rate below 6% would spark inflation similar to that seen in the early 70s, decimating the economy. The current "horrible unemployment rate" was thought of just 7 years ago as "unsustainably high levels of employment that risk pushing the economy into a depression"--obviously alarmist, but the point is that the job situation is still better than it was in 1995, or in the economic boom of the 1980s.
Sure, I know a lot of people who lost their jobs--myself included. The company (a dot-com) I worked at went from 450 to 50 employees. But I don't know anyone who lost their job and didn't find another one at equal or higher pay within 3 months--and that's in the (comparitively) hard-hit software industry. Indeed, without 2 quarters of economic contraction or substantial unemployment, many people don't think we're in a recession at all. It's certainly a much milder one than we've had in my lifetime.
Sumner
Some people call this "liberal" or "liberal arts" education. It's funny because the US is the only country I know of that follows "the great European tradition of liberal education." I don't know of too many foreign universities with a "Great Books" program. French and German engineers don't have to take political philosophy courses in college. In fact, they probably won't take any polisci in the last couple years of high school.
Now, you try to get into a math grad school in the US with your BS and good grades, and you'll have a good chance; but why? Because schools can only take so many Hungarian and Romanian mathematicians who've actually been doing mathematics since high school.
Not really true in my experience. I went to public school (high school) in the U.S. and certainly had my share of mathematics--we finished AP Calculus (AB and BC) and some linear algebra in junior year and did differential equations and more linear algebra senior year. Of course, not everyone opts for the accelerated track which leads to...
The real difference between school in Europe vs. stateside is that in Europe there's no stigma associated with going to a trade school instead of university. Here in the US, trade schools are viewed as an option for the uneducated and so a lot of people who don't have any need or desire for a university education wind up in university anyway, which means that universities have a high percentage of courses covering things that really ought to be covered in high school. But the top US students, those who would go on to university in a European system, tend to have just as good an education. In my experience, anyway. In other words, the problem isn't that US high schools are that bad, it's that too many people are going to college who might be better served with European-style trade educations.
Or to put it delicately, most colleges have two tracks; one is those who are in college because it's the thing to do socially, and one is those who actually want/need a college degree. But the second group is about the same size (slightly larger per capita) as the European university group and in my estimation equally well-prepared.
The reason you stand a good shot at getting into grad school is that the tremendous number of college graduates in the US mostly don't apply to grad school. Only the top few do--there isn't the same social pressure to get a postgraduate degree. Indeed, about the same number of US students apply to grad school each year as the totals for France, Germany, and the UK, which works out to 2/3 as many per capita and less than 1/3 as many per college graduate. Those that apply are among the best prepared and the best qualified and matriculation rates from doctorate programs are actually higher for the Americans than the Europeans (even when grouped by university, though I can only find those numbers for U.S. doctoral programs and not for European or masters programs).
Handful of facts from the NSF's last study on higher education: U.S. university attendance has skyrocketed over the last 50 years, from 2.5 million enrolled to over 14 million enrolled. 35 out of every 100 college age individuals in the US are enrolled in college, as compared to 24 per 100 in Germany and 13 per 100 in France. The UK, Canada, and Australia are the major exceptions (UK sends 35/100).
Sumner
I very much agree with you about spending time with family, but speaking from experience, freelance is not the way to go. I and most of my colleagues work freelance/contract and over the past few years have spent almost every major holiday with last-minute crunches destroying our family time. And with freelance, as opposed to a "job" job, you need to do the time whenever it's required, or you don't work anymore.
The worst part is that no matter how hard you try to be available to go out with the kids on Hallowe'en, you'll get a life-or-death call from a client that afternoon which you have to take seriously.
That depends heavily on the sort of freelance work you do. When I do consulting, I _never_ allow any sort of on-call option. I definitely never let them have my cell phone or home phone number. I also do absolutely zero system admin or web development work, I stick strictly to software development (whether new products, new features, project rescue for failing projects, platform ports, cleanup, optimization, whatever...) with written requirements and I make sure that the contract backs me when it comes to saying "changing requirements==changing time estimates".
This is not to say I'm inflexible, changing requirements is a good thing as you work with the client to understand their needs (and often what they really want bears little resemblance to what they say they want, and they'll usually be very pleased if you can actually nail down something that helps them). Just that as you get a clearer view of what needs to be done, you also get a clearer view of how long it'll take.
I'd say 90% of non-web software development work (as opposed to admin work) has absolutely no reason to call the developer in late nights or weekends. You might need to pull some weekends if you're way off in your time estimate, just to meet deadline. But do that a couple times and you start to get good at providing better estimates. And it's usually way more interesting than web dev.
Sumner
Good call on the single-cluster angle, but the folks at Alesis [alesis.com] already beat you to it.
It's not an issue of getting beat to it, it's a really old idea. Heck, ReiserFS was hardly the first and they've done it for over 5 years. ext2 hasn't, but there is some movement in that direction.
Sumner
Every key name/value pair--being a file--will take up one cluster, regardless of how much data it actually contains
Not if your filesystem has decent support for small files, like e.g. ReiserFS or some of the recent ext2 devel branches. This is one thing ReiserFS definitely got right (though I'm less sure about the direction they're taking for ReiserFSv4).
As I said, it's a workable approach for grafting extensible metadata onto an existing FS, but it's not how you would design an ideal from-scratch FS.
It's definitely how I would design an ideal from-scratch FS. Metadata in the FS is a really horrible hack that goes against all the namespace unification efforts that OS designers have been making in the last 10-15 years.
Sumner
Bear in mind that one of the core driving philosophies that has allowed Unix to survive and evolve is "everything's a file", where a file is a stream of bytes. Attaching metadata directly to a file gets to be a nightmare when you start copying files over the network, dealing with editors that create new files rather than modify in place, backing things up, etc.
NeXT had a great solution to this: instead of just having a file "mysong.mp3", have a directory "mysong" that includes files like data.mp3, title.txt, etc. The file browser presented this to the user in a simple UI (e.g. the directory would have the "mp3" icon in the file browser, double-clicking it would play the song in the mp3 player, displaying the title.txt, etc). But traditional file transfer and backup tools worked just fine, and you could use standard tools (your regular editor, copy program, etc) instead of crufty special purpose ones (regedit, or binhex, or whatever) to work with the metadata.
Think of it this way: you don't want to just have 2-3 defined metadata fields, you would like to be able to add arbitrary new ones in a flexible way. And you want the contents of each to be flexible, without size limits or such. So you want a way of associating names with values--wait, we have that, it's called a filesystem. And you want a way of grouping a bunch of name-value pairs together--wait, we have that, it's called a directory. You'll either end up reimplementing a filesystem, badly, as your metadata, but need special tools to view/edit the metadata or move it around (a la resource forks or the registry), or you'll end up going with a NeXT-like solution and use the filesystem itself (which deals quite nicely with things like this).
Sumner
GPCL = GTK Pattern Creation Lab
GTK = Gimp Toolkit (original expansion)
GIMP = GNU Image Manipulation Program
GNU = GNU's not Unix
you also lose the one thing that makes gnome slower than tar
Actually, tar is pretty fast--it's bzip2 that makes it seem slow. Try gzip or lzop instead, or don't compress if you are storing compressed files--though maybe cpio is somewhat faster than tar.
(Sorry, couldn't resist)
Sumner
I didn't notice anything about online backups, point-in-time recovery, or standby database
Online backups: yes, for quite some time
point-in-time recovery: postgres uses WAL undo/redo logging, but I'm not sure what the state of rollback tools is at the moment.
Standby database: Assuming you mean Master/Slave replication, this is one of the major features planned for 7.3; 7.x has added a lot of the infrastructure needed for replication, and by 7.4 they hope to have multimaster replication (ie a fully distributed database).
SONY. Because caucasians are just too damn tall.
Crazy People. Hysterical movie.
Sumner
Worth keeping in mind that you can form a corporation for less than $300, which still excludes the poor but it's not just the super-rich that can exploit corporate law.
KDE only uses Xrender for its fonts.
It cant do true alpha channel effects.
Qt 3.0 and later supports XRender for alpha transparency, etc. Requires XFree86 4.1 or later. KDE 3.0 supports this (in CVS), using it for transparent menus and other stuff. Not released yet, and added 29 December (which was after the first KDE 3.0 Beta release, I think). The transparency is hardware accelerated on supported platforms (including lots of them, e.g. nvidia, ATI rage128, Matrox, etc)
Speed isnt as important as you think when you have a fast computer, frame buffer is slower? SLOWER? Its frame buffer, if you have nvidia and a pentium 4 is it slower?
For anything where speed really matters (e.g. video playback or other Xv stuff, 3D) yes, fb is definitely slower. It's probably slower for "normal" 2D stuff as well, that's true of most 2D cards but I don't have an nvidia so I can't bench it.
Sumner
This is the worst kind of offtopic nitpick, but I have to point out that Vi does run on a dumb terminal.
:-)
Y'know, I originally had written "you can't even run vi on a true dump terminal except in open mode", but I figured people would quibble over what a dumb terminal was and so changed to the glass tty aside. Should've left the open mode comment in too, I guess.
Sumner
While I agree that the toolkits suck, that email is rather irrelevent. If I interpret the email correctly, backing is referring to double-buffering.
:-O
No, it's referring to duplicate work being done by the toolkit. e.g. X sends notices saying "the lower-left needs a redraw", "the mid-left needs a redraw", and "the top-left needs a redraw. A properly written app would coalesce those into a single call to redraw the left side (ie clear the event queue of all exposure events before repainting), rather than repaint several times. gtk 2 takes some steps toward this.
Even as it is, cards with good driver support don't show the flicker problem. But well-written toolkit code would do the coalesce and work fine even on your old Cirrus Logic without flicker (or the new whiz-bang card with crappy unaccelerated drivers).
and hell, Photon is even more flexible than X. It even runs the graphics driver as a seperate process
How does this make it more flexible than X? X keeps the driver as a loadable object file, and puts the high-level GUI code (from Xlib on up) into a seperate address space from the X server. If you like, you can run a framebuffer video driver and fb X server, putting the low-level hardware support in a seperate address space. None of these really impacts the flexibility in any great way, though, the X extension mechanism is way more crucial to flexibility than which address space various components happen to run in.
QNX does rock, though. Between it and L4, I'm almost tempted to forget the travesty that is Mach. How's the swapper doing?
Sumner
Check the docu, you need to have it compiled into your X server and might need an Option line in XF86Config-4. If it's on, you'll get:
(==) R128(0): Silken mouse enabled
or similar in the XFree86 output.
If you're running a kernel with good latency (e.g. 2.4.17 + Andrew Morton's LL patches, I use 2.4.17-jl11) you'll no longer see any mouse cursor dragging/skipping/etc problems. If the kernel has bad latency, there's nothing X can do about it.
Sumner
I think part of the problem is the fact that there never seems to have been any coherent work done on this. The windowing system oriented people who work on X say "the toolkit authors fault". The toolkit authors would say "it's your drivers or the limitations of X Windows"
Nope, read the thread I quoted and you'll see that gtk developer Owen Taylor agrees and that gtk 2.0 includes some of the optimization mentioned. The toolkit and X11 authors do work together on these things, and the toolkit authors have had a huge amount of input into the design of the XRender extension and the DRI infrastructure.
While I do appreciate the flexibility of X Windows, I honestly DON'T think the windowing system and toolkits should be these totally orthogonal projects, and the toolkits just "draw as they see fit" on a canvas that they expect the windowing system to render dumbly. This is the X model, inherited from the dumb terminal days.
Actually that's not the X model (BTW, X wouldn't run on a dumb terminal--even vi wouldn't run on a true dumb terminal (ie glass tty)). The X model is to provide high-level graphics primitives to the application, which then submits them to the server which can turn them into whichever low-level calls are most efficient on the hardware in question. Not only that, but the library used to submit those request can (and does) batch them together so that the application writer can have a simpler model and still get efficient code--for instance, multiple XDrawLine calls are batched by XLib into a single XDrawLines call that's sent on to the server, saving on round-trips and in some cases saving on bus traffic to the video card by eliminating redundant traffic. Or servicing those high-level requests in whatever manner is most efficient for the hardware in use.
Highly efficient graphics can be done this way. Witness SGI, who were for years the undisputed leaders in the graphics field. They used X11.
But think of X as being more of a device-driver with a unified API, the GUI is to be built on top of that. It's a highly reasonable and well considered model that is ideal for building the high-performance GUIS of the future on. Far better than e.g. a framebuffer, which is already obsolete (doesn't handle many 2D features like overlays & alpha blending, doesn't do 3D acceleration, doesn't allow for hardware security a la SGI, doesn't handle hardware video decoding, etc) and is low-level enough that you can't have the driver do intelligent optimizations without rewriting the apps. And designed with the foresight to be extremely flexible.
Sumner
If you had read the thread I mentioned in the article you replied to, you'd see the anser to this one:
> > Not to be too non-technical...
> > > > If the protocol overhead is so small, why can't my 1200 mips (600mhzPIII)
> > machine resize windows without widgets streaking? My 486 could do
> > this fine running MS Windows. Is this because many widget toolkits (GTK,
> > QT) use XPutImage? There must be some way to speed things up.
blame your widget set. basically (sorry owen and co on the list) gtk a
(and i presume qt) dont render optimially at all. the do a semi-decent
job.. EXCEPt for opaque resizing, and when redrawing is more than a few
lines and boxes... this is a toolkit issue and imho the current set of
toolkits (motif, qt, gtk etc.) do a god-awful job of this kind of
stuff. right now i have silky smooth "opaque resize" stuff working here
with enlightenment 17 - but i do the rendering completely differently
to gtk/qt - its all a canvas and thus the rendering happens in a
"backing" so updates are smooth. on todays hardware this is the best
way to do it and have almots no artifacts ANd retain speed.
> "Streaking"? Are these opaque resizes? Alot of apps aren't doing
> event compression. They repaint the whole damn window every time they
> get an event. They could have at least checked that there weren't
> more events in the queue and got rid of them instead of handling each
> one in turn.
true. its a very bad thing that there are a LOT of apps that behave like
this... a LOT. some of the most commonly used are guilty of this
(netscape for one....)
If you enable Silken Mouse in XFree86 4.0 and later, this should be fixed. Certainly an implementation issue and not an architectural issue (i.e. not a reason to throw out X and start over)
These aren't X11 problems but GUI problems, GUI standardization is certainly a huge issue. But, gtk-2.0's accessibility enhancements include excellent keyboard support and some steps toward simplifying and unifying look&feel. KDE is moving in that direction as well. Obviously you need to use a single unified UI on your desktop, but having two decent ones available to choose from is not a bad thing (not to say that either is decent yet, but they're both heading there rapidly).
Sumner
In fact, when X was originally developed Jim Gettys et al considered putting _all_ graphics rendering in an extension (leaving just the core windowing w/o rendering in the core). They fully expected the original rendering model to be replaced fairly soon, but that's taken a long time. XRender hopes to do that and probably will largely supplant the old rendering primitives for new apps in a few years. Maybe sooner for gtk/Qt/other whizzbang bleeding-edge stuff.
--Jim Gettys, 2001
--Keith Packard, 2001
From the thread
Proposal for server-side Anti-Aliased fonts
Sumner
The problem with X11 is, in part, the separation of client/server; this causes extra latency and a heap of context switches
The context switches aren't a significant overhead. They weren't even a significant overhead in 1986 when Sun first started spreading FUD about this (at the time, Sun was trying to push NeWs over X11). See e.g. Jim Gettys' posts in the "rendering model in X" thread in the Xrender mailing list archives
It's not all sunshine, he's willing to own up to places where X needs improvement (exposure lists are a big one, througput for e.g. texture mapping is another), but it's way better than a lot of people claim. And Xrender and DRI address the vast majority of the problem cases very effectively.
Sumner
Most services stopped responding, and the administrative port wasn't turned on when they installed the new hardware. H.P Anvin posted to lkml about this. They're working on getting it back up and running.
Sumner
Including the patches, which are not enabled unless you choose the option while configuring the kernel, wouldn't mess up kernel debugging, would it?
But if a user has preempt enabled and submits a bug report, that bug report becomes very difficult for kernel developers to decipher. Not to mention that hard-to-reproduce bugs may show up under preempt, then you have to build a non-preempt kernel and try for a potentially long time to reproduce the error.
I am not a kernel hacker, and I was not commenting on the code; I feel that the idea of leveraging the SMP locks is elegant.
It would be, if the locking requirements for SMP were sufficient to deal with the locking requirements for preempt. But they're not, these are really two different beasts--see my other message under this topic for a laundry list of reasons and linux-kernel topic to read.
Could you explain what "live locks" are, please? I'm not familiar with the term.
Definition of livelock
Sumner