I have bad news for you, if you are worried about blocking. You are guilty of it. You do it all the time. Every day, with every breath you take, you block. It might be that you choose not to read the latest news about the Kosovo crisis, or you choose not to find out how one would give one's poodle a perm.
Humans encounter more information today than they ever have before, and as in the past, the method of dealing with it is to filter it. As a matter of fact, all five of your senses are actively filtering the world around you with every breath that you take.
This is *not* the same as censorship. Censorship is when someone *else* filters for you. When I put up my proxy filter and choose which ads to block out, I am simply filtering a barrage of information which has little use to me in my life, or at least I *think* it has little use to me. Your eyes and ears, your sense of smell all do this on a regular basis. If I'm wrong -- well, that is my loss. However it is my right to choose what I see. And I do, and you do it too.
Some forms of "blocking", like slashdot, come very close to censorship, particularly when the moderators come into play. You could almost percieve it as having a *huge* committee of people who choose what you see -- if you opt for moderated content. But it is important to remember that you are still able to choose whether you want moderated content or not, so in my opinion even this is not censorship. It is much like watching the news on TV. You "vote with your dollar" for someone to filter information for you.
If the persons you select to filter information for you do it in a way that you didn't think they were doing it -- that's censorship. But we all know how moderators here work, so if you opt for it.. well, you picks yer filters, and you installs them.
Think wall-filling, crisp beautiful display. Seriously, these are not that expensive these days, and they are certainly less than people are paying for those massive HDTV's (most of which are rear projection screens, and much cruddier in terms of display.
This is what I plan to do anyway. Not to mention the possibility of playing Descent on the thing... zowie!
Linux needs formal engaged testing? Don't be sure.
on
Slate Takes on Linux
·
· Score: 2
snipped from the article... Furthermore, as I've written before in Slate, software companies spend a surprising fraction of their resources testing software, not writing it. In my experience, this is the ultimate problem with Open Source development: not enough formal engaged testing. Developers want to write code, they don't want to solve all the niggling little problems that users come up with.
Oooh, very bad. Companies spend a surprising fraction of their resources on testing for one reason, and one reason only. That reason is that the development process is all messed up. Requirements and specifications are not laid out properly, then design is not thorough. The development process is supposed to be front-end loaded, with over 50% of the time being spent before a single line of code is written. Test should be about 10% of the time. Why? What does it matter where you spend your time? Data shows that the further along in the development process a bug is found, the more it costs - exponentially! If a bug costs $100 to repair in design, it will cost $1000 to detect and repair in coding, and $10,000 to detect and repair in test. It only gets worst after release. This has been definatively proven (to my mind) by the Software Engineering Institute (SEI), and is the topic of countless software engineering and design papers (all coming to the same conclusion). A great deal of data has been gathered to support this, and you only have to read the software engineering papers out there.
Okay, all that said -- does that mean that linux has 10% testing, and it is the right 10%? I'm not sure, but I don't think so. I think that typically linux programmers (like most software engineers) fall into the same trap -- leaping into coding before a proper design has been done... because coding is fun, and that is what they like to do. So how does the "huge chunk" of testing get done? Linux software is quality software.. somehow. I believe that the testing gets done by the users and beta testers, and while it isn't formal (which is always good) it is one of the ultimate forms of testing -- if the software works for the user, then the software works. Even if there are bugs. The user is the definition of the requirements, and of what meets the requirements. Most of this paragraph is just conjecture though, because not much data has been gathered to support it.
Copper may be a problem, but optic, certainly not? It is a bit expensive now, but compare the expense of fiber to the expense of your wireless solution.. I bet the fiber compares quite well. I bet if you do a cost/bandwith analysis, it compares even better.
And even if we went wireless what would we have then -- great bandwidth... on clear days. What happens when it rains and snows? Your bandwith goes out the the window (if you even still have a connection).
Wireless may be good for one thing as you said -- distributed LANS. Lets keep it out of the home.
So, It would be MS GPL Linux 2.2.7. Big hairy deal. A simple UNIX utility called "diff" will tell you exactly what changes MS made, and a good developer will probably be able to guess why. Then the changes can be merged right back in, or flatly ignored.
A beowulf cluster is good and cool for what it does because it has a good bang for buck ratio. Now these may have a fairly good dollar/spec(int/fp) ratio, but stupid things like Imacs don't. Get real people.
Well, this is confusing to me. We have a lot of mission-critical systems, and all the major vendors do their support this way:
Digital(Compaq), SGI, IBM, HP, Sun (actually we use Polaris, and have a software support contract with sun), Network Appliance...
Typically you get some commitment, like 4 or 8 hour response time, on hardware and/or software. I don't see why linux should be any different, or why Compaq, HP or any of the others would change their mode of support for linux. In fact, I'd be surprised if they didn't simply integrate linux into their current support infrastructures -- it would be wasteful and difficult to track trouble tickets otherwise.
An ideal OS would be one that does fit all. You are all confusing the operating system with the software that comes with it (applications, utilities).
Modern operating systems, heck, even modern versions of very old operating systems (UNIX) have modern kernels which are fast, fairly compact, and tunable to all heck. Most support some sort of modules, so you literally can use the same *compile* of the kernel for two different purposes. Your server kernel can simply load the raid drivers, or gigabit ethernet drivers, and your client kernel can load the sound drivers.
The other silly stuff, the applications -- of course it makes sense to have different distributions with different applications (for server or desktop), if only so that you can install off of one CD.
Why the heck you would want to have a different core operating system (kernel), and a "mostly compatible" API, and "mostly compatible" libraries, I don't know. Oh wait -- yes, I do know -- that glorious god of backwards compatability.
Carry on being stupid, and incapable of writing a decent operating system M$, people will carry on migrating elsewhere.
The EULA (at least as Chris has on her site) does not specify a full refund, just a refund, and she got exactly that. The arguement could be made in court that a full refund is implied, but it would be difficult to back up. So she is lucky she didn't get $0.01.
A friend of mine went to a theme park trade show and got to play a game where you sit in a motorized chair and you are the pinball. You get to control the bumpers still, but the screen perspective is either from the pinball itself, or above it. So you can see the bumper looming towards you then smack you in the face.
Slashdot Mirror
I have bad news for you, if you are worried about blocking. You are guilty of it. You do it all the time. Every day, with every breath you take, you block. It might be that you choose not to read the latest news about the Kosovo crisis, or you choose not to find out how one would give one's poodle a perm.
Humans encounter more information today than they ever have before, and as in the past, the method of dealing with it is to filter it. As a matter of fact, all five of your senses are actively filtering the world around you with every breath that you take.
This is *not* the same as censorship. Censorship is when someone *else* filters for you. When I put up my proxy filter and choose which ads to block out, I am simply filtering a barrage of information which has little use to me in my life, or at least I *think* it has little use to me. Your eyes and ears, your sense of smell all do this on a regular basis. If I'm wrong -- well, that is my loss. However it is my right to choose what I see. And I do, and you do it too.
Some forms of "blocking", like slashdot, come very close to censorship, particularly when the moderators come into play. You could almost percieve it as having a *huge* committee of people who choose what you see -- if you opt for moderated content. But it is important to remember that you are still able to choose whether you want moderated content or not, so in my opinion even this is not censorship. It is much like watching the news on TV. You "vote with your dollar" for someone to filter information for you.
If the persons you select to filter information for you do it in a way that you didn't think they were doing it -- that's censorship. But we all know how moderators here work, so if you opt for it.. well, you picks yer filters, and you installs them.
Deal.
Think wall-filling, crisp beautiful display. Seriously, these are not that expensive these days, and they are certainly less than people are paying for those massive HDTV's (most of which are rear projection screens, and much cruddier in terms of display.
This is what I plan to do anyway. Not to mention the possibility of playing Descent on the thing... zowie!
snipped from the article...
Furthermore, as I've written before in Slate, software companies spend a surprising fraction of their resources testing software, not writing it. In my experience, this is the ultimate problem with Open Source development: not enough formal engaged testing. Developers want to write code, they don't want to solve all the niggling little problems that users come up with.
Oooh, very bad. Companies spend a surprising fraction of their resources on testing for one
reason, and one reason only. That reason is that the development process is all messed up. Requirements
and specifications are not laid out properly, then design is not thorough. The development process
is supposed to be front-end loaded, with over 50% of the time being spent before a single line of code
is written. Test should be about 10% of the time. Why? What does it matter where you spend your time? Data
shows that the further along in the development process a bug is found, the more it costs - exponentially! If a bug costs
$100 to repair in design, it will cost $1000 to detect and repair in coding, and $10,000 to detect and repair in test. It only
gets worst after release. This has been definatively proven (to my mind) by the Software Engineering Institute (SEI),
and is the topic of countless software engineering and design papers (all coming to the same conclusion). A great deal
of data has been gathered to support this, and you only have to read the software engineering papers out there.
Okay, all that said -- does that mean that linux has 10% testing, and it is the right 10%? I'm not sure, but
I don't think so. I think that typically linux programmers (like most software engineers) fall into the same trap --
leaping into coding before a proper design has been done... because coding is fun, and that is what they like to do.
So how does the "huge chunk" of testing get done? Linux software is quality software.. somehow. I believe that
the testing gets done by the users and beta testers, and while it isn't formal (which is always good) it is one of the
ultimate forms of testing -- if the software works for the user, then the software works. Even if there are bugs.
The user is the definition of the requirements, and of what meets the requirements. Most of this paragraph is just
conjecture though, because not much data has been gathered to support it.
Copper may be a problem, but optic, certainly not? It is a bit expensive now, but compare the expense of fiber to the expense of your wireless solution.. I bet the fiber compares quite well. I bet if you do a cost/bandwith analysis, it compares even better.
And even if we went wireless what would we have then -- great bandwidth... on clear days. What happens when it rains and snows? Your bandwith goes out the the window (if you even still have a connection).
Wireless may be good for one thing as you said -- distributed LANS. Lets keep it out of the home.
So, It would be MS GPL Linux 2.2.7. Big hairy deal. A simple UNIX utility called "diff" will tell you exactly what changes MS made, and a good developer will probably be able to guess why. Then the changes can be merged right back in, or flatly ignored.
Grr. I meant a *grin*, since slashdot filters out the greater/less thans.
Did I forget to add a to the end of that? :-P
Okay, I guess I missed something... who won a harley, and when, at what conference? I'm guessing Linus...
Pity it couldn't have been a decent bike.. but at least he can sell it for a tidy profit and buy himself a nice bike.. or two.
A beowulf cluster is good and cool for what it does because it has a good bang for buck ratio. Now these may have a fairly good dollar/spec(int/fp) ratio, but stupid things like Imacs don't. Get real people.
Digital(Compaq), SGI, IBM, HP, Sun (actually we use Polaris, and have a software support contract with sun), Network Appliance...
Typically you get some commitment, like 4 or 8 hour response time, on hardware and/or software. I don't see why linux should be any different, or why Compaq, HP or any of the others would change their mode of support for linux. In fact, I'd be surprised if they didn't simply integrate linux into their current support infrastructures -- it would be wasteful and difficult to track trouble tickets otherwise.
Hey cool. I actually did the search at Avnet, and this is the pricing info that came up:
Part #/Description Piece Price in US Dollars
MTX604-070 0
DUAL
604R,333MHZ,7PCI
SLOTS,DUAL
ETHERNET
Modern operating systems, heck, even modern versions of very old operating systems (UNIX) have modern kernels which are fast, fairly compact, and tunable to all heck. Most support some sort of modules, so you literally can use the same *compile* of the kernel for two different purposes. Your server kernel can simply load the raid drivers, or gigabit ethernet drivers, and your client kernel can load the sound drivers.
The other silly stuff, the applications -- of course it makes sense to have different distributions with different applications (for server or desktop), if only so that you can install off of one CD.
Why the heck you would want to have a different core operating system (kernel), and a "mostly compatible" API, and "mostly compatible" libraries, I don't know. Oh wait -- yes, I do know -- that glorious god of backwards compatability.
Carry on being stupid, and incapable of writing a decent operating system M$, people will carry on migrating elsewhere.
The EULA (at least as Chris has on her site) does not specify a full refund, just a refund, and she got exactly that. The arguement could be made in court that a full refund is implied, but it would be difficult to back up. So she is lucky she didn't get $0.01.
A friend of mine went to a theme park trade show
and got to play a game where you sit in a motorized chair and you are the pinball. You get to control the bumpers still, but the screen perspective is either from the pinball itself, or above it. So you can see the bumper looming towards you then smack you in the face.