I'm sorry, but I'm forced to disagree with your characterizations of both McCain and Obama.
I spoke with McCain about a year and a half ago, before he was an official presidential candidate but long after he had decided to run. He gave what has now become his usual justification of his candidacy, followed by a brief discussion of the issues of the war in Iraq, the economy, school vouchers, torture, and state's rights (I live in South Carolina). I then asked him what, given the deeply conservative agenda he had laid out, would he say to those who had called him a maverick or independent? His response was as follows:
"I am sorry, but I am a deeply conservative candidate"
His voting record backs this up- where he breaks with the Republican party, he does not typically break left. He breaks in favor of what he calls 'higher conservative principles', and in practice this has nearly always meant going further right.
I consider myself an independent. I've supported candidates on both sides of the aisle for president and local office both, and have spoken (admittedly briefly) with both the men you are characterizing. Senator McCain is undoubtedly one of the most conservative candidates in recent electoral history, if such a label has any meaning anymore.
As for Obama, I am forced once again to disagree. You should know that I support the Obama candidacy.
I presume that when you say that Obama is further left than Hillary, you are relying upon the oft-quoted and little understood study conducted by the National Review (a publication that, incidentally, cannot by any reasonable person be called 'unbiased') indicating that Obama was the "most liberal Senator", with Hillary occupying the "16th most liberal Senator" spot. The methodology was simple: classify all substantive votes as either liberal or conservative, and tabulate the results.
As a person who fundamentally disagrees with the "you're a liberal or you're a conservative" zero-sum-game approach to politics, I have to disagree with this methodology. It is also worth pointing out that somehow John Kerry earned the coveted top spot on the study as soon as his candidacy was announced, making it harder to believe that the individuals deciding what constitutes 'liberal' and 'conservative' votes are looking strictly at politics.
As a second point, Obama has made it a cornerstone of his campaign that there is a need to end the process of politics as currently practiced. There are those who say that it's empty rhetoric, but to my ears those people aren't listening very well. It means that he has built his candidacy on how he will do what he will do as president, rather than what he will do. We have a President right now who, while undoubtedly flawed, has committed no political sin more grave than trying to go it alone. I might even advance an argument that without that, his presidency would have been disliked but not hated. This is a crucially important point: that while what you do in office is important, the Presidency is a politically weak office, but a highly visible one. How you do things is sometimes more important than what you do.
This is really freaking obvious. I wrote a p2p application that cached based on search requests and then fetched based on router hops years ago, and presumed it was nothing new then. I strongly doubt this will be an unencumbered technology if it ever sees the light of day.
I'm sorry, but you're discussing things in a reasonable and intelligent manner. Your Intertubes license is officially revoked. Turn in your e-peen at the counter.
If you're putting a week's worth up and selling out in a day, that's not the end of the world. If you're putting two month's worth up and selling out that afternoon, you have a big problem.
I somewhat doubt both of those predictions.
It's ok, but it would improve if you sold Windows-based PCs instead Lot of competition in that space. On the low end machines, a quarter of the cost can be software licensing. My experience is that that cost has to be passed on to the customer, and that it leads to lost sales. Your mileage may vary, of course.
At the Wal-Mart scale, it's millions more to make by axing the Linux PC in favor of cheap, Windows-based machines, so it becomes a simple business decision I don't dispute that Wal-Mart is making a business decision here, or that it's a simple one, but I think people are overlooking the probable reason- Wal-Mart loves good suppliers, and hates bad ones. It wouldn't be the first time Wal-Mart axed a profitable product because the supply chain just wasn't able to support demand, and TFA clearly states that these repeatedly sold out.
Because you clearly do not understand the PC business, the article, or the events in question, allow me to point out a few small facts you appear to have overlooked:
1) PCs are a commodity business. You don't stick around selling 3 of them in a month.
2) If you're making $1000 off of 3 PCs, I want whoever your marketing guy is.
3) Wal-Mart doesn't order things in 50s. It orders them in thousands.
4) TFA clearly states that Wal-Mart repeatedly sold out of the machines.
Put it together. Wal-Mart has sold thousands of these machines out repeatedly- which means that it has a product whose supplier cannot meet demand. If you're a company that size and want to lose a lot of money, the way to do it is to have to deal with somebody else's god awful supply chain.
This is mostly what I've been saying. I know I could eventually get what I need, but in today's computers and GUI applications, if I need to go as far as going to the source, the application has failed from a usability standpoint. That's my issue, and it's likely an issue for many others.
Depends on what you mean by 'usability'. End users should never have to go further than the forums or IRC. Administrators should never have to go further than the man pages. Developers should be looking at either API documentation or source, and preferably both. All of the above are addressed in a good program, and most of the *major* projects do so. You've already said that you're not really a fan of the major avenue of assistance in Linux, which is the community forums and IRC, and I guess if that's how you feel about it that's the way it is, but I don't think that without community assistance you're comparing apples to apples in terms of support.
I said that the options, which may be dangerous, are still accessable and that adequate warnings and explainations are given. Right there, in the dialog to set the option (or by pressing F1). I found that in Linux, I would find the option, but the various settings of that option aren't adequately explained where you set the option. You need to go digging. Again, it's a usablity failure.
I'm not sure if there's a particular piece of software than you're referring to here, but most man pages seem exhaustively complete to me, especially on major projects./shrug.
Again, quite a lot of steps. Does Unison setup both client and servers? When building a new workstation do I need to use Unison to configure the client again, or is there a way to say "this server controls your settings" and leave it at that? I'm not sure I would have ever found or heard of Unison if not for this discussion... which again I think points to a problem.
If you already have rsync and ssh on your server, you just need to install Unison on the client, tell it which directories to sync which way (can be saved in a config file, which can also be shared safely) and point it at your server.
As for not having heard of it without my help, well- yeah. That's why the community exists. That's what I mean by living documentation. If you ask somebody a question- and we're easy to find- we'll answer if we can, and know where to look if we can't.
Hmm... options can be good and bad. In this case, I lean toward bad, because I now have to evaluate one of these packages before I can even start administering my network. If I choose poorly, I have to undo everything and switch to another package.
Hmmm. I'm afraid I don't agree with you here- I tend to believe that competition in a market segment tends to improve the quality of the competitors and increase the power of the consumer- but I guess to each their own.
No, my knowledge of it was that it is a scripting language (at least, when I first heard of it, it was) and I try to stear clear of those. Maybe it's changed, or maybe I'm confusing it with TCL/TK. I dunno, it's been a while. I tend to stick to languages that are similar to C / C++ sytax. I don't know why, but I've always liked it more than Basic, and other languages just haven't appealed to me as much. Not that I haven't used a few, at least a couple of times each.. lisp, snobol, eiffel, VAX assembly and some other's I'm forgetting.:-)
It's a scripting language in the same sense as lisp is- interpreted, quick to write, blurs the line between data and code- but it's pretty common for a lot of general purpose programming anymore. A lot of the syntax is pretty much C++ sans punctuation, which I find handy, but I figured since you mentioned both.NET and Java that I would mention Python since the three major implementations of Python are C,.NET CLR, and Java.
So you've never really used Windows then (the NT line.. 2000 or higher)? I've found lots of people say Windows is for morons or its overly simply. The fact is that it's complex, and much of the complexity is hidden. But its there if you really need to dig in. Given that I've seen many comments about virtualization and supercomputing that mirror yours, I'm not suprised that Windows doesn't fit your needs. I would check back now and then though (maybe now, because Server 2008 is out), because MS is trying to move into these areas. I've found that they usually succeed; not on the first or second attempts, but they have a good record. I know.. that's not what most/.ers will say.
No, I've used Windows extensively. First, my high school days aren't all that far behind me, and XP was already out. Second, as part of my business- many legacy business apps cannot be ported without monumental effort, and so it is frequently the case that we set up a Windows box with a Linux passthrough.
As for Windows supercomputing, I wish them great good luck but don't see it happening, as the UNIX world's tools for HPC etc are just incredible. It would take a LONG time and a lot of money to come up with a comparable feature set.
Well, you've probably had much more time. All of my Linux experience was me, all by myself, trying to be inbetween. Wanting to administer my own network like an expert would, yet not having the resources provided if I worked with Linux all day. Again, maybe you're right about home users, and my experience is because I'm in this weird limbo. As far as making unsafe options available, well, we're not talking about a gun, so lets not go down that route. You can reformat a machine that gets owned, you can't bring someone back from the dead. Also, if the option is adequately explained and the security implications are stated, I don't see what the problem is. In the end, you can set that unsafe option on either system. The problem with Linux seems to be they are just hiding it, so if you find the option you may be less aware of the implications. Windows points it out, saying "hey, this could be really bad, here's why, but if you want, go for it." As someone trying to learn more about administring systems, I find that is more helpful.
Its all there, it just happens to be in either the man files or the ultimate documentation. Once you learn how to use a command's basic invocation (from irc or wherever), then check the man pages for reference documentation, and you're generally fine. Again, all of this is anecdotal- my experience is x, yours is y, ergo, we run around in circles, but like I say, between talking directly with the developers, reading the man pages, and, if you're really masochistic, perusing the source code, you'll figure it out. In the end, there's really nothing in Linux that's totally opaque, which is, once again, a major advantage for those willing to take advantage of it.
As far as security is concerned, we probably don't want to get into that flamewar, but there are a lot of things you can do to a system that simply can't be undone. If your customer database gets cracked, its game over. If someone installs a malicious utility, its game over. God forbid they be able to insert a kernel module. The assumption that box security is more important that infrastructure or data security is one of the major problems with home use security models, and really highlights a major difference in approach between Windows and Linux. It doesn't help that much of security research is very convoluted stuff, and that many elements of a security strategy cannot be considered in isolation. So, I can't really agree that every insecure option should be just as easily available as any given security-neutral option, any more than I think that every option available on the CLI needs to be available in the GUI.
Your suggestion is to send out configuration files via rsync? Fair enough. Seems like this should be made easier to do tho
I understand your experience differs, and that's fine as you said. However, a couple of things. Did you learn much of what you needed for Linux because you have a job doing so? I'm certainly not afraid of Linux by any means. Its just that its more difficult to learn. I do fine with shells and what not, after all I used Unix all through my five year college program. That's when I first came into contact with Linux as well. "Unix on my PC, awesome!"
My areas of interest in computing are supercomputing, CPR, and (more recently) virtualization. Windows does not give me the kind of power as a developer that I need to pursue my research in those areas. After a brief flirtation with Solaris when I was in high school (the school had a Sun server), I moved to Linux full time and haven't looked back.
Your user experience is limited (it seems) to business users. That's fine as I said, but I don't think that the experience a business user of linux (who likely doesn't manage the computer they use in any way) translates to a home user well (because they have to "manage" the system themselves).
We set up all kinds of systems, in all kinds of offices (including home offices), but most of them are not centrally managed, and in many of them- especially former Mac shops- the users have a tremendous degree of latitude in what they do and how they do it. You are absolutely right that the comparison between a large corporate desktop and a standard home machine is pretty much apples and oranges, but our installs on a lot of machines are basically stock Ubuntu, Mint, or (if requested) Fedora. Nobody seems to have a big problem, and in fact many people only notice that the background has changed and that the start button is in a different place. After that, most don't *seem* to care, although of course they know me as 'the Linux guy', and people aren't usually in a hurry to criticize others' livelihoods. Again, however, I have no data with which to back up my claims, so it could just be my admittedly biased perception.
In that sense, market share IS important, because home users also add hardware to their PC. You can't just buy anything and assume it will work on Linux.
The only devices I've had real problems with on Linux- besides really strange things like one-off industrial controllers- are blackberries. They really just don't work the way they're supposed to, although the Barry project tries really hard. Besides that, most hardware does work now, and again, I have nothing but anecdotes to hand you.
As far as/home being on an NFS mount; I think that was recommended when I first setup my server. At any rate, the idea was to keep user data on the server. I understand WHY everything locked up, but it didn't lead to a great experience. Contrast this to Windows; a few group policy settings (which affect more than the location of user documents) and user files and settings are stored on the server. With Windows, its smart enough to use the local store and sync later. No configuration needed.
Well, its not the lockout issues I was commenting on so much as security, since/home is frequently scanned by UID 0 processes and UID 0 access is badly handled by NFS, leading to serious security issues- but then I'm the guy who includes TEMPEST scanning as part of a security audit. The better solution would probably have been SSHFS overmounting a known safe/home on the local machine. It probably would have spared you a lot of headaches and would definitely have been more secure. It is also *far* easier to set up, since it requires less in the way of cooperation from your server. Another possible route (and much more comparable to Windows' group policy) would have been rsync, which is quite easy to set up.
You've said several times now that you think that Windows is easier to use than Linux. I don't think that's correct at very many levels of user experience, but again, we're throwing anecdotes
Sounds good. Presuming that there is merit in the methodology here, we may see some more competition in the market. I'd like it more if there were more code sharing going on, but maybe if we don't get tempers too hot we'll see a little more of that too.
I may be being stupid here, but I'm trying to follow this and not getting it. Please correct where my chain of logic is going awry.
1) the point of GPL that you can't distribute code that's derived from GPL code under anything other than the GPL.
2) Linking against code is considered 'derivative'. Using a public interface is not.
3) NDISWrapper links against the GPL'd kernel.
4) NDISWrapper is required to release under GPL and does so.
5) NDISWrapper uses the public interface of the binary driver
6) This imposes no license requirements on NDISwrapper
7) No license violations have occurred.
Even if Linus doesn't like it, it doesn't seem like there's anything in the GPL that specifically stipulates that you can't both link against GPL'd code and use the public interface of non-GPL'd code, and if there is, it will cause a LOT of problems in LinuxLand. Am I wrong?
Well, you're giving them training. Are you also taking over managing the servers from then on? As I attempted to be a self taught admin (somewhat) sometimes I didn't care about the details of why something worked, I just wanted it to work. I can dig into the whys later. I found on Linux, I had to spend an inornate amount of time to figure out it, and the why details usually drowned out the how.
We do give them training, but we don't take over management, preferring instead to offer break-n-bake support plans and service guarantees. We do both desktops and servers, and training for both IT staff and end users. We find that end users typically cause no more problems on Linux than they do on Windows after an adjustment period, and depending on the setup, sometimes less. IT is always a hassle, but that's just the nature of the beast- people do not like retraining away from a skill set that they have based their careers around.
You are correct that Linux as a server OS does not encourage you to perform operations you do not fully understand. Linux on the desktop, I think we may have to agree to disagree. My personal experience (no more or less valid as evidence than your experiences, of course) is that we have few problems with desktop installs, and that what problems we have usually come down to unsupported hardware (industrial controls, etc), and even that has become quite infrequent now. To the best of my knowledge, we haven't had any major complaints about the (admittedly limited) applications we install.
Which is why I don't think OSS will get anywhere; someone HAS to be trained, they can't just sit down with very basic knowledge and off they go. In a business setting, you usually have IT to call on to answer questions. Home users don't, and they sometimes find certain tasks in Windows hard to do. Imagine home users now trying to muddle through Linux.
Again, I think we may have to agree to disagree. I don't think that very many people would have argued with you a few years ago- I know I wouldn't have- but I also think that OSS has come a long way since then. Of course, at this point we are using OSS as a synonym for Linux, which I in particular ought to know better than to do, but I hope you'll understand if I use it as shorthand where I don't otherwise elaborate.
My personal experience (again, not valid data) is not broadly representative. It is based on our service to a region that is both desperately poor and possessed of an educational system that would shame Somalia. Most of our end users treat computers as bewildering, terrifying, and capricious lesser gods, whose power to help you is only exceeded by its ability and desire to do you harm. They have little Windows experience, and have never heard of Linux or, indeed, an operating system, before. They don't understand, and don't care, as long as it keeps the gray box happy. We have little difficulty in training them to do basic office and worksite tasks on Linux, and life goes on. Our more experienced Windows users are another story. They do not like to retrain, and frequently try to treat Linux as "Windows for cheapskates". They run everything as root to avoid the inconvenience of passwords, they complain bitterly that things are not in exactly the same place "as they used to be", and they generally experience higher levels of frustration and greater difficulties than the above mentioned group, even if they do not do significantly more with those machines. I suppose you can turn that into an argument for Windows- that people are already trained on it and do not need retraining- but while that makes a good business case I don't think it addresses the fundamental question we're talking about, which seems to be the quality of the software in question.
If the project continues, and statistically speaking no one is using it, does it matter? There's no competition at that point.
I think I wasn't as clear as I wanted to be. When I say competitive, I mean feature-for-featu
Do you run Linux on your desktop? I've found this to be patently false. Yes, I do- and more importantly, on my clients' desktops, since a major part of my job is windows to Linux migrations and the accompanying training. We consistently find that after an acclimation period productivity sits right about where it did before. It sounds like you had a bad experience, and as a developer I really am sorry about that, but it doesn't parallel what our clients find or we wouldn't be in business.
Part of my point is that the FF team has a big player backing them with money. Should that company no longer fund the primary developers, would FF survive? Would the project continue? Yes. Would it have anywhere near the market share it does? No. But the amount of cash involved has nothing to do with whether a project is open source or not.
GNUCash is a more appropriate model; the developers think I want all the gory details of double entry accounting, which I don't. Things like MS Money or Quicken hide those details for me. GNUCash also was lacking in any reporting of any kind. They even admit it: "I'd guess that GnuCash's biggest problem is that programmers don't use the software." These are the majority of OSS projects. Again, I'm sorry to hear that you had a bad experience, but part of what I was saying earlier is the fact that the vast majority of software is bad software, open or closed source. For every bad open source program that you point out, I can point out a bad closed source program, and we're back to the "my anecdote is better than your anecdote" game. In the absence of data, we are just opining.
Well, the problem is that without objective qualifications we can't really ever evaluate if your statement is true. As I've tried to carefully explain *we*have*no*data*. All we have are perceptions and opinions. I, naturally, think my opinion is correct; that the open source model provides specific advantages, that much open source software is of high quality and that it is in many respects a level playing field, with the leaders in some market segments being closed source and the leaders in others being open. You, naturally, think yours is correct, and while I disagree with it, that fact doesn't make it any more or less valid.
Well, I imagine we could forever too, since we don't have anything objective to go on. Given the level of sarcasm, I presume you have a nice simple 10-or-so word definition of 'good software' that I can whip up a test for and be done? I didn't think so. People who live in glass houses...
It was harder to use. GNUcash wasn't cutting it (no reports), I still had random crashes. The documentation is non-existent in most cases, and managing my network is easier with Windows Server 2003 (SBS). And of course RPM hell. I felt like I was spending more time fighting to get Linux to do what I want than actually having it do what I wanted. And I put the time in too; I had run a linux server for 10 years and on the desktop for three. It was honestly more frustrating than Windows. I'm not sure about GNUcash, as I don't use it (we have an accountant for that) but I'm amazed that you had so much difficulty with the desktop, given that level of experience with Linux. We get individuals who have no technical background to the point where IT calls are down to normal within a few days, and with one major exception have never had a persistent problem. Any issues in particular? How long ago was this?
You wouldn't happen to know of any good, accessible resources for somebody just getting into object recognition would you? So far I seem to have a knack for selecting the most obtuse references I can and it's starting to pick my pocket- any help would be much appreciated.
I'm not your strawman, so put down the torch. I don't think this is the year of the Linux desktop, and have been quite explicit in saying that I think that OSS is very good at some things, while closed source models seem to do better at others. I think that many Linux desktop applications are at or approaching the point of competitiveness with their closed-source brethren, but I also think that its going to be a long time before some parts of the industry come under contention.
My initial question was simply whether OSS will wind up dominating arenas with disjoint user and developer bases the way that it does where those two groups are closely aligned, and you ask when we'll know the answer to that. My general feeling is that we already know that it is possible for projects like Firefox to contest some segments of the market, but that we probably won't see just one model in very many parts of the industry unless an application that simply obliterates its competition winds up emerging, and I don't think its possible to predict whether such a rogue entity would be open or closed.
You ask what my standards for making that judgment would be, and if you're driving at the point that its a subjective estimation, you're right. Market position is very difficult to correctly ascertain, and I don't particularly think its the kind of thing either of us really wants to spend our lives doing. As a result, I hope you'll forgive me when I borrow the words of Anthony Comstock and give you the deeply unsatisfying answer that "I'll know it when I see it".
As far as it not having been settled, well, we're sitting here arguing about it. Sure, I'm writing this from a Linux machine running Firefox, but from your earlier statements I envision you doing the same from a Windows box and IE. With any luck, competition will help to improve the quality of all the contenders.
If you don't mind, I'd like to hear more about your experience with open source. What projects did you think were most or least competitive?
My point was that for most software, the OSS model will fail. The fact that it can work (and not demonstratably better than the closed source model) is pretty irrelevent.
This is your problem. Had you simply stated that this was your point all along, instead of pretending to be part of a reasonable discussion, we could have been done with this much more quickly. OSS has proven that it can at least enter most markets. The question of whether it can dominate those whose developer and user bases are disjoint has yet to be settled. Saying anything more -or less- is simply irrational.
Slashdot Mirror
huh. skip it and read the next one?
Got about 20 words into the first one and never found time to pick it back up again. Worth it?
Not to the best of my knowledge. Am I missing a meme?
and, having solved all OTHER problems, the brightest minds of our world turn to figuring out spit.
hell, meet handbasket....
IANAEM (I Am Not An English Major), but wouldn't that be "IAAB"?
I'm willing to try /What?
I'm sorry, but I'm forced to disagree with your characterizations of both McCain and Obama.
I spoke with McCain about a year and a half ago, before he was an official presidential candidate but long after he had decided to run. He gave what has now become his usual justification of his candidacy, followed by a brief discussion of the issues of the war in Iraq, the economy, school vouchers, torture, and state's rights (I live in South Carolina). I then asked him what, given the deeply conservative agenda he had laid out, would he say to those who had called him a maverick or independent? His response was as follows:
"I am sorry, but I am a deeply conservative candidate"
His voting record backs this up- where he breaks with the Republican party, he does not typically break left. He breaks in favor of what he calls 'higher conservative principles', and in practice this has nearly always meant going further right.
I consider myself an independent. I've supported candidates on both sides of the aisle for president and local office both, and have spoken (admittedly briefly) with both the men you are characterizing. Senator McCain is undoubtedly one of the most conservative candidates in recent electoral history, if such a label has any meaning anymore.
As for Obama, I am forced once again to disagree. You should know that I support the Obama candidacy.
I presume that when you say that Obama is further left than Hillary, you are relying upon the oft-quoted and little understood study conducted by the National Review (a publication that, incidentally, cannot by any reasonable person be called 'unbiased') indicating that Obama was the "most liberal Senator", with Hillary occupying the "16th most liberal Senator" spot. The methodology was simple: classify all substantive votes as either liberal or conservative, and tabulate the results.
As a person who fundamentally disagrees with the "you're a liberal or you're a conservative" zero-sum-game approach to politics, I have to disagree with this methodology. It is also worth pointing out that somehow John Kerry earned the coveted top spot on the study as soon as his candidacy was announced, making it harder to believe that the individuals deciding what constitutes 'liberal' and 'conservative' votes are looking strictly at politics.
As a second point, Obama has made it a cornerstone of his campaign that there is a need to end the process of politics as currently practiced. There are those who say that it's empty rhetoric, but to my ears those people aren't listening very well. It means that he has built his candidacy on how he will do what he will do as president, rather than what he will do. We have a President right now who, while undoubtedly flawed, has committed no political sin more grave than trying to go it alone. I might even advance an argument that without that, his presidency would have been disliked but not hated. This is a crucially important point: that while what you do in office is important, the Presidency is a politically weak office, but a highly visible one. How you do things is sometimes more important than what you do.
God must love idiots, because He made so many of them...
This is really freaking obvious. I wrote a p2p application that cached based on search requests and then fetched based on router hops years ago, and presumed it was nothing new then. I strongly doubt this will be an unencumbered technology if it ever sees the light of day.
I'm sorry, but you're discussing things in a reasonable and intelligent manner. Your Intertubes license is officially revoked. Turn in your e-peen at the counter.
We think it's a good idea because there would be no free alternative otherwise. That's important to a lot of people.
If you're putting a week's worth up and selling out in a day, that's not the end of the world. If you're putting two month's worth up and selling out that afternoon, you have a big problem.
It's ok, but it would improve if you sold Windows-based PCs instead Lot of competition in that space. On the low end machines, a quarter of the cost can be software licensing. My experience is that that cost has to be passed on to the customer, and that it leads to lost sales. Your mileage may vary, of course.
At the Wal-Mart scale, it's millions more to make by axing the Linux PC in favor of cheap, Windows-based machines, so it becomes a simple business decision I don't dispute that Wal-Mart is making a business decision here, or that it's a simple one, but I think people are overlooking the probable reason- Wal-Mart loves good suppliers, and hates bad ones. It wouldn't be the first time Wal-Mart axed a profitable product because the supply chain just wasn't able to support demand, and TFA clearly states that these repeatedly sold out.
Because you clearly do not understand the PC business, the article, or the events in question, allow me to point out a few small facts you appear to have overlooked:
1) PCs are a commodity business. You don't stick around selling 3 of them in a month.
2) If you're making $1000 off of 3 PCs, I want whoever your marketing guy is.
3) Wal-Mart doesn't order things in 50s. It orders them in thousands.
4) TFA clearly states that Wal-Mart repeatedly sold out of the machines.
Put it together. Wal-Mart has sold thousands of these machines out repeatedly- which means that it has a product whose supplier cannot meet demand. If you're a company that size and want to lose a lot of money, the way to do it is to have to deal with somebody else's god awful supply chain.
I sell 'em. Make a good bit of money doing so. Care to explain why my business is, in defiance of all evidence to the contrary, not viable?
This is mostly what I've been saying. I know I could eventually get what I need, but in today's computers and GUI applications, if I need to go as far as going to the source, the application has failed from a usability standpoint. That's my issue, and it's likely an issue for many others.
Depends on what you mean by 'usability'. End users should never have to go further than the forums or IRC. Administrators should never have to go further than the man pages. Developers should be looking at either API documentation or source, and preferably both. All of the above are addressed in a good program, and most of the *major* projects do so. You've already said that you're not really a fan of the major avenue of assistance in Linux, which is the community forums and IRC, and I guess if that's how you feel about it that's the way it is, but I don't think that without community assistance you're comparing apples to apples in terms of support.
I said that the options, which may be dangerous, are still accessable and that adequate warnings and explainations are given. Right there, in the dialog to set the option (or by pressing F1). I found that in Linux, I would find the option, but the various settings of that option aren't adequately explained where you set the option. You need to go digging. Again, it's a usablity failure.
I'm not sure if there's a particular piece of software than you're referring to here, but most man pages seem exhaustively complete to me, especially on major projects. /shrug.
Again, quite a lot of steps. Does Unison setup both client and servers? When building a new workstation do I need to use Unison to configure the client again, or is there a way to say "this server controls your settings" and leave it at that? I'm not sure I would have ever found or heard of Unison if not for this discussion... which again I think points to a problem.
If you already have rsync and ssh on your server, you just need to install Unison on the client, tell it which directories to sync which way (can be saved in a config file, which can also be shared safely) and point it at your server.
As for not having heard of it without my help, well- yeah. That's why the community exists. That's what I mean by living documentation. If you ask somebody a question- and we're easy to find- we'll answer if we can, and know where to look if we can't.
Hmm... options can be good and bad. In this case, I lean toward bad, because I now have to evaluate one of these packages before I can even start administering my network. If I choose poorly, I have to undo everything and switch to another package.
Hmmm. I'm afraid I don't agree with you here- I tend to believe that competition in a market segment tends to improve the quality of the competitors and increase the power of the consumer- but I guess to each their own.
No, my knowledge of it was that it is a scripting language (at least, when I first heard of it, it was) and I try to stear clear of those. Maybe it's changed, or maybe I'm confusing it with TCL/TK. I dunno, it's been a while. I tend to stick to languages that are similar to C / C++ sytax. I don't know why, but I've always liked it more than Basic, and other languages just haven't appealed to me as much. Not that I haven't used a few, at least a couple of times each.. lisp, snobol, eiffel, VAX assembly and some other's I'm forgetting. :-)
It's a scripting language in the same sense as lisp is- interpreted, quick to write, blurs the line between data and code- but it's pretty common for a lot of general purpose programming anymore. A lot of the syntax is pretty much C++ sans punctuation, which I find handy, but I figured since you mentioned both .NET and Java that I would mention Python since the three major implementations of Python are C, .NET CLR, and Java.
By the way, I'm not continuing to post to cha
So you've never really used Windows then (the NT line.. 2000 or higher)? I've found lots of people say Windows is for morons or its overly simply. The fact is that it's complex, and much of the complexity is hidden. But its there if you really need to dig in. Given that I've seen many comments about virtualization and supercomputing that mirror yours, I'm not suprised that Windows doesn't fit your needs. I would check back now and then though (maybe now, because Server 2008 is out), because MS is trying to move into these areas. I've found that they usually succeed; not on the first or second attempts, but they have a good record. I know.. that's not what most /.ers will say.
No, I've used Windows extensively. First, my high school days aren't all that far behind me, and XP was already out. Second, as part of my business- many legacy business apps cannot be ported without monumental effort, and so it is frequently the case that we set up a Windows box with a Linux passthrough.
As for Windows supercomputing, I wish them great good luck but don't see it happening, as the UNIX world's tools for HPC etc are just incredible. It would take a LONG time and a lot of money to come up with a comparable feature set.
Well, you've probably had much more time. All of my Linux experience was me, all by myself, trying to be inbetween. Wanting to administer my own network like an expert would, yet not having the resources provided if I worked with Linux all day. Again, maybe you're right about home users, and my experience is because I'm in this weird limbo. As far as making unsafe options available, well, we're not talking about a gun, so lets not go down that route. You can reformat a machine that gets owned, you can't bring someone back from the dead. Also, if the option is adequately explained and the security implications are stated, I don't see what the problem is. In the end, you can set that unsafe option on either system. The problem with Linux seems to be they are just hiding it, so if you find the option you may be less aware of the implications. Windows points it out, saying "hey, this could be really bad, here's why, but if you want, go for it." As someone trying to learn more about administring systems, I find that is more helpful.
Its all there, it just happens to be in either the man files or the ultimate documentation. Once you learn how to use a command's basic invocation (from irc or wherever), then check the man pages for reference documentation, and you're generally fine. Again, all of this is anecdotal- my experience is x, yours is y, ergo, we run around in circles, but like I say, between talking directly with the developers, reading the man pages, and, if you're really masochistic, perusing the source code, you'll figure it out. In the end, there's really nothing in Linux that's totally opaque, which is, once again, a major advantage for those willing to take advantage of it.
As far as security is concerned, we probably don't want to get into that flamewar, but there are a lot of things you can do to a system that simply can't be undone. If your customer database gets cracked, its game over. If someone installs a malicious utility, its game over. God forbid they be able to insert a kernel module. The assumption that box security is more important that infrastructure or data security is one of the major problems with home use security models, and really highlights a major difference in approach between Windows and Linux. It doesn't help that much of security research is very convoluted stuff, and that many elements of a security strategy cannot be considered in isolation. So, I can't really agree that every insecure option should be just as easily available as any given security-neutral option, any more than I think that every option available on the CLI needs to be available in the GUI.
Your suggestion is to send out configuration files via rsync? Fair enough. Seems like this should be made easier to do tho
I understand your experience differs, and that's fine as you said. However, a couple of things. Did you learn much of what you needed for Linux because you have a job doing so? I'm certainly not afraid of Linux by any means. Its just that its more difficult to learn. I do fine with shells and what not, after all I used Unix all through my five year college program. That's when I first came into contact with Linux as well. "Unix on my PC, awesome!"
My areas of interest in computing are supercomputing, CPR, and (more recently) virtualization. Windows does not give me the kind of power as a developer that I need to pursue my research in those areas. After a brief flirtation with Solaris when I was in high school (the school had a Sun server), I moved to Linux full time and haven't looked back.
Your user experience is limited (it seems) to business users. That's fine as I said, but I don't think that the experience a business user of linux (who likely doesn't manage the computer they use in any way) translates to a home user well (because they have to "manage" the system themselves).
We set up all kinds of systems, in all kinds of offices (including home offices), but most of them are not centrally managed, and in many of them- especially former Mac shops- the users have a tremendous degree of latitude in what they do and how they do it. You are absolutely right that the comparison between a large corporate desktop and a standard home machine is pretty much apples and oranges, but our installs on a lot of machines are basically stock Ubuntu, Mint, or (if requested) Fedora. Nobody seems to have a big problem, and in fact many people only notice that the background has changed and that the start button is in a different place. After that, most don't *seem* to care, although of course they know me as 'the Linux guy', and people aren't usually in a hurry to criticize others' livelihoods. Again, however, I have no data with which to back up my claims, so it could just be my admittedly biased perception.
In that sense, market share IS important, because home users also add hardware to their PC. You can't just buy anything and assume it will work on Linux.
The only devices I've had real problems with on Linux- besides really strange things like one-off industrial controllers- are blackberries. They really just don't work the way they're supposed to, although the Barry project tries really hard. Besides that, most hardware does work now, and again, I have nothing but anecdotes to hand you.
As far as /home being on an NFS mount; I think that was recommended when I first setup my server. At any rate, the idea was to keep user data on the server. I understand WHY everything locked up, but it didn't lead to a great experience. Contrast this to Windows; a few group policy settings (which affect more than the location of user documents) and user files and settings are stored on the server. With Windows, its smart enough to use the local store and sync later. No configuration needed.
Well, its not the lockout issues I was commenting on so much as security, since /home is frequently scanned by UID 0 processes and UID 0 access is badly handled by NFS, leading to serious security issues- but then I'm the guy who includes TEMPEST scanning as part of a security audit. The better solution would probably have been SSHFS overmounting a known safe /home on the local machine. It probably would have spared you a lot of headaches and would definitely have been more secure. It is also *far* easier to set up, since it requires less in the way of cooperation from your server. Another possible route (and much more comparable to Windows' group policy) would have been rsync, which is quite easy to set up.
You've said several times now that you think that Windows is easier to use than Linux. I don't think that's correct at very many levels of user experience, but again, we're throwing anecdotes
Sounds good. Presuming that there is merit in the methodology here, we may see some more competition in the market. I'd like it more if there were more code sharing going on, but maybe if we don't get tempers too hot we'll see a little more of that too.
I may be being stupid here, but I'm trying to follow this and not getting it. Please correct where my chain of logic is going awry.
1) the point of GPL that you can't distribute code that's derived from GPL code under anything other than the GPL.
2) Linking against code is considered 'derivative'. Using a public interface is not.
3) NDISWrapper links against the GPL'd kernel.
4) NDISWrapper is required to release under GPL and does so.
5) NDISWrapper uses the public interface of the binary driver
6) This imposes no license requirements on NDISwrapper
7) No license violations have occurred.
Even if Linus doesn't like it, it doesn't seem like there's anything in the GPL that specifically stipulates that you can't both link against GPL'd code and use the public interface of non-GPL'd code, and if there is, it will cause a LOT of problems in LinuxLand. Am I wrong?
Well, you're giving them training. Are you also taking over managing the servers from then on? As I attempted to be a self taught admin (somewhat) sometimes I didn't care about the details of why something worked, I just wanted it to work. I can dig into the whys later. I found on Linux, I had to spend an inornate amount of time to figure out it, and the why details usually drowned out the how.
We do give them training, but we don't take over management, preferring instead to offer break-n-bake support plans and service guarantees. We do both desktops and servers, and training for both IT staff and end users. We find that end users typically cause no more problems on Linux than they do on Windows after an adjustment period, and depending on the setup, sometimes less. IT is always a hassle, but that's just the nature of the beast- people do not like retraining away from a skill set that they have based their careers around.
You are correct that Linux as a server OS does not encourage you to perform operations you do not fully understand. Linux on the desktop, I think we may have to agree to disagree. My personal experience (no more or less valid as evidence than your experiences, of course) is that we have few problems with desktop installs, and that what problems we have usually come down to unsupported hardware (industrial controls, etc), and even that has become quite infrequent now. To the best of my knowledge, we haven't had any major complaints about the (admittedly limited) applications we install.
Which is why I don't think OSS will get anywhere; someone HAS to be trained, they can't just sit down with very basic knowledge and off they go. In a business setting, you usually have IT to call on to answer questions. Home users don't, and they sometimes find certain tasks in Windows hard to do. Imagine home users now trying to muddle through Linux.
Again, I think we may have to agree to disagree. I don't think that very many people would have argued with you a few years ago- I know I wouldn't have- but I also think that OSS has come a long way since then. Of course, at this point we are using OSS as a synonym for Linux, which I in particular ought to know better than to do, but I hope you'll understand if I use it as shorthand where I don't otherwise elaborate.
My personal experience (again, not valid data) is not broadly representative. It is based on our service to a region that is both desperately poor and possessed of an educational system that would shame Somalia. Most of our end users treat computers as bewildering, terrifying, and capricious lesser gods, whose power to help you is only exceeded by its ability and desire to do you harm. They have little Windows experience, and have never heard of Linux or, indeed, an operating system, before. They don't understand, and don't care, as long as it keeps the gray box happy. We have little difficulty in training them to do basic office and worksite tasks on Linux, and life goes on. Our more experienced Windows users are another story. They do not like to retrain, and frequently try to treat Linux as "Windows for cheapskates". They run everything as root to avoid the inconvenience of passwords, they complain bitterly that things are not in exactly the same place "as they used to be", and they generally experience higher levels of frustration and greater difficulties than the above mentioned group, even if they do not do significantly more with those machines. I suppose you can turn that into an argument for Windows- that people are already trained on it and do not need retraining- but while that makes a good business case I don't think it addresses the fundamental question we're talking about, which seems to be the quality of the software in question.
If the project continues, and statistically speaking no one is using it, does it matter? There's no competition at that point.
I think I wasn't as clear as I wanted to be. When I say competitive, I mean feature-for-featu
Part of my point is that the FF team has a big player backing them with money. Should that company no longer fund the primary developers, would FF survive? Would the project continue? Yes. Would it have anywhere near the market share it does? No. But the amount of cash involved has nothing to do with whether a project is open source or not.
GNUCash is a more appropriate model; the developers think I want all the gory details of double entry accounting, which I don't. Things like MS Money or Quicken hide those details for me. GNUCash also was lacking in any reporting of any kind. They even admit it: "I'd guess that GnuCash's biggest problem is that programmers don't use the software." These are the majority of OSS projects. Again, I'm sorry to hear that you had a bad experience, but part of what I was saying earlier is the fact that the vast majority of software is bad software, open or closed source. For every bad open source program that you point out, I can point out a bad closed source program, and we're back to the "my anecdote is better than your anecdote" game. In the absence of data, we are just opining.
Well, the problem is that without objective qualifications we can't really ever evaluate if your statement is true. As I've tried to carefully explain *we*have*no*data*. All we have are perceptions and opinions. I, naturally, think my opinion is correct; that the open source model provides specific advantages, that much open source software is of high quality and that it is in many respects a level playing field, with the leaders in some market segments being closed source and the leaders in others being open. You, naturally, think yours is correct, and while I disagree with it, that fact doesn't make it any more or less valid. Well, I imagine we could forever too, since we don't have anything objective to go on. Given the level of sarcasm, I presume you have a nice simple 10-or-so word definition of 'good software' that I can whip up a test for and be done? I didn't think so. People who live in glass houses...
It was harder to use. GNUcash wasn't cutting it (no reports), I still had random crashes. The documentation is non-existent in most cases, and managing my network is easier with Windows Server 2003 (SBS). And of course RPM hell. I felt like I was spending more time fighting to get Linux to do what I want than actually having it do what I wanted. And I put the time in too; I had run a linux server for 10 years and on the desktop for three. It was honestly more frustrating than Windows. I'm not sure about GNUcash, as I don't use it (we have an accountant for that) but I'm amazed that you had so much difficulty with the desktop, given that level of experience with Linux. We get individuals who have no technical background to the point where IT calls are down to normal within a few days, and with one major exception have never had a persistent problem. Any issues in particular? How long ago was this?
You wouldn't happen to know of any good, accessible resources for somebody just getting into object recognition would you? So far I seem to have a knack for selecting the most obtuse references I can and it's starting to pick my pocket- any help would be much appreciated.
I'm not your strawman, so put down the torch. I don't think this is the year of the Linux desktop, and have been quite explicit in saying that I think that OSS is very good at some things, while closed source models seem to do better at others. I think that many Linux desktop applications are at or approaching the point of competitiveness with their closed-source brethren, but I also think that its going to be a long time before some parts of the industry come under contention.
My initial question was simply whether OSS will wind up dominating arenas with disjoint user and developer bases the way that it does where those two groups are closely aligned, and you ask when we'll know the answer to that. My general feeling is that we already know that it is possible for projects like Firefox to contest some segments of the market, but that we probably won't see just one model in very many parts of the industry unless an application that simply obliterates its competition winds up emerging, and I don't think its possible to predict whether such a rogue entity would be open or closed.
You ask what my standards for making that judgment would be, and if you're driving at the point that its a subjective estimation, you're right. Market position is very difficult to correctly ascertain, and I don't particularly think its the kind of thing either of us really wants to spend our lives doing. As a result, I hope you'll forgive me when I borrow the words of Anthony Comstock and give you the deeply unsatisfying answer that "I'll know it when I see it".
As far as it not having been settled, well, we're sitting here arguing about it. Sure, I'm writing this from a Linux machine running Firefox, but from your earlier statements I envision you doing the same from a Windows box and IE. With any luck, competition will help to improve the quality of all the contenders.
If you don't mind, I'd like to hear more about your experience with open source. What projects did you think were most or least competitive?
My point was that for most software, the OSS model will fail. The fact that it can work (and not demonstratably better than the closed source model) is pretty irrelevent.
This is your problem. Had you simply stated that this was your point all along, instead of pretending to be part of a reasonable discussion, we could have been done with this much more quickly. OSS has proven that it can at least enter most markets. The question of whether it can dominate those whose developer and user bases are disjoint has yet to be settled. Saying anything more -or less- is simply irrational.