Mozilla/Gecko overused XPCOM heavily, which added bloat and was a performance/code-complexity hit. For instance internal to the rendering engine you don't really need it, it doesn't give (m)any advantages over straight C++. Nobody in their right mind would try to write a rendering subsystem in JavaScript, and anyway you can't, Gecko isn't a very pluggable renderer. You need to special case code like I think nsCSSFrameConstructor if you want to add support for new document rendering types (unless you use XBL).
The fact is that the parent poster doesn't know what he's talking about.
The reason SVG isn't included in the default build is nothing to do with "politics" unless you have a very broad definition of the term, it's not in because it's not complete.
Netscape/Mozilla have been burned before when they included half-assed support for a standard. It's bad for a ton of reasons. People don't know what features they can use and what they can't, if mistakes are made they get frozen into the defacto standard and so on. So, until Mozillas SVG support matches a W3C standard, it won't be switched on.
The main problem is that SVG is really huge and complicated. I think last time I checked they were aiming for "SVG Static" which is a cut down version (no animation for instance). Because that's also a recognised standard they could switch it on at that point.
I don't know how Konquerors SVG support matches against Mozillas, but I'd be surprised if they'd implemented the whole thing (with the required KHTML/DOM integration). If they haven't done the whole thing then I'd not suggest they switch it on, it's that simple.
MNG support was dropped because MNG is another huge, (bloated?) spec. It's not just GIF-with-PNG you know. If anything it competes with Flash. The code for it was huge and it the person who owned the relevant module didn't care about it, so it got dropped. Now, whether you agree with this decision or not is somewhat irrelevant, you aren't the maintainer of that part of Mozilla (feel free to fork the beast). You have to question though - if MNG had been 100x simpler it'd probably still be in there today. As it is, nobody uses MNG at all.
Yes, but only Windows XP has things like the Secure Audio Path (SAP) in which audio passes encrypted into the kernel. IE the only way to rip is to have two computers or a minidisc player etc (as the drivers disable recording at the same time as playing in DRM mode).
It looks like it uses WMA, which has fairly good DRM. Screws people who aren't on Windows 2000/XP of course, but I guess they consider people using Windows 98, Linux or MacOS not mainstream enough (or more likely, the underlying OS not DRM-secure enough).
I'm 19, and I'm lucky enough to (currently) be working part time for a well known free software company.
I'm still at university, in fact, I'm in my first year. While I've never had a letter like this, my parents have of course raised the same issues. This is what I told them. Hopefully it will be of use to other young people in my situation.
The fact is, that making money by selling software is hard. Damn hard. Even if you just work every day for a paycheque and go home at 5pm after you added a new feature to Photoshop, you're in the minority. Most programmers (I've seen statistics that say 80% but I have no idea how accurate that is) don't write software to sell, they write software to solve peoples problems.
Let's review why writing software and selling it is hard, from the perspective of the guy who had the idea and is trying to capitalize on it rather than the 9-5 hired hand.
Firstly, it's not just a matter of writing a program and sitting back while the cash rolls in. You are expected, at minimum, to release new versions every so often, have professional packaging, probably you will be required to support it and deal with the random problems people come to you with. This is not a short term commitment. Your software may be around and have users for years. In other words, selling software requires a considerable investment of effort and time.
Market conditions in software are not favourable. Software competes on a global market - this isn't a grocery store you're running. If the guy on the other side of the planet has a better product for a better price, you are in direct competition with them. It's not even a case of better product better price often - you think you can write a better word processor than Word? Go for it. Just don't expect to sell more than a few copies even if it is better. Life isn't fair, and the "unfree market" especially so.
No. Why would I want to work day after day on the same product, being a cog in the machine? I want to try for a better way.
What I'm currently valued for is not what I've written, you see, but what I can write. When people hire me (and I've worked for quite a few well known companies by now), they are hiring my knowledge and expertise which I sell to them typically at an hourly or monthly rate.
They purchase my skills because I can solve their problems. Ultimately this is what it's all about. One way programmers can solve peoples problems is by writing a product, setting up retail channels and then hoping that enough people have the same problem that they can strike it rich, but this is a high risk endevour and I'm not naturally somebody who likes high risk. I'd rather go to them directly (or in the case of one of the last jobs I did, went to a consortium of people), and solve their problem directly then move onto a different problem.
This is how I intend to earn my living, and so far it's working out pretty nicely.
Claiming that software has to be proprietary, that it has to be bought and sold as if it were a physical thing is a gross distortion of both economics and common sense. People tend to look at software as a machine, as a black box, and so it's natural to draw an analogy to a physical thing (hence copy protection) but really it's little more than a series of instructions for how to solve a problem.
If you asked me, "How do I make a pasta bake?" would I write down a recipe for a pasta bake then sell it to you on the condition that you didn't give anybody else a copy? Would I try and sell that as a physical product? Of course not. Just phrasing it in english and writing it down doesn't make it a product. It's simply an encapsulation of knowledge.
A better idea, if people ask you that often, is to teach people cooking, or alternatively become a chef, ie people pay you to excercise your skills (cooking) to solve their problem (hunger) and if you happen to invent new recipes and share them with fellow chefs at the same time then so what? Nobody loses.
... In the usability realm, open source software on any platform often sucks.
So? A lot of commercial software does too. Have you ever seen the usability halls of shame? Check out this one and this one. Notice all the stupidities that are allowed into commercial software, especially Microsofts. You'd think with their budgets they could hire UI experts, but apparently not.
My point is that usability is hard, and you can't make sweeping generalizations like "open source has bad usability, proprietary software has good usability".
OpenOffice, Evolution and Mozilla have completely different interfaces from each other, in fact there are few simularities among them.
Erm, because they do different things?
Perhaps you mean the widget toolkits they use. Let's see what the competition has to offer:
I don't know about you, but to me those apps all look entirely different. They use different artwork and even widget sets! Yet, they are from the same company and are all flagship products produced within the last few years. Microsoft don't even have history as an excuse. Apple are just as bad - Aqua today, brushed metal tomorrow.
So, I don't see what your point is. Given that FireFox and Evolution both use the same widget themes, and recent builds of OpenOffice can do the same trick, it looks like in terms of UI and widget consistancy Linux beats Windows hands down.
That's wrong. The 80% figure appears doesn't actually appear in the Gartner report and seems to have been pulled out of thin air. In fact, little to nothing has been confirmed or is known about the whole operation....
..if Linux really is to blame (and I haven't seen any specifics on what problems they are having), then they can fix them themselves
The little information the community has about this seems to be that the main problems are purely related to the monopoly status of Windows, ie app compatibility, lack of knowledge, admins who don't want to have to retrain, infighting and so on... not much that can be fixed with changes to Linux, it's purely a matter of economics and inertia
What do Munch and Ballmer expect? Moving from one version of Windows to another has been the cause of god only knows how many IT budget overruns, and surprise surprise, they are finding that moving everything to Linux is not easy.
Well, duh. That's why Microsoft has a monopoly, right? Ballmer likes to bitch and whine about how it was a "political decision" and how such things are somehow dirty and rare, but he seems to have missed the fact that every decision is political. There's no such thing as a pure business decision.
No matter how many TCO studies you do, no matter how many reports are written by an IT dept doing an evaluation, the final decision is going to be made based on how comfortable somebody is with an idea. Going with Microsoft is safe, it's easy, because everybody else does it. That's a political decision. It's the old "nobody got fired for buying IBM" thing.
The problem with Ballmer is that he sees what he wants to see. Somehow he has to reconcile his beliefs (that Microsoft is better) with reality (people are chomping at the bit to leave them). He does this by saying:
The people who are making business decisions based on where are the applications, what is the value, what is the lowest cost of ownership, we're not losing them.
... while apparently ignoring that TCO includes future costs such as forced upgrades, complying with license audits, working around the inflexibility of their software and so on. The hard to value, intangible costs. So he smears former (and possibly) future customers by writing off their decisions as "political" - a thinly disguised euphamism for "irrational".
There's another thing. Does anybody else have questions about the competency of the Munich guys to be doing such a migration? Why are they doing a crash switch, which is bound to end in tears? Why are there persistant rumours of them using VMware rather than bringing Wine up to speed on their products (which I'd guess works out cheaper in the long run and certainly provides a better desktop experience).
Finally, is it just me or does Ballmer look really evil in that photo?
Erm, if you read the whole mail you'd realise that was kind of a joke. And the developer agreed that there was a problem, but wasn't sure what the solution was.
He didn't write CUPS so why should he feel obligated to fix it? He's a USER. He didn't write the code. He didn't design the interface. As a USER he's in a position to criticize. It's what users do.
Well, no. That's not what users do. At least not the kind we want.
Smart users write thoughtful, constructive criticism, file bugs, suggest better (in their opinion) UI designs and so on. Stupid users write rants on their webpage about how bad free lunch X/Y/Z is and how the authors OWE it to them to answer their every whim.
What a lot of people seem to forget here is that Eric S Raymond did not pay for CUPS, nor Fedora. That changes the rules a bit - it means you can't just insult the people who made it and tell them they suck. Instead, he should act politely and with civility. Until he does that, I don't see anybody rushing to address his concerns.
Now, you might say "but then Linux will never overtake Windows". But you'd be wrong, because the free software community has made great strides in usability and UI design lately - and it's not because of people like Raymond or Zawinski spouting off, it's because people who care sat down and got things done.
By broadly categorizing "users" as assholes you've not only insulted a great many people, but you've also painted a woefully inaccurate picture.
Hmmm, so you're saying that a company should be allowed to flagrantly break the law for marketing reasons and get away with it because a break-even online shop might make Eminem some money, maybe, one day, if he's lucky?
Make a significantly better product and communicate this to your target market. Do this, and you'll win.
You seem to have overlooked the fact that to be significantly better than Windows requires being able to run Windows software too (without a copy, otherwise that kind of defeats the point).
Doing that requires cloning the Windows API, which is a task of such magnitude that no corporate entity can do it. Eventually, as many programs expect these things to be there, it'll require cloning Internet Explorer and maybe WMP as well.
You brush such minor details aside as if it's merely a case of "come on chaps! let's make a better mousetrap and it'll all sort itself out!", when that clearly isn't the case.
The same thing happened in the US after the first Microsoft antitrust case. They were ordered to produce a version of Windows 95 (IIRC) without IE. They did, it didn't work properly and nobody wanted it.
Yeah, I don't really see what Slashdot finds so hard to understand about this. Unintegrating IE now is, quite simply, impossible. Microsoft didn't lie when they said that was the case
Do you guys have any idea at all of how many apps expect Internet Explorer and related DLLs to be installed? Working on Wine brings this point home in a really fundamental way.
Oh sure. You could remove iexplore.exe. That would remove like 0.1% of IE from your system. It'd be a pretty hollow gesture.
I'm not just talking about things like the MSHTML component. I'm talking about things like the SHLWAPI DLL - a utility library which wraps the Win32 API to some extent developed by the IE team partly to make portability between Win31, Win2K and Win98 simpler. It's only half documented, a lot of the functions are exported only by ordinal, yet a surprising number of programs expect it to be there.
What about the URL monikers implementation? What about all the installers that assume the presence of Favourites? What about all the programs that embed the Trident engine to render parts of their UI, their online help - in the case of one game that shall remain nameless even the games main menu!
Windows shipped without Internet Explorer would effectively break so many apps nobody would buy it, even if they could. Quite a lot of apps don't even complain, they just crash. Win95 not supported.
Now, this stuff is mostly academic. Shipping Windows without IE on the desktop would have made a difference - 5 years ago. Nowadays many (most?) people have never heard of Netscape, think that the Internet is the blue E icon, and so on.
The only way IE will ever disappear in other words is when Linux starts kicking Windows' ass on the desktop, which last I checked was still a year or two away just on the corporate desktop let alone the home user desktop.
If anything, Linux is worse. Linus has stated numerous times he makes no effort whatsoever to retain binary compatibility even between minor point releases of the kernel.
You're not distinguishing between Linux the kernel and Linux the OS. The parent was almost certainly talking about the latter.
Userland backwards compatability on Linux is OK but we've certainly had our fair share of cockups. The rollout of the new threading systems (NPTL and the new TLS system) was pretty much a backwards compatability disaster. I currently have to run XMMS of all things with LD_ASSUME_KERNEL because of NPTL. Oh sure, I tried to debug it. Doesn't work when gdb pukes and dies - again due to threading. Don't even get me started on the breakage Wine has had to deal with.
The glibc/kernel guys claim they know how to write backwards compatible software but in reality they don't. They don't, because unlike Microsoft they treat backwards compatability as a science, as a fixed set of rules that if they follow they think can be held blameless. Of course when you get situations like NPTL where the old system was so broken everybody had workarounds which stop working in obscure ways when the underlying bugs were fixed, this logic breaks. They still break backwards compatability, they just end up playing the blame game instead, which is stupid.
We could have much better backwards compat without the huge hacks Microsoft use with even a few small changes to process, but I'm not seeing people interested in making those changes.
Given the intensity with which it is probably being scrutinised, I'd imagine any "obvious" embarassments would have already surfaced by now.
Hell, if you want a laugh (and if you're a win32 dev you will learn something too) go read Raymond Chens blog. Not only is this guy a near-genius level coder, but he's been working on Win32 (and USER in particular) for a very long time now. Yes, I know some of you think that's an oxymoron. He often posts interesting stories about its development and about the bizarre hacks they put in to work around broken apps (some parts of windows even go so far as to detect and correct stack corruption).
Can you please join #autopackage on freenode during GMT hours? I'd like to get to the bottom of this. It sounds like you have a broke libvorbisfile, so what distro is this?
Apparently they are going to import the XFree DDX wholesale - hard but doable. This is just going from the words of one guy though. I'd want the Word Of Keith before believing it myself.
And then the engineers will go "Gah, this binary API is changing *all* the time. Can't we simply release a GPL driver and be done with it?" And then, after a little talk with legal, and probably the PHB again, hopefully it'll happen.
I think that's pretty naive. Alan Cox himself has said he can't think of a good argument to make nVidia release their drivers - their fears as to 3rd party IP rights and performance relative to competitors are, apparently, justified.
In much the same way that it's unrealistic for Linux to not be able to run proprietary software, it's unrealistic to not be able to (adequately) use proprietary drivers. While I'd love for all drivers to be GPLd experience has shown that making the installation annoying for the end user does not lead to more GPLd drivers.
At some point, somebody will simply take the matter out of Linus' hands by writing a module that provides a stable API/ABI to other drivers. Sure, it'll be less flexible than the one available to GPLd drivers, but they'll have to live with that.
That's wrong, Wine is a compatability layer for Windows programs. Programs that only run on 2K or XP are harder to emulate because by definition they use more functionality than your average app which restricts itself to only features available in Win98, but there's no fundamental reason why they can't be run too - and in fact they are run all the time.
Wow. That has to be the longest post I've ever seen on Slashdot! Let's see....
* Software Packaging. This is a huge pain in the ass for most commercial vendors of any Linux software
Actually it's not so bad - commercial software tends to be mostly statically linked or to come with the right DSOs included. Programs like Loki Setup or BitRock can make pretty NSIS/InstallShield type installers.
The problems we have with software packaging are mostly problems for open source projects.
* Linux Binary Portability. This is the bane of all vendors releasing any closed-source software for Linux.
I'd venture that this isn't as problematic as you think it is - having done a ton of research into this, Linux binary portability is not really any worse than the difference between Win98 and WinXP, for instance.
Yes, the rollout of NPTL is being done badly, and this is unfortunate, but the vast majority of apps don't suffer any problems at all.
We're really nearly there.... the default/stock Fedora Core install has built in CD writing - just insert a blank CD, drop files into the folder that pops up and hit "Burn" (built into gnome). Couldn't be easier. Right click ISOs to burn them. Just need to nail burning audio CDs in an integrated fasion, maybe with RhythmBox.
Totem is a great DVD player. The only problem is that it doesn't work out of the box thanks to the DMCA problems with libdvdcss. I think a better "start here" trail for new users would help this, as it's not hard to install the needed RPM (just download and click it, really).
2. Friends who are familiar with the OS/Distro
I think this isn't so bad. You know, nearly everybody I talk to at my university has heard of Linux. I mean, 99%. Most of them know a friend, or friend-of-a-friend, who uses it.
7. Some packaging system with less dependency problems. [Yes, there are a few that show very good promise, with only occasional issues surfacing.]
Yup!:) Autopackage 0.4 will be out in the next week, promise!
The compatibility of programs is much better between os x/linux then it is between os x/windows.
Not really. What you mean is software developed for Linux can run on both Windows and MacOS X without too much difficulty, because they are typically based purely on free software. I'd note that while some Windows apps can be run on Linux, no MacOS X apps can be run on Linux, nor on Windows.
So it's better to say that Linux software is more compatible, than to bring MacOS into the picture (it doesn't even have dlopen as part of the core, just like Windows doesn't).
I find it sad that the Mac's marketshare is represented so low, but I find OS X and Linux users on the same side of the bigger war, and the enemy of my enemy is my friend. w00t!:)
Ever considered that some people don't see it that way? A mac is just like a PC running Windows in terms of economics and philosophy, swapping Bill Gates for Steve Jobs isn't a useful trade to make. So in reality it's free software, vs non-free software (or platforms, to be more accurate).
People who think it's Windows vs everything else are just shortsighted IMHO
Slashdot Mirror
Mozilla/Gecko overused XPCOM heavily, which added bloat and was a performance/code-complexity hit. For instance internal to the rendering engine you don't really need it, it doesn't give (m)any advantages over straight C++. Nobody in their right mind would try to write a rendering subsystem in JavaScript, and anyway you can't, Gecko isn't a very pluggable renderer. You need to special case code like I think nsCSSFrameConstructor if you want to add support for new document rendering types (unless you use XBL).
The reason SVG isn't included in the default build is nothing to do with "politics" unless you have a very broad definition of the term, it's not in because it's not complete.
Netscape/Mozilla have been burned before when they included half-assed support for a standard. It's bad for a ton of reasons. People don't know what features they can use and what they can't, if mistakes are made they get frozen into the defacto standard and so on. So, until Mozillas SVG support matches a W3C standard, it won't be switched on.
The main problem is that SVG is really huge and complicated. I think last time I checked they were aiming for "SVG Static" which is a cut down version (no animation for instance). Because that's also a recognised standard they could switch it on at that point.
I don't know how Konquerors SVG support matches against Mozillas, but I'd be surprised if they'd implemented the whole thing (with the required KHTML/DOM integration). If they haven't done the whole thing then I'd not suggest they switch it on, it's that simple.
MNG support was dropped because MNG is another huge, (bloated?) spec. It's not just GIF-with-PNG you know. If anything it competes with Flash. The code for it was huge and it the person who owned the relevant module didn't care about it, so it got dropped. Now, whether you agree with this decision or not is somewhat irrelevant, you aren't the maintainer of that part of Mozilla (feel free to fork the beast). You have to question though - if MNG had been 100x simpler it'd probably still be in there today. As it is, nobody uses MNG at all.
Yes, but only Windows XP has things like the Secure Audio Path (SAP) in which audio passes encrypted into the kernel. IE the only way to rip is to have two computers or a minidisc player etc (as the drivers disable recording at the same time as playing in DRM mode).
It looks like it uses WMA, which has fairly good DRM. Screws people who aren't on Windows 2000/XP of course, but I guess they consider people using Windows 98, Linux or MacOS not mainstream enough (or more likely, the underlying OS not DRM-secure enough).
Hey, 25 isn't bad ;)
I'm 19, and I'm lucky enough to (currently) be working part time for a well known free software company.
I'm still at university, in fact, I'm in my first year. While I've never had a letter like this, my parents have of course raised the same issues. This is what I told them. Hopefully it will be of use to other young people in my situation.
The fact is, that making money by selling software is hard. Damn hard. Even if you just work every day for a paycheque and go home at 5pm after you added a new feature to Photoshop, you're in the minority. Most programmers (I've seen statistics that say 80% but I have no idea how accurate that is) don't write software to sell, they write software to solve peoples problems.
Let's review why writing software and selling it is hard, from the perspective of the guy who had the idea and is trying to capitalize on it rather than the 9-5 hired hand.
Firstly, it's not just a matter of writing a program and sitting back while the cash rolls in. You are expected, at minimum, to release new versions every so often, have professional packaging, probably you will be required to support it and deal with the random problems people come to you with. This is not a short term commitment. Your software may be around and have users for years. In other words, selling software requires a considerable investment of effort and time.
Market conditions in software are not favourable. Software competes on a global market - this isn't a grocery store you're running. If the guy on the other side of the planet has a better product for a better price, you are in direct competition with them. It's not even a case of better product better price often - you think you can write a better word processor than Word? Go for it. Just don't expect to sell more than a few copies even if it is better. Life isn't fair, and the "unfree market" especially so.
No. Why would I want to work day after day on the same product, being a cog in the machine? I want to try for a better way.
What I'm currently valued for is not what I've written, you see, but what I can write. When people hire me (and I've worked for quite a few well known companies by now), they are hiring my knowledge and expertise which I sell to them typically at an hourly or monthly rate.
They purchase my skills because I can solve their problems. Ultimately this is what it's all about. One way programmers can solve peoples problems is by writing a product, setting up retail channels and then hoping that enough people have the same problem that they can strike it rich, but this is a high risk endevour and I'm not naturally somebody who likes high risk. I'd rather go to them directly (or in the case of one of the last jobs I did, went to a consortium of people), and solve their problem directly then move onto a different problem.
This is how I intend to earn my living, and so far it's working out pretty nicely.
Claiming that software has to be proprietary, that it has to be bought and sold as if it were a physical thing is a gross distortion of both economics and common sense. People tend to look at software as a machine, as a black box, and so it's natural to draw an analogy to a physical thing (hence copy protection) but really it's little more than a series of instructions for how to solve a problem.
If you asked me, "How do I make a pasta bake?" would I write down a recipe for a pasta bake then sell it to you on the condition that you didn't give anybody else a copy? Would I try and sell that as a physical product? Of course not. Just phrasing it in english and writing it down doesn't make it a product. It's simply an encapsulation of knowledge.
A better idea, if people ask you that often, is to teach people cooking, or alternatively become a chef, ie people pay you to excercise your skills (cooking) to solve their problem (hunger) and if you happen to invent new recipes and share them with fellow chefs at the same time then so what? Nobody loses.
So? A lot of commercial software does too. Have you ever seen the usability halls of shame? Check out this one and this one. Notice all the stupidities that are allowed into commercial software, especially Microsofts. You'd think with their budgets they could hire UI experts, but apparently not.
My point is that usability is hard, and you can't make sweeping generalizations like "open source has bad usability, proprietary software has good usability".
Erm, because they do different things?
Perhaps you mean the widget toolkits they use. Let's see what the competition has to offer:
I don't know about you, but to me those apps all look entirely different. They use different artwork and even widget sets! Yet, they are from the same company and are all flagship products produced within the last few years. Microsoft don't even have history as an excuse. Apple are just as bad - Aqua today, brushed metal tomorrow.
So, I don't see what your point is. Given that FireFox and Evolution both use the same widget themes, and recent builds of OpenOffice can do the same trick, it looks like in terms of UI and widget consistancy Linux beats Windows hands down.
That's wrong. The 80% figure appears doesn't actually appear in the Gartner report and seems to have been pulled out of thin air. In fact, little to nothing has been confirmed or is known about the whole operation....
The little information the community has about this seems to be that the main problems are purely related to the monopoly status of Windows, ie app compatibility, lack of knowledge, admins who don't want to have to retrain, infighting and so on... not much that can be fixed with changes to Linux, it's purely a matter of economics and inertia
Well, duh. That's why Microsoft has a monopoly, right? Ballmer likes to bitch and whine about how it was a "political decision" and how such things are somehow dirty and rare, but he seems to have missed the fact that every decision is political. There's no such thing as a pure business decision.
No matter how many TCO studies you do, no matter how many reports are written by an IT dept doing an evaluation, the final decision is going to be made based on how comfortable somebody is with an idea. Going with Microsoft is safe, it's easy, because everybody else does it. That's a political decision. It's the old "nobody got fired for buying IBM" thing.
The problem with Ballmer is that he sees what he wants to see. Somehow he has to reconcile his beliefs (that Microsoft is better) with reality (people are chomping at the bit to leave them). He does this by saying:
There's another thing. Does anybody else have questions about the competency of the Munich guys to be doing such a migration? Why are they doing a crash switch, which is bound to end in tears? Why are there persistant rumours of them using VMware rather than bringing Wine up to speed on their products (which I'd guess works out cheaper in the long run and certainly provides a better desktop experience).
Finally, is it just me or does Ballmer look really evil in that photo?
Erm, if you read the whole mail you'd realise that was kind of a joke. And the developer agreed that there was a problem, but wasn't sure what the solution was.
Well, no. That's not what users do. At least not the kind we want.
Smart users write thoughtful, constructive criticism, file bugs, suggest better (in their opinion) UI designs and so on. Stupid users write rants on their webpage about how bad free lunch X/Y/Z is and how the authors OWE it to them to answer their every whim.
What a lot of people seem to forget here is that Eric S Raymond did not pay for CUPS, nor Fedora. That changes the rules a bit - it means you can't just insult the people who made it and tell them they suck. Instead, he should act politely and with civility. Until he does that, I don't see anybody rushing to address his concerns.
Now, you might say "but then Linux will never overtake Windows". But you'd be wrong, because the free software community has made great strides in usability and UI design lately - and it's not because of people like Raymond or Zawinski spouting off, it's because people who care sat down and got things done.
By broadly categorizing "users" as assholes you've not only insulted a great many people, but you've also painted a woefully inaccurate picture.
I don't really use taskbar buttons at all these days. Virtual desktops are more efficient and easier to use by a long, long way.
That takes Apple fanboyism to a whole new level.
You seem to have overlooked the fact that to be significantly better than Windows requires being able to run Windows software too (without a copy, otherwise that kind of defeats the point).
Doing that requires cloning the Windows API, which is a task of such magnitude that no corporate entity can do it. Eventually, as many programs expect these things to be there, it'll require cloning Internet Explorer and maybe WMP as well.
You brush such minor details aside as if it's merely a case of "come on chaps! let's make a better mousetrap and it'll all sort itself out!", when that clearly isn't the case.
Yeah, I don't really see what Slashdot finds so hard to understand about this. Unintegrating IE now is, quite simply, impossible. Microsoft didn't lie when they said that was the case
Do you guys have any idea at all of how many apps expect Internet Explorer and related DLLs to be installed? Working on Wine brings this point home in a really fundamental way.
Oh sure. You could remove iexplore.exe. That would remove like 0.1% of IE from your system. It'd be a pretty hollow gesture.
I'm not just talking about things like the MSHTML component. I'm talking about things like the SHLWAPI DLL - a utility library which wraps the Win32 API to some extent developed by the IE team partly to make portability between Win31, Win2K and Win98 simpler. It's only half documented, a lot of the functions are exported only by ordinal, yet a surprising number of programs expect it to be there.
What about the URL monikers implementation? What about all the installers that assume the presence of Favourites? What about all the programs that embed the Trident engine to render parts of their UI, their online help - in the case of one game that shall remain nameless even the games main menu!
Windows shipped without Internet Explorer would effectively break so many apps nobody would buy it, even if they could. Quite a lot of apps don't even complain, they just crash. Win95 not supported.
Now, this stuff is mostly academic. Shipping Windows without IE on the desktop would have made a difference - 5 years ago. Nowadays many (most?) people have never heard of Netscape, think that the Internet is the blue E icon, and so on.
The only way IE will ever disappear in other words is when Linux starts kicking Windows' ass on the desktop, which last I checked was still a year or two away just on the corporate desktop let alone the home user desktop.
You're not distinguishing between Linux the kernel and Linux the OS. The parent was almost certainly talking about the latter.
Userland backwards compatability on Linux is OK but we've certainly had our fair share of cockups. The rollout of the new threading systems (NPTL and the new TLS system) was pretty much a backwards compatability disaster. I currently have to run XMMS of all things with LD_ASSUME_KERNEL because of NPTL. Oh sure, I tried to debug it. Doesn't work when gdb pukes and dies - again due to threading. Don't even get me started on the breakage Wine has had to deal with.
The glibc/kernel guys claim they know how to write backwards compatible software but in reality they don't. They don't, because unlike Microsoft they treat backwards compatability as a science, as a fixed set of rules that if they follow they think can be held blameless. Of course when you get situations like NPTL where the old system was so broken everybody had workarounds which stop working in obscure ways when the underlying bugs were fixed, this logic breaks. They still break backwards compatability, they just end up playing the blame game instead, which is stupid.
We could have much better backwards compat without the huge hacks Microsoft use with even a few small changes to process, but I'm not seeing people interested in making those changes.
Given the intensity with which it is probably being scrutinised, I'd imagine any "obvious" embarassments would have already surfaced by now.
Hell, if you want a laugh (and if you're a win32 dev you will learn something too) go read Raymond Chens blog. Not only is this guy a near-genius level coder, but he's been working on Win32 (and USER in particular) for a very long time now. Yes, I know some of you think that's an oxymoron. He often posts interesting stories about its development and about the bizarre hacks they put in to work around broken apps (some parts of windows even go so far as to detect and correct stack corruption).
Can you please join #autopackage on freenode during GMT hours? I'd like to get to the bottom of this. It sounds like you have a broke libvorbisfile, so what distro is this?
Also what is the output of this command:
objdump -x /usr/lib/libvorbisfile.so.2|grep NEEDED
thanks -mike
Apparently they are going to import the XFree DDX wholesale - hard but doable. This is just going from the words of one guy though. I'd want the Word Of Keith before believing it myself.
I think that's pretty naive. Alan Cox himself has said he can't think of a good argument to make nVidia release their drivers - their fears as to 3rd party IP rights and performance relative to competitors are, apparently, justified.
In much the same way that it's unrealistic for Linux to not be able to run proprietary software, it's unrealistic to not be able to (adequately) use proprietary drivers. While I'd love for all drivers to be GPLd experience has shown that making the installation annoying for the end user does not lead to more GPLd drivers.
At some point, somebody will simply take the matter out of Linus' hands by writing a module that provides a stable API/ABI to other drivers. Sure, it'll be less flexible than the one available to GPLd drivers, but they'll have to live with that.
That's wrong, Wine is a compatability layer for Windows programs. Programs that only run on 2K or XP are harder to emulate because by definition they use more functionality than your average app which restricts itself to only features available in Win98, but there's no fundamental reason why they can't be run too - and in fact they are run all the time.
* Software Packaging. This is a huge pain in the ass for most commercial vendors of any Linux software
Actually it's not so bad - commercial software tends to be mostly statically linked or to come with the right DSOs included. Programs like Loki Setup or BitRock can make pretty NSIS/InstallShield type installers.
The problems we have with software packaging are mostly problems for open source projects.
* Linux Binary Portability. This is the bane of all vendors releasing any closed-source software for Linux.
I'd venture that this isn't as problematic as you think it is - having done a ton of research into this, Linux binary portability is not really any worse than the difference between Win98 and WinXP, for instance.
Yes, the rollout of NPTL is being done badly, and this is unfortunate, but the vast majority of apps don't suffer any problems at all.
We're really nearly there .... the default/stock Fedora Core install has built in CD writing - just insert a blank CD, drop files into the folder that pops up and hit "Burn" (built into gnome). Couldn't be easier. Right click ISOs to burn them. Just need to nail burning audio CDs in an integrated fasion, maybe with RhythmBox.
Totem is a great DVD player. The only problem is that it doesn't work out of the box thanks to the DMCA problems with libdvdcss. I think a better "start here" trail for new users would help this, as it's not hard to install the needed RPM (just download and click it, really).
2. Friends who are familiar with the OS/Distro
I think this isn't so bad. You know, nearly everybody I talk to at my university has heard of Linux. I mean, 99%. Most of them know a friend, or friend-of-a-friend, who uses it.
7. Some packaging system with less dependency problems. [Yes, there are a few that show very good promise, with only occasional issues surfacing.]
Yup! :) Autopackage 0.4 will be out in the next week, promise!
Not really. What you mean is software developed for Linux can run on both Windows and MacOS X without too much difficulty, because they are typically based purely on free software. I'd note that while some Windows apps can be run on Linux, no MacOS X apps can be run on Linux, nor on Windows.
So it's better to say that Linux software is more compatible, than to bring MacOS into the picture (it doesn't even have dlopen as part of the core, just like Windows doesn't).
Ever considered that some people don't see it that way? A mac is just like a PC running Windows in terms of economics and philosophy, swapping Bill Gates for Steve Jobs isn't a useful trade to make. So in reality it's free software, vs non-free software (or platforms, to be more accurate).
People who think it's Windows vs everything else are just shortsighted IMHO