Listen, user feedback is a valuable way to make your software better. If you don't want to do what it takes to make your software good, then don't try to promote your software, and don't get upset when no one wants to use it. Don't sit back, shooting down user feedback because the users aren't elite enough, and then complain that stupid users aren't using your software.,
The question becomes "What OSS developers are complaining that nobody is using their software?" It certainly isn't teh GIMP developers. Plenty of people use the GIMP. Unless you can show that the OSS developers (GIMP developers in particular) that are complaining AND refuse to accept suggestions, this whole topic is moot.
Again, this ISN'T a complaint about OSS in general, but only a minority in the community. There is a lot of OSS that is quite good. However, those pieces of software usually come out of a community that is open to user complaints and receptive to input. They clearly were not employing the "If you don't like it, go f$#% yourself," method of customer service.
You are making a very general argument using a specific subset of people that is only vaguely defined here. It might help if you identify the unresponsive part of the OSS community. If you mean the GIMP developers, well that doesn't fit your idea that it is usually only bad projects that are unreponsive.
Are the GIMP developers happy with their small user-base, or would they like to see their program used more widely?
Insofar as the *nix world is a "small user-base," I am pretty sure they are happy with it. Last I checked, the GIMP developers have explicity stated that their target audience is *nix users. A Windows or OS X port of The GIMP is little more than a side-effect of the GTK+ libraries being ported/portable.
Do they want it to be considered a PITA to use by graphic designers, or would they prefer that a Photoshop user sits down in front of the GIMP and says, "Wow, this is nice."
Considering that The GIMP is targetted to *nix users, I doubt developers care much about Photoshop/Windows/Mac users.
That said, when I was first learning to use the GIMP, I was able to follow Photoshop tutorials almost word for word. I don't think the differences are that great. But of course, I am not a "graphics designer."
The GIMP is just an example of something larger that we're talking about, though. The real question is, is OSS just supposed to be something that a small number of geeks tinker with, or is it supposed to be something my mom can use too?
OSS is "supposed to be" something that any number of geeks can tinker with unless otherwise stated. If you can find a project whose explicit goal is broad acceptance, but refuses to accept input from users, then you might have a point. Otherwise, you can take it or leave it.
We'll never have very rigid standards in anything OSS because, I believe, programmers let their egos get in the way of creating the most usable program possible. They resent the notion of someone telling them how their project should function, and offen interpret any feedback as an attempt to stifle their creativity.
Wah, wah, wah! Quit whining. These people are working for free to give you software. You should be grateful.
A lot of people like doing things their way, and that's fine! But when we see such fragmentation, forks, redundancy, etc. in OSS projects, we can't be surprised when interoperability is next to impossible.
There is plenty of interoperability. Maybe it isn't all polished and marketable as you would like, but there is interoperability galore in the form of starndard file formats, protocols, etc. Where interoperability is difficult, I doubt any OSS developers are particularly surprised.
So if you need to make your project work in a way that only you want it to work, don't be surprised when nobody else uses it.
Doesn't seem to be a problem with The GIMP or OSS in general. We wouldn't even be talking about The GIMP right now if nobody was using it. The fact is that The GIMP is extremely powerful and works well for those who want to use it. The only people I see complaining in this thread are users, not the developers. I'm sure the GIMP developers are quite pleased with its popularity. Same with Linus Torvalds and Linux. Same with the Apache group, etc, etc.
(BTW, this isn't not a complaint about OSS as a whole, but only a rebuttal of the "You don't like it, fix it" attitude expressed by some.)
How about "you don't like it, don't use it?" I guess the argument against this attitude is that there is so much hype about OSS saying that we SHOULD be using it as a general rule. All I can say is that I, personally, never suggest anyone use OSS if it doesn't fit their needs. GIMP doesn't fit everyone's needs, but it is still a pretty damn nice piece of OSS. Why should OSS developers try to make software that is fit for general consumption? Isn't the main complaint about WIndows and many products that run on it that they pander to the lowest common denomonator while simutaneously trying to pack in every concevable bell and whistle? Isn't it good enough to say that The GIMP is very powerful and works really well fo the people who use it?
State of the art search technologies. Longhorn is the mother of the technology also used for Gnome's Beagle and Dashboard, and it will ship with an excellent interface for their search tools (probably much better than Apple's Spotlight).
You missed my main question. Is this enough to warrant an upgrade? I know Microsoft will push it all as amazing must have stuff, but really, are people demanding this kind of stuff like they are demanding security and freedom from spyware? As far as I can tell, Longhorn looks to be much more of the same in that department. Just another initial release full of unexpected security problems. At least Apple has the sense to grandualy introduce new features without changing the basic structure a whole lot (aside from the jump from OS 9 to OS X).
People I know that upgraded from 98 to XP didn't do it because they wanted XP's "features." They upgraded because that was the only way they could get any semblence of stability. That or XP just came with the new computer they got. If you remember, even XP didn't catch on as quickly as Microsoft had hoped it would.
Microsoft's only hope is that they can dupe the user into thinking that Longhorn will address their main concerns merely by virtue of being radically different than previous releases.
But then again, I have been proven to be out of touch with "regular" users on more than one occassion.
Make sure you aren't confusing "consutant" with "temp." I find it amusing how so many IT workers think they discovered some new work paradigm. When in fact, clerical workers have been doing this kind of thing for quite some time now. There is consulting, where one tackles large projects which a company could not be expected to keep the appropriate skills onhand for, and then there is temporary/part time work where a company simply doesn't want to pay employee benefits.
Personally, I have done some mon-wed-fri work for a large financial institution and I'll tell you what, it SUCKS ASS. I hate going there to be their bitch for a day. One minute I'll be configuring their corporate firewall and the next I will be physically installing some asshole stockbroker's Windows PC. That is not consulting. At least when you are working for a smaller company they understand the value of your time. You come in, tackle the toughest jobs, and you are done. As long as you are strict about billing for your time, few small companies are going to abuse your services like many larger companies will.
Linux was designed from the ground up to be secure. Most of the problems in Linux are bugs, not features that were deliberately designed insecure.
I was right with you up until this point. Linux was most certainly not designed from the gorund up to be secure. It was built largely as a learning tool for Linus Torvalds (and others). The only reason Linux lacks hooks for viruses and spyware is because few developers have cared enough to implement the kinds of esoteric hidden features that are used by Windows to make it "friendly" to both programmers and users alike.
So, if Firefox is to combat IE on Longhorn, they will have to push and take as much of the marketplace as they can before Longhorn hits the market.
One has to wonder if users will even pick up Longhorn all the quickly. The upgrade from ME to XP was a nobrainer. 95/98/ME barely worked at all for many people. XP is stable enough that I think a good chunk of people will either stick with it or maybe try a Mac. What is Longhorn's major selling point? What does it stand to offer the average user? MORE features? MS is already hiding most of the control panels from the average user. Seems like more features will just be more control panels to hide.
On the other hand, it is so much easier to get into the buisiness working with the little fish. It can be good money. No mid to large sized company with half a "brain" is going to hire an independant consultant. If they need that kind of help, they should hire someone that can get to know their systems. If they have a really big problem, they need real consultants with the resources to do the job right and bring in a team if necessary. Not that all or even most larger consulting companies are necessarily competent, of course. But it is a better bet than the CEO's son's friend.
An 8088 11-12 years ago??? Dude, that was 1993. I was running Linux on a 80486 in 1993. You were a whole 3 CPU generations behind the state of the art. No wonder you had such a dramatic experience.
This is what what Andrew Tanenbaum was talking about in 1992 and if Linus Torvalds hadn't overreacted so furiously when Linux was only one year old, today, 13 years later, we might not face many of the problems you are talking about.
I don't really see how this has a whole lot to do with the current topic. Adny was talking about microkernels and portability; claiming that Linux was "obsolete." Considering that Andy turned out to be completely wrong on just about every count, I am not surprised that Linus "overreacted." Linux is not obsolete, the monolithic design has not significantly affected stability or portability, and where is Minix or GNU/HURD now? It seems as though Linux's fast development has been what kept Linux viable. Linus was right.
As far as potentially unstable binary drivers... well, they don't generally ship with the default kernel in most distributions. . At least not without specifically installing a package. And how would one support the features of some of the newer hardware without them? I have a nice 3D NVIDIA card. Yeah, the kernel module is huge and it can cause instability, but how else would I run my favorite OpenGL apps as fast as I do? The answer isn't "install NetBSD." And it certainly isn't "install GNU/HURD or Minix." My only other option is to install a worse operating system, Windows. Which, BTW, is microkernel based. So what the hell was Andy talking about?
At best you can say that any given new kernel release is potentially unstable for reasons of quick development, but there is nothing stopping an individual or a distribution from standardizing on one known stable kernel and only taking security updates from newer development. I know this is what Debian does. Debian stable lags well behind the bleeding edge in both kernel version an package versions. All security updates get backported, but the basic system remains solid. If you feel adventurous or absolutely need the bleeding edge kernel with all the cool new features, you can compile and install it. So what is the big deal? The choice in Linux distributions compensates for the problems of quick development. If one decides to use linux, one is not obligated to keep up with the latest and coolest deveopment. You can pick a conservative distribution or a liberal one. Whatever suits you. You even have this to some degree in the BSD world. One could say that, in general, *BSD is more conservative. And that is cool. In the end, it is all about choice.
Great, so now they are reinventing the Debian project itself? Why not just run Debian stable? It already has all the proposed features of Ubuntu Enterprise.
18 months of support is not good enough. I have servers that sit around for years which need to be up to date with all current security patches. Although apt-get dist-upgrade is not difficult to do, it is not practical to do every 18 months or so.
What about security releases? Is Ubuntu as good about security as Debian? You know, there is a reason Debian takes so long to release stable. It is very well tested. Ubuntu is a little too bleeding edge for most serious server environments. In my experience, the latest greatest packages are not generally necessary on servers. I'd stick with Debian, personally.
The question is: Why are you, a supposed IT expert, asking a PHB to make a technical decision? A good manager wants the *right* answer to all IT problems... not just one answer.
Are you saying that Red Hat aren't responsible for the Linux Kernel that they package and sell? If not, why not? It's part of their distro.
The real question here is, "Does any software company take responsiblity for any of the software that they sell?" According to most of the EULAs that I have skimmed, the answer is a resounding "no." If a Microsoft based network is hacked or installing the software somehow damages the hardware, how does Microsoft take responsibility for that? They don't and maybe they shouldn't, but the point is that no software company really takes responsibility beyond releasing updates and patches which both Redhat and Microsoft do. Redhat does so despite not being the owner of the code. That is what you pay them for. Even when you are not paying someone, such with the Debian project, you STILL get timely and relavent updates to OSS softwrare. The Microsoft argument is bullshit.
When most people talk about using commercial software for accountability, they are not talking about having someone who will take responsibility, but rather just someone to point a finger at to cover their own incompitent asses! The Suits know that Microsoft won't step in to pay reparations or fix their software when it breaks. It is all about politics and finger pointing, not real responsibility.
Well, amd64 optimizations are not optimizations in the same sense as, say, P4 optimizations. It is more like a different architecture with backwards compatability. Other Linux distributions actually have versions specific for amd64. So running Gentoo to get optimizations should be unnecessary.
Well, I've been running Linux almost exclusively for 8 years. I guess I'm probably beyond whatever Gentoo could show me. To me it was just a big pain the ass. Between the merging of etc-updates and compiling every damn piece of software I want to run. Also, Gentoo seemed to use an unusually large amount of disk space just for a regular install... even after making sure all the dist files were cleaned. And then there was building OpenOffice. I'll never do that again.
I agree. Ubuntu doesn't have many applications outside of the standard gnome applications. I found myself pulling many packages from the Debian respositories. Soon enough I had a full blown hybrid on my hands. So I scrapped it and went back to full Debian sarge.
Gentoo with a friendly UI? That is... wrong, obscene somehow. Like LFS with an XFCE-4.2 style graphical installer. A major selling point of Gentoo is that it breaks you, kicks you, and forces you to understand how a Linux distro is built - like a marine bootcamp for the OS-impaired.
Not to mention Gentoo being a lesson in patience! I got tired of waiting for stuff to complile every time I wanted some misc. utility or application or game (not to mention openoffice!) I'm pretty sure the 10% more total CPU I gained from -march=pentium4 optimizations was eaten up by gcc long before I even got up and running.
Also, I don't really buy the idea that Gentoo makes people understand how an OS works. Just because it is a PITA, doesn't mean someone is learning. Most new Gentoo users just follow the receipes for getting stuff done. What is to learn?
Slashdot Mirror
The question becomes "What OSS developers are complaining that nobody is using their software?" It certainly isn't teh GIMP developers. Plenty of people use the GIMP. Unless you can show that the OSS developers (GIMP developers in particular) that are complaining AND refuse to accept suggestions, this whole topic is moot.
Again, this ISN'T a complaint about OSS in general, but only a minority in the community. There is a lot of OSS that is quite good. However, those pieces of software usually come out of a community that is open to user complaints and receptive to input. They clearly were not employing the "If you don't like it, go f$#% yourself," method of customer service.
You are making a very general argument using a specific subset of people that is only vaguely defined here. It might help if you identify the unresponsive part of the OSS community. If you mean the GIMP developers, well that doesn't fit your idea that it is usually only bad projects that are unreponsive.
Are the GIMP developers happy with their small user-base, or would they like to see their program used more widely?
Insofar as the *nix world is a "small user-base," I am pretty sure they are happy with it. Last I checked, the GIMP developers have explicity stated that their target audience is *nix users. A Windows or OS X port of The GIMP is little more than a side-effect of the GTK+ libraries being ported/portable.
Do they want it to be considered a PITA to use by graphic designers, or would they prefer that a Photoshop user sits down in front of the GIMP and says, "Wow, this is nice."
Considering that The GIMP is targetted to *nix users, I doubt developers care much about Photoshop/Windows/Mac users.
That said, when I was first learning to use the GIMP, I was able to follow Photoshop tutorials almost word for word. I don't think the differences are that great. But of course, I am not a "graphics designer."
The GIMP is just an example of something larger that we're talking about, though. The real question is, is OSS just supposed to be something that a small number of geeks tinker with, or is it supposed to be something my mom can use too?
OSS is "supposed to be" something that any number of geeks can tinker with unless otherwise stated. If you can find a project whose explicit goal is broad acceptance, but refuses to accept input from users, then you might have a point. Otherwise, you can take it or leave it.
-matthew
Wah, wah, wah! Quit whining. These people are working for free to give you software. You should be grateful.
A lot of people like doing things their way, and that's fine! But when we see such fragmentation, forks, redundancy, etc. in OSS projects, we can't be surprised when interoperability is next to impossible.
There is plenty of interoperability. Maybe it isn't all polished and marketable as you would like, but there is interoperability galore in the form of starndard file formats, protocols, etc. Where interoperability is difficult, I doubt any OSS developers are particularly surprised.
So if you need to make your project work in a way that only you want it to work, don't be surprised when nobody else uses it.
Doesn't seem to be a problem with The GIMP or OSS in general. We wouldn't even be talking about The GIMP right now if nobody was using it. The fact is that The GIMP is extremely powerful and works well for those who want to use it. The only people I see complaining in this thread are users, not the developers. I'm sure the GIMP developers are quite pleased with its popularity. Same with Linus Torvalds and Linux. Same with the Apache group, etc, etc.
-matthew
How about "you don't like it, don't use it?" I guess the argument against this attitude is that there is so much hype about OSS saying that we SHOULD be using it as a general rule. All I can say is that I, personally, never suggest anyone use OSS if it doesn't fit their needs. GIMP doesn't fit everyone's needs, but it is still a pretty damn nice piece of OSS. Why should OSS developers try to make software that is fit for general consumption? Isn't the main complaint about WIndows and many products that run on it that they pander to the lowest common denomonator while simutaneously trying to pack in every concevable bell and whistle? Isn't it good enough to say that The GIMP is very powerful and works really well fo the people who use it?
-matthew
You missed my main question. Is this enough to warrant an upgrade? I know Microsoft will push it all as amazing must have stuff, but really, are people demanding this kind of stuff like they are demanding security and freedom from spyware? As far as I can tell, Longhorn looks to be much more of the same in that department. Just another initial release full of unexpected security problems. At least Apple has the sense to grandualy introduce new features without changing the basic structure a whole lot (aside from the jump from OS 9 to OS X).
People I know that upgraded from 98 to XP didn't do it because they wanted XP's "features." They upgraded because that was the only way they could get any semblence of stability. That or XP just came with the new computer they got. If you remember, even XP didn't catch on as quickly as Microsoft had hoped it would.
Microsoft's only hope is that they can dupe the user into thinking that Longhorn will address their main concerns merely by virtue of being radically different than previous releases.
But then again, I have been proven to be out of touch with "regular" users on more than one occassion.
-matthew
Make sure you aren't confusing "consutant" with "temp." I find it amusing how so many IT workers think they discovered some new work paradigm. When in fact, clerical workers have been doing this kind of thing for quite some time now. There is consulting, where one tackles large projects which a company could not be expected to keep the appropriate skills onhand for, and then there is temporary/part time work where a company simply doesn't want to pay employee benefits.
Personally, I have done some mon-wed-fri work for a large financial institution and I'll tell you what, it SUCKS ASS. I hate going there to be their bitch for a day. One minute I'll be configuring their corporate firewall and the next I will be physically installing some asshole stockbroker's Windows PC. That is not consulting. At least when you are working for a smaller company they understand the value of your time. You come in, tackle the toughest jobs, and you are done. As long as you are strict about billing for your time, few small companies are going to abuse your services like many larger companies will.
-matthew
Working nearly full time in a temporary position with no benefits isn't the same thing as being a consultant. It is called "temping." ;-)
-matthew
I was right with you up until this point. Linux was most certainly not designed from the gorund up to be secure. It was built largely as a learning tool for Linus Torvalds (and others). The only reason Linux lacks hooks for viruses and spyware is because few developers have cared enough to implement the kinds of esoteric hidden features that are used by Windows to make it "friendly" to both programmers and users alike.
-matthew
One has to wonder if users will even pick up Longhorn all the quickly. The upgrade from ME to XP was a nobrainer. 95/98/ME barely worked at all for many people. XP is stable enough that I think a good chunk of people will either stick with it or maybe try a Mac. What is Longhorn's major selling point? What does it stand to offer the average user? MORE features? MS is already hiding most of the control panels from the average user. Seems like more features will just be more control panels to hide.
-matthew
On the other hand, it is so much easier to get into the buisiness working with the little fish. It can be good money. No mid to large sized company with half a "brain" is going to hire an independant consultant. If they need that kind of help, they should hire someone that can get to know their systems. If they have a really big problem, they need real consultants with the resources to do the job right and bring in a team if necessary. Not that all or even most larger consulting companies are necessarily competent, of course. But it is a better bet than the CEO's son's friend.
-matthew
An 8088 11-12 years ago??? Dude, that was 1993. I was running Linux on a 80486 in 1993. You were a whole 3 CPU generations behind the state of the art. No wonder you had such a dramatic experience.
-matthew
Nobody said that the trend was recent.
-matthew
Some asses are actually pretty nice to look at. Although I am thinking more along the lines of "J-Lo" than GNOME.
I don't really see how this has a whole lot to do with the current topic. Adny was talking about microkernels and portability; claiming that Linux was "obsolete." Considering that Andy turned out to be completely wrong on just about every count, I am not surprised that Linus "overreacted." Linux is not obsolete, the monolithic design has not significantly affected stability or portability, and where is Minix or GNU/HURD now? It seems as though Linux's fast development has been what kept Linux viable. Linus was right.
As far as potentially unstable binary drivers... well, they don't generally ship with the default kernel in most distributions. . At least not without specifically installing a package. And how would one support the features of some of the newer hardware without them? I have a nice 3D NVIDIA card. Yeah, the kernel module is huge and it can cause instability, but how else would I run my favorite OpenGL apps as fast as I do? The answer isn't "install NetBSD." And it certainly isn't "install GNU/HURD or Minix." My only other option is to install a worse operating system, Windows. Which, BTW, is microkernel based. So what the hell was Andy talking about?
-matthew
At best you can say that any given new kernel release is potentially unstable for reasons of quick development, but there is nothing stopping an individual or a distribution from standardizing on one known stable kernel and only taking security updates from newer development. I know this is what Debian does. Debian stable lags well behind the bleeding edge in both kernel version an package versions. All security updates get backported, but the basic system remains solid. If you feel adventurous or absolutely need the bleeding edge kernel with all the cool new features, you can compile and install it. So what is the big deal? The choice in Linux distributions compensates for the problems of quick development. If one decides to use linux, one is not obligated to keep up with the latest and coolest deveopment. You can pick a conservative distribution or a liberal one. Whatever suits you. You even have this to some degree in the BSD world. One could say that, in general, *BSD is more conservative. And that is cool. In the end, it is all about choice.
-matthew
Great, so now they are reinventing the Debian project itself? Why not just run Debian stable? It already has all the proposed features of Ubuntu Enterprise.
-matthew
I haven't gotten to the point where I am actually installing Linux on other people's desktops, but I'll consider Ubuntu when I do. ;-)
-matthew
18 months of support is not good enough. I have servers that sit around for years which need to be up to date with all current security patches. Although apt-get dist-upgrade is not difficult to do, it is not practical to do every 18 months or so.
What about security releases? Is Ubuntu as good about security as Debian? You know, there is a reason Debian takes so long to release stable. It is very well tested. Ubuntu is a little too bleeding edge for most serious server environments. In my experience, the latest greatest packages are not generally necessary on servers. I'd stick with Debian, personally.
-matthew
The question is: Why are you, a supposed IT expert, asking a PHB to make a technical decision? A good manager wants the *right* answer to all IT problems... not just one answer.
-matthew
Umm, isn't Ubuntu primarily a desktop distribution? Debian proper would be a much better choice for servers.
-matthew
The real question here is, "Does any software company take responsiblity for any of the software that they sell?" According to most of the EULAs that I have skimmed, the answer is a resounding "no." If a Microsoft based network is hacked or installing the software somehow damages the hardware, how does Microsoft take responsibility for that? They don't and maybe they shouldn't, but the point is that no software company really takes responsibility beyond releasing updates and patches which both Redhat and Microsoft do. Redhat does so despite not being the owner of the code. That is what you pay them for. Even when you are not paying someone, such with the Debian project, you STILL get timely and relavent updates to OSS softwrare. The Microsoft argument is bullshit.
When most people talk about using commercial software for accountability, they are not talking about having someone who will take responsibility, but rather just someone to point a finger at to cover their own incompitent asses! The Suits know that Microsoft won't step in to pay reparations or fix their software when it breaks. It is all about politics and finger pointing, not real responsibility.
-matthew
Well, amd64 optimizations are not optimizations in the same sense as, say, P4 optimizations. It is more like a different architecture with backwards compatability. Other Linux distributions actually have versions specific for amd64. So running Gentoo to get optimizations should be unnecessary.
-matthew
Well, I've been running Linux almost exclusively for 8 years. I guess I'm probably beyond whatever Gentoo could show me. To me it was just a big pain the ass. Between the merging of etc-updates and compiling every damn piece of software I want to run. Also, Gentoo seemed to use an unusually large amount of disk space just for a regular install... even after making sure all the dist files were cleaned. And then there was building OpenOffice. I'll never do that again.
-matthew
So why run Ubuntu instead of Debian then? Is it just a more "stable" version of Debian sid with a better installer?
I agree. Ubuntu doesn't have many applications outside of the standard gnome applications. I found myself pulling many packages from the Debian respositories. Soon enough I had a full blown hybrid on my hands. So I scrapped it and went back to full Debian sarge.
-matthew
Not to mention Gentoo being a lesson in patience! I got tired of waiting for stuff to complile every time I wanted some misc. utility or application or game (not to mention openoffice!) I'm pretty sure the 10% more total CPU I gained from -march=pentium4 optimizations was eaten up by gcc long before I even got up and running.
Also, I don't really buy the idea that Gentoo makes people understand how an OS works. Just because it is a PITA, doesn't mean someone is learning. Most new Gentoo users just follow the receipes for getting stuff done. What is to learn?
-matthew