"A single person wanting to romp off with somebody's IP is going to be quickly clobbered by inability to execute his plan, or deliver on his promises of product and distribution."
Perhaps, but what about a big company trying to romp off with an individual's, or and open source group's, IP? Patent law protects both the big guy and the little guy.
Sedgewick's books (Algorithms in C, Pascal, C++ -- the same book transated to different languages) are very good.
Slightly off-topic, but anyone interested in bread-and-butter algorithms should definitely check out the STL (Standard Template Library, part of the C++ standard library) for a brilliant, original and incredibly useful algorithms library.
Some people say that communism as a theory and communism as it was practiced by Lennon, Stalin, Mao, etc. are two different things -- that pure communism is a wonderful, but possibly impractical, idea that has never been put into practice. But, eliminating private property and having the state own the means of production don't seem like such nice ideas to me, even if they were practical. If Webster's definition is correct, there are fundamental problems with communism as an ideology, not just with the way it has been practiced.
Some of RMS' statements do hint at mild communism. For example, he said in a magazine interview that he didn't care whether proprietary software went away, or was banned by law. Well, I certainly do care! Banning proprietary software (a form of private property, I believe) would be a form of communism. Also, I think I read somewhere that RMS once proposed a hardware tax, with the profits going to support free software development. Presumably the government or some quasi-gonvernmental body would control the means of such software production.
I do not, however, uphold the principles of absolute and unrestricted liberty. Promoting the virtue of liberty above all others, including life and community, just doesn't make sense. Justice requires more of a balancing act.
Pure capitalism does not provide an adequate program for society, either. Fortunately, it has probablly never been practiced in its pure form. Capitalism needs to be augmented with citizens who have a sense of civic duty, and private institutions to support charitable activity. The government can encourage this activity up to a point, but if the government conducts too much of this activity itself, it takes away people's freedom for meaningful communal activity, and often botches the job.
Giving away software is one such activity -- meaningful when you do it on your own initiative, restrictive when everyone is forced to do it.
I've worked as a contractor for several years; now I'm an employee.
While I always tried to write maintainable code as a contractor, not everyone does. Right now, I'm trying to clean up some messes some contractors made. The contractors (including the owner of the contracting firm) were brought in as experts on a particular product. Two of the contractors were fine, but one (the owner of the firm!) produced absolute crap. It got the job done at the time (sort of -- bugs are cropping up), but his code is totally unmaintainable. A module of his I'm looking at now has nine goto's and three labels! Of course, he got out before he had to maintain it.
True enough -- an interpreted language and gui is usually more portable than a compiled language. But Java is not the only interpreted (yeah, I know byte-compiled -- also not a new concept) language out there. Look at Perl/Tk, Python/Tk, Tcl/Tk, etc. for cross platform gui solutions.
The author says that SGI is stable, rock-solid, but offers no evidence. He then dismisses people who say SGI as unstable as spreading FUD. Well, making blanket statements with no evidence is FUD (or reverse FUD).
Nothing against SGI, but this is a really dumb article. Layoffs and restructuring do warrant concern about the future of a company. Saying that people who are concerned about SGI's future are spreading FUD is stupid.
The original poster may have exagerated, but I have to agree with him that teacher training is a joke, and, statistically, they score a lot lower on standardized tests. This does not mean, of course, that every teacher is stupid. A very bright friend of mine is working to get his teaching certificate so that he can teach in the inner city schools. He constantly complains about how inane the teacher training curriculum is, and how lacking in basic reading, writing and arithmetic skills some of his classmates and working teachers he has observed are. True, teachers don't have to be Einsteins, but if a significant percentage of them don't even understand what they are supposed to be teaching, then we have a problem.
1. SuSE does have a graphical configuration tool, Yast. I don't care for it much myself, but they've spent a lot of time and money on it. (As an SuSE user, I'm sure you know this).
2. It seems like the people who purchase Microsoft subscriptions are mostly corporate developers. We have a full cabinet full where I work. True, we don't actually upgrade our NT servers and whatnot that often, but management and the developers like to know that if they have a problem that requires a patch, they have the CD on hand. In fact, most (all?) of our software purchases here at Enterprise Rent-A-Car are subscriptions. So I think that, based on empirical evidence, subscriptions appeal to corporate users as much or more than hobbyists.
Fundamentally, though, I agree that RedHat should think of itself as a services company, not a software company. Which makes the whole RedHat Labs/Gnome/Enlightment fiasco puzzling. If RedHat wants to focus on servers, why spend money developing a desktop, especially an instable window manager (Enlightenment) that excells at providing eye candy and hogging resources? Don't get me wrong -- I like Enlightment, and Gnome is OK, but would you want to run it on a server? (Of course, Raster left RedHat, so presumably E will not be the default RedHat window manager in the future).
-------------------------------------------
RedHat should NOT tie itself to specific products
on
Open Letter to Red Hat
·
· Score: 2
I don't think it serves RedHat's interests to buy or take over particular projects or technology companies. Indirectly support, yes, but buy or take over, no. Here's why:
As a distribution, RedHat's best bet is to skim off the cream from all the Linux projects, maybe support some of the key ones indirectly by giving money, and integrate everything. If RedHat tries to take over a project, it hurts the project's credibility, since it could be seen as serving only RedHat's interests, and limits RedHat's perceived options, since RedHat would be seen as having a vested interest in their own projects.
For example, RedHat's connection with Gnome has hurt both Gnome, Enlightenment and RedHat. Rightly or wrongly, Gnome 1.0 was perceived by many as a buggy product, released early to fit RedHat's schedule. Conversely, RedHat's choice of Gnome as their default desktop provided an opening for Mandrake to step in. Corporate and newbie buyers want the most stable, easiest to use desktop. They don't want to think that their distributor has a vested interest in pushing their own, possibly inferior product. They want to think that their distributor is objective, and chooses the best stuff for them, so they don't have to think about it.
(Gnome has gotten better since the 1.0 release, but the damage to their reputation still lingers. My point is not that Gnome is good or bad, but that it's close association with RedHat hurts both RedHat's and Gnome's reputation.)
For similar reasons, I don't think RedHat should buy Cygnus. It's bad enough that people think of gcc as the "Cygnus" compiler; it would be worse if it became the "RedHat" compiler.
Instead of looking at itself as a software development company, RedHat should look at itself as a service company. Don't buy Borland, SCO, Cygnus, or whatever. Instead, spend money on relationships with major PC vendors, beef up your support staff, improve documentation, consider subscriptions, etc. Also, look at IBM's model, and consider developing custom solutions for companies, but using whatever tools are the best, or what the customer wants, not pushing your own stuff.
In short, the standard software will take care of itself in the long run. Focus instead on the added value you can offer as an integrator, store house of Linux expertise, support provider, etc.
I've been wondering about this myself. (Although personally I'm interested in using with KDevelop).
Maybe it could also be given a Corba OpenParts and/or Corba Baboon wrapper. I've thought of doing an OpenParts wrapper for XEmacs myself, but I've never gotten around to it (haven't looked into it to see if it's feasable).
Of course a Qt or Gtk+ implementation of Emacs would also be nice, but a lot more work.
I agree with you that RAD development is not a substitute for design. But I do think that visual programming tools are particularly suited to gui development. It's very intuitive to move controls around and see what they look like.
Also, there is definitely a place for tools to throw together throw-away apps quickly. Witness perl, python, and other scripting languages. Admittely, they can be misused, but would it be a better world if everyone was forced to use compiled languages for everything? I think the same thing applies with gui tools. People like Tk because it's quick to write for. RAD environments offer the same advantage.
Having said all of that, I personally prefer writing C++/Qt apps with emacs.
I tend to agree. I do view works of programming as works of art (artifacts), but to call the whole "Open Source" process, as opposed a specific artifact like the Linux kernel, a work of art is streching it.
I think this guy has a post-modernist agenda that he is trying to sneak in under the Open Source hype. He compares patches to "sampling", a sees a lack of authorial viewpoint in OpenSource.
He may have a point, but it's overstated. Many of the more famous open source projects were in fact the results of the vision of one or a few persons -- Linux and Linus, Emacs and RMS, etc.
Programming is definitely art in the traditional sense, not just the post-modernists (non)-sense.
The term "revolutionary" is thrown around way too much, for things that are really evolutionary. On a technical level, at least, Linux was definitely not revolutionary.
I'm not against certification. It's useful for non-programmers to figure out if someone is qualified or not. I'm not sure that an computer-engineering degree means that you're qualified, however. (But MSCE is a total joke). In California, you can take the bar without a law degree. If you pass, you're a lawyer. Also, in all states, just getting a law degree doesn't mean you can legally practice law. You have to pass the bar. Maybe the "software enginering" profession needs something similar.
Personally, I call my self a "hacker," or, for those who associate hacker with cracker, a programmer.
How do you define engineer? As someone who has an engineering degree? That's like saying a musician is someone who has a music degree. Some great musicians have music degrees, others don't. Many famous symphonic musicians only have high school degrees. The proof is in the pudding -- either you can play or you can't, and when you audition for an orchestra, they ask you to play. They don't ask to see your diploma.
Here's Webster's definition of engineer:
"A person who carries through an enterprise by skillful or artful contrivance."
I also read the letter in IEEE criticising Ellen Ulman's credentials. I thought it was unfair. Ulman has supervised and designed many successfull software enterprises through her skillful and artful contrivances. I think that fits the definition of engineer.
I prefer the term "hacker," however. To me, a hacker is someone who a person who carries through a software enterprise by skillful or artful contrivance.
What about Emacs? Are hackers moving away from it for hipper, newer editing environments, like Crisp, Visual Slick Edit, et al? Or the vi fanatics, are they moving on? What about Unix gurus in general? Did they all switch to OS/2, or Be, or whatever? Some of them, maybe, but for other people, when they get really comfortable with an old tool, they stick with it forever, and keep customizing and extending it.
I heard some people complain about the X consortium's control over X. Could we fork off a different branch of X, a 'la [X]?Emacs, and would that be a good idea?
A friend of mine says he doesn't like X's client/server over the network model. He says that this consumes unnecessary resources when everything (the X server and clients) are running locally. I said I thought any wasted resources were marginal. Anyone know for sure?
Slashdot Mirror
You said:
"A single person wanting to romp off with somebody's IP is going to be quickly clobbered by inability to execute his plan, or deliver on his promises of product and distribution."
Perhaps, but what about a big company trying to romp off with an individual's, or and open source group's, IP? Patent law protects both the big guy and the little guy.
Sedgewick's books (Algorithms in C, Pascal, C++ -- the same book transated to different languages) are very good.
Slightly off-topic, but anyone interested in bread-and-butter algorithms should definitely check out the STL (Standard Template Library, part of the C++ standard library) for a brilliant, original and incredibly useful algorithms library.
---------------------------------------
Good dictionary citation!
Some people say that communism as a theory and communism as it was practiced by Lennon, Stalin, Mao, etc. are two different things -- that pure communism is a wonderful, but possibly impractical, idea that has never been put into practice. But, eliminating private property and having the state own the means of production don't seem like such nice ideas to me, even if they were practical. If Webster's definition is correct, there are fundamental problems with communism as an ideology, not just with the way it has been practiced.
Some of RMS' statements do hint at mild communism. For example, he said in a magazine interview that he didn't care whether proprietary software went away, or was banned by law. Well, I certainly do care! Banning proprietary software (a form of private property, I believe) would be a form of communism. Also, I think I read somewhere that RMS once proposed a hardware tax, with the profits going to support free software development. Presumably the government or some quasi-gonvernmental body would control the means of such software production.
I do not, however, uphold the principles of absolute and unrestricted liberty. Promoting the virtue of liberty above all others, including life and community, just doesn't make sense. Justice requires more of a balancing act.
Pure capitalism does not provide an adequate program for society, either. Fortunately, it has probablly never been practiced in its pure form. Capitalism needs to be augmented with citizens who have a sense of civic duty, and private institutions to support charitable activity. The government can encourage this activity up to a point, but if the government conducts too much of this activity itself, it takes away people's freedom for meaningful communal activity, and often botches the job.
Giving away software is one such activity -- meaningful when you do it on your own initiative, restrictive when everyone is forced to do it.
---------------------------------
I agree with the above.
- --
I've worked as a contractor for several years; now I'm an employee.
While I always tried to write maintainable code as a contractor, not everyone does. Right now, I'm trying to clean up some messes some contractors made. The contractors (including the owner of the contracting firm) were brought in as experts on a particular product. Two of the contractors were fine, but one (the owner of the firm!) produced absolute crap. It got the job done at the time (sort of -- bugs are cropping up), but his code is totally unmaintainable. A module of his I'm looking at now has nine goto's and three labels! Of course, he got out before he had to maintain it.
-----------------------------------------------
True enough -- an interpreted language and gui is usually more portable than a compiled language. But Java is not the only interpreted (yeah, I know byte-compiled -- also not a new concept) language out there. Look at Perl/Tk, Python/Tk, Tcl/Tk, etc. for cross platform gui solutions.
What's really unique about Java is the hype.
-----------------------------------------
What's wrong with ksh for Linux?
The author says that SGI is stable, rock-solid, but offers no evidence. He then dismisses people who say SGI as unstable as spreading FUD. Well, making blanket statements with no evidence is FUD (or reverse FUD).
Nothing against SGI, but this is a really dumb article. Layoffs and restructuring do warrant concern about the future of a company. Saying that people who are concerned about SGI's future are spreading FUD is stupid.
----------------------------------
The original poster may have exagerated, but I have to agree with him that teacher training is a joke, and, statistically, they score a lot lower on standardized tests. This does not mean, of course, that every teacher is stupid. A very bright friend of mine is working to get his teaching certificate so that he can teach in the inner city schools. He constantly complains about how inane the teacher training curriculum is, and how lacking in basic reading, writing and arithmetic skills some of his classmates and working teachers he has observed are. True, teachers don't have to be Einsteins, but if a significant percentage of them don't even understand what they are supposed to be teaching, then we have a problem.
-----------------------------------
I agree with most of what you said.
A few points:
1. SuSE does have a graphical configuration tool, Yast. I don't care for it much myself, but they've spent a lot of time and money on it. (As an SuSE user, I'm sure you know this).
2. It seems like the people who purchase Microsoft subscriptions are mostly corporate developers. We have a full cabinet full where I work. True, we don't actually upgrade our NT servers and whatnot that often, but management and the developers like to know that if they have a problem that requires a patch, they have the CD on hand. In fact, most (all?) of our software purchases here at Enterprise Rent-A-Car are subscriptions. So I think that, based on empirical evidence, subscriptions appeal to corporate users as much or more than hobbyists.
Fundamentally, though, I agree that RedHat should think of itself as a services company, not a software company. Which makes the whole RedHat Labs/Gnome/Enlightment fiasco puzzling. If RedHat wants to focus on servers, why spend money developing a desktop, especially an instable window manager (Enlightenment) that excells at providing eye candy and hogging resources? Don't get me wrong -- I like Enlightment, and Gnome is OK, but would you want to run it on a server? (Of course, Raster left RedHat, so presumably E will not be the default RedHat window manager in the future).
-------------------------------------------
I don't think it serves RedHat's interests to buy or take over particular projects or technology companies. Indirectly support, yes, but buy or take over, no. Here's why:
As a distribution, RedHat's best bet is to skim off the cream from all the Linux projects, maybe support some of the key ones indirectly by giving money, and integrate everything. If RedHat tries to take over a project, it hurts the project's credibility, since it could be seen as serving only RedHat's interests, and limits RedHat's perceived options, since RedHat would be seen as having a vested interest in their own projects.
For example, RedHat's connection with Gnome has hurt both Gnome, Enlightenment and RedHat. Rightly or wrongly, Gnome 1.0 was perceived by many as a buggy product, released early to fit RedHat's schedule. Conversely, RedHat's choice of Gnome as their default desktop provided an opening for Mandrake to step in. Corporate and newbie buyers want the most stable, easiest to use desktop. They don't want to think that their distributor has a vested interest in pushing their own, possibly inferior product. They want to think that their distributor is objective, and chooses the best stuff for them, so they don't have to think about it.
(Gnome has gotten better since the 1.0 release, but the damage to their reputation still lingers. My point is not that Gnome is good or bad, but that it's close association with RedHat hurts both RedHat's and Gnome's reputation.)
For similar reasons, I don't think RedHat should buy Cygnus. It's bad enough that people think of gcc as the "Cygnus" compiler; it would be worse if it became the "RedHat" compiler.
Instead of looking at itself as a software development company, RedHat should look at itself as a service company. Don't buy Borland, SCO, Cygnus, or whatever. Instead, spend money on relationships with major PC vendors, beef up your support staff, improve documentation, consider subscriptions, etc. Also, look at IBM's model, and consider developing custom solutions for companies, but using whatever tools are the best, or what the customer wants, not pushing your own stuff.
In short, the standard software will take care of itself in the long run. Focus instead on the added value you can offer as an integrator, store house of Linux expertise, support provider, etc.
-------------------------------------
I've been wondering about this myself. (Although personally I'm interested in using with KDevelop).
Maybe it could also be given a Corba OpenParts and/or Corba Baboon wrapper. I've thought of doing an OpenParts wrapper for XEmacs myself, but I've never gotten around to it (haven't looked into it to see if it's feasable).
Of course a Qt or Gtk+ implementation of Emacs would also be nice, but a lot more work.
---------------------------
Well, you can use Qt/Python.
Ellen Ullman
-----------------------------
Do you have any evidence for this claim? Or is it just libel?
I guess it's convenient your anonymous.
----------------------------------------------
I agree with you that RAD development is not a substitute for design. But I do think that visual programming tools are particularly suited to gui development. It's very intuitive to move controls around and see what they look like.
Also, there is definitely a place for tools to throw together throw-away apps quickly. Witness perl, python, and other scripting languages. Admittely, they can be misused, but would it be a better world if everyone was forced to use compiled languages for everything? I think the same thing applies with gui tools. People like Tk because it's quick to write for. RAD environments offer the same advantage.
Having said all of that, I personally prefer writing C++/Qt apps with emacs.
-----------------------------------------------
I tend to agree. I do view works of programming as works of art (artifacts), but to call the whole "Open Source" process, as opposed a specific artifact like the Linux kernel, a work of art is streching it.
I think this guy has a post-modernist agenda that he is trying to sneak in under the Open Source hype. He compares patches to "sampling", a sees a lack of authorial viewpoint in OpenSource.
He may have a point, but it's overstated. Many of the more famous open source projects were in fact the results of the vision of one or a few persons -- Linux and Linus, Emacs and RMS, etc.
Programming is definitely art in the traditional sense, not just the post-modernists (non)-sense.
-----------------------------------------
Click and Clack, the Tap'it brothers (the car guys on NPR) are hawking Y2K compliant needle nose pliars, coffee cups, etc.
------------------------------------------
Doh!
Message to the humor impaired:
I think they wrote Feb 31 on purpose!
----------------------------------
Good post.
The term "revolutionary" is thrown around way too much, for things that are really evolutionary. On a technical level, at least, Linux was definitely not revolutionary.
---------------------------------
I'm not against certification. It's useful for non-programmers to figure out if someone is qualified or not. I'm not sure that an computer-engineering degree means that you're qualified, however. (But MSCE is a total joke). In California, you can take the bar without a law degree. If you pass, you're a lawyer. Also, in all states, just getting a law degree doesn't mean you can legally practice law. You have to pass the bar. Maybe the "software enginering" profession needs something similar.
Personally, I call my self a "hacker," or, for those who associate hacker with cracker, a programmer.
---------------------------
How do you define engineer? As someone who has an engineering degree? That's like saying a musician is someone who has a music degree. Some great musicians have music degrees, others don't. Many famous symphonic musicians only have high school degrees. The proof is in the pudding -- either you can play or you can't, and when you audition for an orchestra, they ask you to play. They don't ask to see your diploma.
Here's Webster's definition of engineer:
"A person who carries through an enterprise by skillful or artful contrivance."
I also read the letter in IEEE criticising Ellen Ulman's credentials. I thought it was unfair. Ulman has supervised and designed many successfull software enterprises through her skillful and artful contrivances. I think that fits the definition of engineer.
I prefer the term "hacker," however. To me, a hacker is someone who a person who carries through a software enterprise by skillful or artful contrivance.
-----------------------------------
What about Emacs? Are hackers moving away from it for hipper, newer editing environments, like Crisp, Visual Slick Edit, et al? Or the vi fanatics, are they moving on? What about Unix gurus in general? Did they all switch to OS/2, or Be, or whatever? Some of them, maybe, but for other people, when they get really comfortable with an old tool, they stick with it forever, and keep customizing and extending it.
Will the frame-buffer driver eventually address this?
I heard some people complain about the X consortium's control over X. Could we fork off a different branch of X, a 'la [X]?Emacs, and would that be a good idea?
A friend of mine says he doesn't like X's client/server over the network model. He says that this consumes unnecessary resources when everything (the X server and clients) are running locally. I said I thought any wasted resources were marginal. Anyone know for sure?