Re:Comment from a real PERL programmer
on
Perl6 for Mortals
·
· Score: 2
If you were to play towards java's strengths, you'd write a larger program where encapsulation maintainability, readability, and robustness are the most important factors.
All of which are conveniently unquantifiable. The code snippet provided is quite readable to and maintainable by Perl programmers. (By the way,/. seems to have swallowed a <> in the code). Contrary to the TIMTOWDI rhetoric, the code presented is pretty much how any experienced Perl programmer would do it. I'm not looking for a language that will use 3x more code to express the same ideas. I'm looking for a language that uses less code to express the same ideas. And I think Perl 6 will achieve that in some cases.
Perl is a higher level language than Java, just as Java is a higher level language than assembler. If you think there's something inherently good (or readable or robust) about using a lower level language (which means more explicit actions, more steps chosen from a smaller vocabulary) why not go all the way and use assembler?
Actually, for substantial parts of a web browser Perl would be a good choice. That includes fetching and caching resources, managing a cookie database and browse history, and implementing user policies (like ad blocking). It's really only the page rendering and GUI that would be awkward in Perl.
I agree that the average quality of posts has declined. However, this is a side effect of the massively increased user base. I look on the bright side, which is that slashdot propogates certain beneficial ideas to a much larger audience. Inevitably, more time is spent simply correcting the misconceptions (and, unfortunately, refusal to read) of newbies. But it's worth it. I sometimes think a small quiz should be required before posting. The quiz would be geared to the material at hand, and would include hyperlinks to places where you'd find the answers. As for the people who have switched to shiny new languages with less mindshare, I suppose they feel a bit of anxiety at the massive scope and momentum of Perl 6, just as we feel a bit of anxiety when Microsoft hypes.NET or whatever. It's quite possible that Perl 6 will negate whatever advantages $superlanguage once offered, just as Microsoft is nibbling away at the advantages Unix offers.
Granted, this was another legacy feature, but is still around because of the huffman encoding.
The huffman coding consideration hasn't gone away - therefore I don't see why $_ is legacy. It is an intuitive, human-centric way of recognizing focus. Which of the following is more intuitive:
Wash, wax and vacuum each car in the lot.
For each car CAR in the lot, wash CAR, wax CAR and vacuum CAR.
I don't see how (1) is more "cryptic" than (2). Anyone who objects to $_ ought to also object to having a current directory in his shell. If I type 'ls', isn't that cryptic? I'm not specifying which directory to ls.
I find $_ clean and elegant, especially when it's used implicitly. It removes the visual noise of variable names that didn't matter anyway.
And by the same token, I am pleased with the unary '.' operator as described in the article. I really don't like typing '$self->{ whatever }'; it's more repetitive noise.
Sounds like you are comparing apples and oranges. A leased line is a business-grade product, and hopefully has margins that can support some customization. When it comes to the retail customer, the internet is indeed maturing, but this means the opposite of what you said. Retail ISPs run on narrow margins and therefore must sell a cheap, reliable, acceptable service that meets the needs of the majority. If you find this patronizing (and you may be right) you probably need to escape the "retail customer" category. As a retail ISP customer, you can increasingly expect: no meaningful tech support, port 25 outbound blocked, dynamic address, port 80 inbound blocked when they feel like it, and mail filtered as the ISP sees fit to reduce spam. It has nothing to do with the egos of techies. The techies would love to give everyone a wide open T1. It has to do with offering a financially viable service and staying alive.
In practice this is going to mean the website is going to be up for a day or two after the spams start arriving in mail boxes.
I don't know what an actual reasonable time frame is. But it looks like spammers are increasingly drawn to "bulletproof" hosting - the kind that gets listed on MAPS RBL. That must mean that normal hosting reacts too quickly for their liking.
Which means that I will NEVER use an ISP which uses spews - and nor will most other business users.
First of all, most ISPs do not discuss specifics of spam blocking. Generally, large ISPs do not even tell their tech support. So I don't know how you propose to find out what lists or techniques an ISP is using before becoming their customer. Second, this is pretty far down on the list of things to care about. You are far more likely to suffer an outage because your ISP replaced its networking folks with trained monkeys than to lose money because you can't communicate with a listed IP address. Also, I don't know if PacBell is implementing the list on mailservers or on routers. I would guess the former, because I haven't seen a BGP feed for spews. In that case, as a business running your own mail server, you wouldn't be affected.
It's tricky enough making the case for blocking when the blocking is done openly and accountably...
I disagree. I think that all large ISP's maintain private blocklists. To whom do they need to make a case? To executives - they are cutting network traffic substantially, thus saving money. To users? Users generally don't know or care.
The mere fact spews feels they have to self-censor to avoid such problems means they're irrelevant.
Far from it. I don't get much spam, but occasionally I will run the originating IP of a spam past MAPS, spews, and the other DNS lists. Spews is the most accurate. MAPS almost never flags an IP as spamming. I have never gotten spam from a "mainsleazer". I think this is an entirely separate problem from "chickenboner" spam, which is what spews addresses. If you think spews is irrelevant you should read some of the pleas for delisting on nanae. They seem to have got the message across to some ISPs that were quite deaf to complaints.
You point out why a perfect blocklist is impossible. You're right. But meanwhile, people continue to improve these imperfect blocklists.
"We must boycott Sony. We must refuse to buy their tasty electronic gadgets, however tempting. We must - Oh! They're releasing the PS2 Linux kit in the US? Gotta go; I'm going to camp outside Fry's."
Seriously, we need to remember what Sony is next time they wave their enticements in our face.
Well, it's an interesting suggestion, but I don't quite see why the government would want to repeal the laws. Massive sneaky disobedience, especially of a nonviolent sort, sounds like a law enforcement wet dream. A huge reservoir full of tasty fish. They just apply the tactics of the War on Drugs (searches, informants, giving people reduced sentences in exchange for information) and they've got a steady stream of risk-free convictions, rising budgets, and increased public appreciation. And the "sneaky" nature of the disobedience makes it easy to villify the copiers.
Of course our booming prison industry would also benefit from a new stream of inmates.
However my approval is not required. What you described is what's going to happen, like it or not.
Ultimately I don't think it is possible for web hosting companies to offer economically viable "user" services which are also completely incapable of being exploited by spammers.
I don't think MAPS is expecting this. Rather, they ask that ISPs react to spam complaints in a reasonable timeframe and shut down the offending sites if appropriate. MAPS focuses on educating and establishing rapport with ISPs. It only uses RBL listing as a last resort. In fact, MAPS goes so far in giving ISPs the benefit of the doubt that their utility is substantially reduced. Spam ISPs will do a ton of damage before MAPS gets around to listing them. That is stroke one against MAPS.
Stroke two is that they appear to have backed down in the face of wealthy spammers like Experian and Harris. Unfortunately, there is no grassroots solution to that problem. These companies are rich enough to get the ear of decision-makers.
SPEWS solves problem #1 by abandoning all attempts at education. SPEWS does not meet any of your proposed criteria for a spam-block list: they are neither fair nor accountable. However PacBell appears to be using them, and I'm sure other organizations are. When it comes to problem #2, SPEWS appears to avoid listing the big "mainsleazers" like Harris. Therefore, your idea that hard-line lists would not be adopted by large organizations is not necessarily correct. As long as those lists never offend big corporations, they are probably quite palatable to large organizations as cost-savers, which they are.
Unfortunately, MAPS was the only organization that could communicate to the corporate world and convince them not to spam. I think it's become clear that anyone with deep pockets can sue MAPS and get taken off the RBL.
Exactly. I hope someone mods you up. I won't be upset if the lawyers destroy MAPS. It will just push the technical community towards more hard-line services like SPEWS. Eventually we might end up with a Freenet-like system, in which listings are crypto-signed but untraceable. In the mean time, Jamie, Bennet Haselton, and others like them try to whip up anti-MAPS sentiment by telling the story of evil Goliath vs. little David (who just happens to spam a bit on the side, so what?) It mostly impresses people who haven't researched subject much.
Your post presents an incomplete picture. The reason why spam-support IP addresses are listed is that spam is frequently sent either from throwaway accounts or via open relays. Thus, there is no originating IP to blacklist. But the spammers frequently depend on driving traffic to a web site. The most effective way to fight these spammers is to block access to their web sites. ISPs who deliberately harbor such web sites are outcasts. They are intentionally choosing to pollute the internet with unwanted garbage, with the consequence that other networks may refuse to carry their traffic. And when a spam ISP evades an IP address listing by moving the offending site to a different address, MAPS natually tends to list the whole block.
I wish that in answering someone's request for factual information you would include the appropriate context. Seen in that context, MAPS's actions appear more reasonable.
(2) run-away processes that consume resources to no end, until the system crawls...
You could set up a process monitor script that either runs as a daemon or via cron. If the same process is using 80% CPU or 80% memory two samples in a row, it would kill the process and pop an xmessage saying "$program was using too much memory, so I killed it."
I want Linux to continue to run on new, interesting hardware. There is an ongoing battle to get vendors to release their specs. There are also several ominous clouds on the horizon, SSSCA and TCPA (trusted computing platform architecture) which threaten to rain on our parade of cheap, open commodity hardware.
If even a small percentage of "normal" users use Linux, it will be nearly impossible for anyone to marginalize Linux and lock it out of new hardware.
Also, consider protocols. Imagine if Microsoft pushes us to a point where you need.NET to buy airline tickets, make hotel reservations, or file your tax returns. A substantial base of Linux users can apply enough pressure to keep these protocols open.
THe point is that if a user gets a virus, it can only damage their files. not anyone elses.
On most computers, the user's files are the only unique and important ones. And, as I pointed out, that safety vanishes if you have (and use) su or sudo access. Don't assume that the virus would immediately destroy as much as it could. Instead, it could wait to piggyback on your privileges and ride on to greater access.
I can put in last nights backup tapes and bring back my home.
You optimistically assume that the virus would make itself known within the first 24 hours. What if it's been there for six months? What if you only discovered it when you noticed that a program you wrote, which is in production, is periodically making mysterious outbound TCP connections? Unix permissions can buy some margin of safety, but they are not a cure-all.
I agree that Sun may not offer the best value for the money. But two of your comments were (perhaps involuntarily) quite PC-centric.
No decent desktop (do they really think anyone can be bothered with that joke desktop they supply, with no drag n drop, no file associations etc etc)
Why run X on a server at all? Most of what you need to do is command-line stuff. And if you must run a GUI app (like the Oracle installer), wouldn't you rather use your Linux box for the display? That way you can sit in your comfy cube, with your "desktop" set up the way you like instead of standing in a cold, noisy equipment room.
And then the complaint about monitor/keyboard. What you may not realize is that Suns do not favor the Keyboard/video over the serial console. You can do everything from the serial console port. In the largest Sun shop I've seen, there are no keyboards, monitors or KVM hardware connected to the servers. PC hardware has trained us to think that you always need KVM access to get to the BIOS or something, but computers that were designed to be servers don't have this weird limitation.
I don't really understand the point of the E450. As you point out, it can be replaced with a PC. And you have to unrack it to get to the CPU/RAM. But a shop full of 4500's and 6500's is very nice. They take the same CPU/RAM card, which can be installed/removed from the front without surgery. If a machine is running out of either, you can just add another board. Ideally you buy storage from a storage vendor rather than trying Sun's solutions. A datacenter full of 4500's and 6500's is amazingly low maintenance compared to equivalent PC's. Is it worth the cost delta? I don't know.
Many people have pointed out that this is the price of access to Hailstorm, not the price of access to.NET. I don't think the prices are high at all. In fact, they're ridiculously low for entry into the controlled space Microsoft is creating. Why don't you phone up Yahoo and Google and ask what they'd charge to let you put your button on their site? I don't think it will be as cheap as what MS is offering.
I'm seeing a lot of comments like "Programmers will.... " or "Programmers won't.... ". These posters don't seem to realize that the software industry is shaped by businesses, not programmers. If you are hired to work on a.NET-related project, you will not know or care what arrangement the company made with Microsoft. The audience that Microsoft is addressing will not balk at the price.
I agree. It's disturbing how much software lacks man pages. They are the standard documentation on Unix, and everyone making programs or libraries should furnish them. However, I think there are some weaknesses in the man format. I would like to see a more semantically structured format. For example, I'd like to type 'man -x ls' and see the five most common examples of ls invocation. Also, the man page and the program's getopt (or function proto for functions) should be generated from the same mechanism, so they can never be out of sync. Writing the man page shouldn't be a whole separate effort from writing the code - a lot of the info in the man page could be mechanically extracted from the source.
Having said that, you're still stuck browsing GTK. Have you tried using ctags(1)? After generating a ctags "database", you can hit ^] when the cursor is on a function call and you will jump to the function proto. In vim, I mean.
Since you ask, I argue that the people making Microsoft products are smarter than the people using Microsoft products. A framer building a wooden house might drive nails with a heavy framing hammer (emacs). But a home owner replacing a rotted board on his front steps might use a lighter hammer (IDE), even though it's not 'optimally efficient'. His ego is not linked to driving home the nail in two strokes. He just wants to get the job done, preferrably without smashing his thumb.
Can you imagine an average Windows application programmer developing his own elsip?
Of course, if the home owner advocates using a tack hammer to build a house, he's going to get laughed at.
but $HOME probably doesn't contain a lot of possible targets for a virus (that is, executable programs that could be infected).
How about.bashrc and.bash_profile? They are executed whenever you log in. And if that's not enough, imagine: alias su='$HOME/.evilprog su'; alias sudo = '$HOME/.evilprog sudo'. Next time you become root, the password would be transparently trapped. How about.mailcap? Or.vimrc? Face it - your home directory is full of places to plant hooks which will be executed under your userid. If the malicious code can run just once, it can implant itself permanently. From there, it could try to become root (not that it's really worthwhile) or just look for vectors to spread itself.
You're probably using a graphics card that does not have good 2D acceleration under X. The Matrox cards generally rock under X. Cheap cards can be quite horribly slow. Even if the same card is fast under Windows. Of course there are GUI apps that can make anything slow. Netscape, in its darker moments for instance.
Your OS has multi-user security, so that the system binaries aren't affected.
I don't mean to understate the importance of security, but this oft-repeated idea of protecting the OS while losing $HOME is out of sync with modern reality. I can reinstall the OS in 20-30 minutes. But $HOME could contain files I'll never be able to recreate.
Bottom line: when viruses really hit Linux, we're going to go through a major, painful adjustment.
Slashdot Mirror
All of which are conveniently unquantifiable. The code snippet provided is quite readable to and maintainable by Perl programmers. (By the way,
Perl is a higher level language than Java, just as Java is a higher level language than assembler. If you think there's something inherently good (or readable or robust) about using a lower level language (which means more explicit actions, more steps chosen from a smaller vocabulary) why not go all the way and use assembler?
Actually, for substantial parts of a web browser Perl would be a good choice. That includes fetching and caching resources, managing a cookie database and browse history, and implementing user policies (like ad blocking). It's really only the page rendering and GUI that would be awkward in Perl.
I agree that the average quality of posts has declined. However, this is a side effect of the massively increased user base. I look on the bright side, which is that slashdot propogates certain beneficial ideas to a much larger audience. Inevitably, more time is spent simply correcting the misconceptions (and, unfortunately, refusal to read) of newbies. But it's worth it. I sometimes think a small quiz should be required before posting. The quiz would be geared to the material at hand, and would include hyperlinks to places where you'd find the answers. As for the people who have switched to shiny new languages with less mindshare, I suppose they feel a bit of anxiety at the massive scope and momentum of Perl 6, just as we feel a bit of anxiety when Microsoft hypes .NET or whatever. It's quite possible that Perl 6 will negate whatever advantages $superlanguage once offered, just as Microsoft is nibbling away at the advantages Unix offers.
#!/usr/bin/perl -wTn
print ucfirst($_);
exit;
The huffman coding consideration hasn't gone away - therefore I don't see why $_ is legacy. It is an intuitive, human-centric way of recognizing focus. Which of the following is more intuitive:
- Wash, wax and vacuum each car in the lot.
- For each car CAR in the lot, wash CAR, wax CAR and vacuum CAR.
I don't see how (1) is more "cryptic" than (2). Anyone who objects to $_ ought to also object to having a current directory in his shell. If I type 'ls', isn't that cryptic? I'm not specifying which directory to ls.I find $_ clean and elegant, especially when it's used implicitly. It removes the visual noise of variable names that didn't matter anyway.
And by the same token, I am pleased with the unary '.' operator as described in the article. I really don't like typing '$self->{ whatever }'; it's more repetitive noise.
Sounds like you are comparing apples and oranges. A leased line is a business-grade product, and hopefully has margins that can support some customization. When it comes to the retail customer, the internet is indeed maturing, but this means the opposite of what you said. Retail ISPs run on narrow margins and therefore must sell a cheap, reliable, acceptable service that meets the needs of the majority. If you find this patronizing (and you may be right) you probably need to escape the "retail customer" category. As a retail ISP customer, you can increasingly expect: no meaningful tech support, port 25 outbound blocked, dynamic address, port 80 inbound blocked when they feel like it, and mail filtered as the ISP sees fit to reduce spam. It has nothing to do with the egos of techies. The techies would love to give everyone a wide open T1. It has to do with offering a financially viable service and staying alive.
I don't know what an actual reasonable time frame is. But it looks like spammers are increasingly drawn to "bulletproof" hosting - the kind that gets listed on MAPS RBL. That must mean that normal hosting reacts too quickly for their liking.
First of all, most ISPs do not discuss specifics of spam blocking. Generally, large ISPs do not even tell their tech support. So I don't know how you propose to find out what lists or techniques an ISP is using before becoming their customer. Second, this is pretty far down on the list of things to care about. You are far more likely to suffer an outage because your ISP replaced its networking folks with trained monkeys than to lose money because you can't communicate with a listed IP address. Also, I don't know if PacBell is implementing the list on mailservers or on routers. I would guess the former, because I haven't seen a BGP feed for spews. In that case, as a business running your own mail server, you wouldn't be affected.
I disagree. I think that all large ISP's maintain private blocklists. To whom do they need to make a case? To executives - they are cutting network traffic substantially, thus saving money. To users? Users generally don't know or care.
Far from it. I don't get much spam, but occasionally I will run the originating IP of a spam past MAPS, spews, and the other DNS lists. Spews is the most accurate. MAPS almost never flags an IP as spamming. I have never gotten spam from a "mainsleazer". I think this is an entirely separate problem from "chickenboner" spam, which is what spews addresses. If you think spews is irrelevant you should read some of the pleas for delisting on nanae. They seem to have got the message across to some ISPs that were quite deaf to complaints.
You point out why a perfect blocklist is impossible. You're right. But meanwhile, people continue to improve these imperfect blocklists.
"We must boycott Sony. We must refuse to buy their tasty electronic gadgets, however tempting. We must - Oh! They're releasing the PS2 Linux kit in the US? Gotta go; I'm going to camp outside Fry's."
Seriously, we need to remember what Sony is next time they wave their enticements in our face.
Well, it's an interesting suggestion, but I don't quite see why the government would want to repeal the laws. Massive sneaky disobedience, especially of a nonviolent sort, sounds like a law enforcement wet dream. A huge reservoir full of tasty fish. They just apply the tactics of the War on Drugs (searches, informants, giving people reduced sentences in exchange for information) and they've got a steady stream of risk-free convictions, rising budgets, and increased public appreciation. And the "sneaky" nature of the disobedience makes it easy to villify the copiers.
Of course our booming prison industry would also benefit from a new stream of inmates.
However my approval is not required. What you described is what's going to happen, like it or not.
I don't think MAPS is expecting this. Rather, they ask that ISPs react to spam complaints in a reasonable timeframe and shut down the offending sites if appropriate. MAPS focuses on educating and establishing rapport with ISPs. It only uses RBL listing as a last resort. In fact, MAPS goes so far in giving ISPs the benefit of the doubt that their utility is substantially reduced. Spam ISPs will do a ton of damage before MAPS gets around to listing them. That is stroke one against MAPS.
Stroke two is that they appear to have backed down in the face of wealthy spammers like Experian and Harris. Unfortunately, there is no grassroots solution to that problem. These companies are rich enough to get the ear of decision-makers.
SPEWS solves problem #1 by abandoning all attempts at education. SPEWS does not meet any of your proposed criteria for a spam-block list: they are neither fair nor accountable. However PacBell appears to be using them, and I'm sure other organizations are. When it comes to problem #2, SPEWS appears to avoid listing the big "mainsleazers" like Harris. Therefore, your idea that hard-line lists would not be adopted by large organizations is not necessarily correct. As long as those lists never offend big corporations, they are probably quite palatable to large organizations as cost-savers, which they are.
Unfortunately, MAPS was the only organization that could communicate to the corporate world and convince them not to spam. I think it's become clear that anyone with deep pockets can sue MAPS and get taken off the RBL.
Exactly. I hope someone mods you up. I won't be upset if the lawyers destroy MAPS. It will just push the technical community towards more hard-line services like SPEWS. Eventually we might end up with a Freenet-like system, in which listings are crypto-signed but untraceable. In the mean time, Jamie, Bennet Haselton, and others like them try to whip up anti-MAPS sentiment by telling the story of evil Goliath vs. little David (who just happens to spam a bit on the side, so what?) It mostly impresses people who haven't researched subject much.
Probably the subnet. Although MAPS seems to have unlisted it, SPEWS is still listing it. See the file here.
Your post presents an incomplete picture. The reason why spam-support IP addresses are listed is that spam is frequently sent either from throwaway accounts or via open relays. Thus, there is no originating IP to blacklist. But the spammers frequently depend on driving traffic to a web site. The most effective way to fight these spammers is to block access to their web sites. ISPs who deliberately harbor such web sites are outcasts. They are intentionally choosing to pollute the internet with unwanted garbage, with the consequence that other networks may refuse to carry their traffic. And when a spam ISP evades an IP address listing by moving the offending site to a different address, MAPS natually tends to list the whole block.
I wish that in answering someone's request for factual information you would include the appropriate context. Seen in that context, MAPS's actions appear more reasonable.
You could set up a process monitor script that either runs as a daemon or via cron. If the same process is using 80% CPU or 80% memory two samples in a row, it would kill the process and pop an xmessage saying "$program was using too much memory, so I killed it."
I want Linux to continue to run on new, interesting hardware. There is an ongoing battle to get vendors to release their specs. There are also several ominous clouds on the horizon, SSSCA and TCPA (trusted computing platform architecture) which threaten to rain on our parade of cheap, open commodity hardware.
.NET to buy airline tickets, make hotel reservations, or file your tax returns. A substantial base of Linux users can apply enough pressure to keep these protocols open.
If even a small percentage of "normal" users use Linux, it will be nearly impossible for anyone to marginalize Linux and lock it out of new hardware.
Also, consider protocols. Imagine if Microsoft pushes us to a point where you need
I don't think floppies will still be in use by the time most of us are grey.
They seem to accept anything matching
Mozilla.*\(compatible; MSIE [4-6]\.\d+.*.
On most computers, the user's files are the only unique and important ones. And, as I pointed out, that safety vanishes if you have (and use) su or sudo access. Don't assume that the virus would immediately destroy as much as it could. Instead, it could wait to piggyback on your privileges and ride on to greater access.
You optimistically assume that the virus would make itself known within the first 24 hours. What if it's been there for six months? What if you only discovered it when you noticed that a program you wrote, which is in production, is periodically making mysterious outbound TCP connections? Unix permissions can buy some margin of safety, but they are not a cure-all.
Why run X on a server at all? Most of what you need to do is command-line stuff. And if you must run a GUI app (like the Oracle installer), wouldn't you rather use your Linux box for the display? That way you can sit in your comfy cube, with your "desktop" set up the way you like instead of standing in a cold, noisy equipment room.
And then the complaint about monitor/keyboard. What you may not realize is that Suns do not favor the Keyboard/video over the serial console. You can do everything from the serial console port. In the largest Sun shop I've seen, there are no keyboards, monitors or KVM hardware connected to the servers. PC hardware has trained us to think that you always need KVM access to get to the BIOS or something, but computers that were designed to be servers don't have this weird limitation.
I don't really understand the point of the E450. As you point out, it can be replaced with a PC. And you have to unrack it to get to the CPU/RAM. But a shop full of 4500's and 6500's is very nice. They take the same CPU/RAM card, which can be installed/removed from the front without surgery. If a machine is running out of either, you can just add another board. Ideally you buy storage from a storage vendor rather than trying Sun's solutions. A datacenter full of 4500's and 6500's is amazingly low maintenance compared to equivalent PC's. Is it worth the cost delta? I don't know.
Many people have pointed out that this is the price of access to Hailstorm, not the price of access to .NET. I don't think the prices are high at all. In fact, they're ridiculously low for entry into the controlled space Microsoft is creating. Why don't you phone up Yahoo and Google and ask what they'd charge to let you put your button on their site? I don't think it will be as cheap as what MS is offering.
.... " or "Programmers won't .... ". These posters don't seem to realize that the software industry is shaped by businesses, not programmers. If you are hired to work on a .NET-related project, you will not know or care what arrangement the company made with Microsoft. The audience that Microsoft is addressing will not balk at the price.
I'm seeing a lot of comments like "Programmers will
I agree. It's disturbing how much software lacks man pages. They are the standard documentation on Unix, and everyone making programs or libraries should furnish them. However, I think there are some weaknesses in the man format. I would like to see a more semantically structured format. For example, I'd like to type 'man -x ls' and see the five most common examples of ls invocation. Also, the man page and the program's getopt (or function proto for functions) should be generated from the same mechanism, so they can never be out of sync. Writing the man page shouldn't be a whole separate effort from writing the code - a lot of the info in the man page could be mechanically extracted from the source.
Having said that, you're still stuck browsing GTK. Have you tried using ctags(1)? After generating a ctags "database", you can hit ^] when the cursor is on a function call and you will jump to the function proto. In vim, I mean.
Since you ask, I argue that the people making Microsoft products are smarter than the people using Microsoft products. A framer building a wooden house might drive nails with a heavy framing hammer (emacs). But a home owner replacing a rotted board on his front steps might use a lighter hammer (IDE), even though it's not 'optimally efficient'. His ego is not linked to driving home the nail in two strokes. He just wants to get the job done, preferrably without smashing his thumb.
Can you imagine an average Windows application programmer developing his own elsip?
Of course, if the home owner advocates using a tack hammer to build a house, he's going to get laughed at.
How about
You're probably using a graphics card that does not have good 2D acceleration under X. The Matrox cards generally rock under X. Cheap cards can be quite horribly slow. Even if the same card is fast under Windows. Of course there are GUI apps that can make anything slow. Netscape, in its darker moments for instance.
I don't mean to understate the importance of security, but this oft-repeated idea of protecting the OS while losing $HOME is out of sync with modern reality. I can reinstall the OS in 20-30 minutes. But $HOME could contain files I'll never be able to recreate.
Bottom line: when viruses really hit Linux, we're going to go through a major, painful adjustment.