Interoperability with everyone by forcing them to use the formats you use, instead of using formats that were created to be platform independent.
I wrote my resume in HTML and put it on my website. When I link to it from an email to the HR department I can see how many times they view it. It's a great way of judging the effectiveness. That'll work if I apply to a company with Windows computers, or Macs, or Sun desktops. All of which are a possibility, or even like the one animation company I applied to recently as a sysadmin, redhat 8 on every computer. The HR guy was using Open Office to view resumes but said how he liked just clicking on mine in email and loading it up in the browser instead of loading a seperate program. (I've never seen a desktop in the last year where the user didn't have the browser open.)
If nobody fights it we'll all be forced to buy a subscription to Office in order to update our resume to find a job. Or to open files we created last year, before our old copy of the program expired. Monopolies are bad for business and bad for customers.
Also, there are formats designed for platform-independent exchange of data. Word files are not, and they have a ton of annoying features (containing deleted text, being modified by simply opening them, etc) Even in the office where it's mainly Windows to Windows I tend to convert documents to HTML, or PDF where required, and put them on the internal website.
There are ways to setup defaults for new users, but neither is very handy. Either, edit the default user folder and add specific menus and such (but this doesn't do themes and stuff) or if you run a domain and "properly" create users for it, you can do some of this. I'm a little vague on the details but I watched our MS admin do it and he knew a lot of tricks. If you get an MCSE I assume they cover this, but there aren't a lot of howto pages on the net for it, or at least they're not as nicely searchable because I can't find them when I need them.
Yeah, the only times Linux crashed for me recently was when the fan on my server's second CPU stopped spinning, and when I was overclocking my desktop machine and didn't have it completely stable. (My DDR333 wouldn't do 400 for long.)
Write an HTML resume and name it.doc, Word will view it and they won't notice.
But nobody specifies word resumes these days, at least for anything beyond secretarial stuff. Just any email viewable resume. I always send a link to the resume on my website and a an html copy renamed to.doc and zipped up for the people who can't handle the new millenium. I almost always get hits on my website (worth doing just for that) to the resume and I've never heard a complaint. I'll show up for an interview and see a printout with the IE or Mozilla header at the top.
Helps that I'm applying for skills where something other than Word are required though.
There is just such a feature. The printer you have defined changes some of the settings and makes the document appear a bit different. Potentially helpful, but in these days of primarily wanting electronic copies with the printing as a handy fallback, very annoying.
Yeah, PDF is the way to go when alignment must be right. And HTML when alignment isn't as important. PDFs are fairly portable but nowhere near as much (or as easily) as HTMl. HTML is much easier for others to work with too.
IE is only faster because it's pre-loaded and never seems to get completely swapped out. I've seen a situation where coming back from Quake3 at a lan party Notepad swapped to come up, but IE was there instantly. They can't possibly make a browser with less footprint than notepad so I assume they simply tweak swapping to keep IE available. Once IE and Mozilla are both up and fully out of cache the performance is only noticable when rendering test pages (tables in tables in tables, etc). I find that IE is faster at loading large pictures 2500x1800 or so, but that's the only noticable speed difference for me in general usage.
btw, if you're an IE user, perhaps you can tell me how to make IE create new windows full-screened. I always full-screen my browsers and it's a pain that all new IE windows open using like a ninth of my screen. No matter if I close it while full-screened or anything.
Pah, windows still breaks, and in pathetic ways. If the automatic unzip in WinXP locks up and you kill it you'll lose all your explorer (file explorer, not internet, I don't use IE so I don't know) windows and your task bar will vanish for a while. Then it'll come back without any of you tray icons.
Happened to me the same day I installed WinXP. Not great for their most stable OS yet, but lightyears ahead of Win9x I'll give you.
Windows doesn't randomly bluescreen as much but if you do more than just web browse and office work it still dies fairly often. At my work where the Windows users are running Visual Studio and developing an MMC plug-in they tend to get a week or so between crashes or some incapacitating system failure. This is on good hardware too, one of those same batch of machines has been a 24/7 Linux server for the last year.
Don't ignore the planned sales. Some of SCO's statements indicate that they were planning this pump and dump since at least the beginning of last year. Those "planned" sales were to roughly coincide with the announcements.
We need real-world benchmarks then. Like AnandTech's recording of their website's DB traffic. They replay an hour of moderate traffic as fast as possible, seeing how fast easy config can handle it. It's directly relevant to their forum-heavy situation. Similar sites could probably depend on the results. And, even if not, it's a real-world test, not some synthesized benchmark that tests a task nobody'd ever do.
It's not as much that Joe Consumer is stupid, it's that Slashdot isn't popular with the non-tech crowd. How is anyone going to hear about the DRM in Intuit's product, and the fact that the other companies are better in this area, unless a news site carries a story from someone who has already been burned by it?
It takes a mainstream media carrying a "Product X" screws you before it reaches most people. Once it does, or course they'll consider switching. Nobody likes restrictions like using software only on one computer, or annoying dongles that you need to swap from machine to machine.
The news made a big deal of it, but it was the customers who didn't buy the product. As another poster said, not because they give a rat's ass about the customer, but because they like their cushy executive bonuses. I see no need to help the company in the future. Everyone else might be just as bad, but we caught these ones actively trying to sell us out, why on Earth would we want to give them another chance? They're not sorry for bothering people, they're sorry it didn't raise profits.
If every piece of software was GPLed I'd still have a job. I customize existing software so that it fits my company's needs better. One guy I worked with spent a few weeks tweaking Bugzilla, that paid his rent and kept him off welfare, yet it was GPLed software.
And yeah, it doesn't harm me if someone uses my code without paying or helping out the community but it's a freeloading thing to do and I don't like freeloaders. You wouldn't want to support me if I decided to live on welfare, why should I write software and let some company sell it if they aren't willing to give something back? I could, I just don't feel that it benefits anyone who deserves it.
I really think that even if all software was public domain (more then GPLed) with no hope of copyright, that most programmers would still have a job. People want products, Products need code, companies make and sell products. The economics would change but code doesn't write itself, someone would have to and I'm one of those someones.
Besides, if a company can't pay huge salaries to a crack team of coders and come out with something hacked together by a bunch of volunteers, they don't deserve to make money.
The ac kernel isn't a fork because it's based on the main kernel and moves with it. It's like my kernel isn't a fork just because I compile with a few patches. A fork is when you diverge with no intent to go back.
And I don't think RMS had to convince people that corporate is evil, simply that free is the best choice for their own work. I work for a company writing proprietary code by day. It doesn't "want" freedom, nor do I care if it gets it. But the code I write at night I write for people to share. Sharing code to me means that you can add to it, but your changes must remain open. It'd be like if I wrote a song, you made a few good changes to it, and then you sang a song that was 95% mine, but didn't let me sing your version.
That's what companies do with source code, when allowed, because everything they do has to go past legal and the lawyers will always choose to do nothing, when possible, because it's the road of least resistance. Thus, I believe you can't really share with companies (and expect them to share with you) unless you require it.
It's not often you can cut and paste working code from googling on the net, at best it'll use different variable names. You pretty much need to rewrite everything anyways, so not being able to paste from the net isn't much of a pain. (And anything larger than a few lines that you download is probably a library you'll want to have added to the main project.)
If they'd actually code it securely you could hack it but you wouldn't gain extra info because it wouldn't send anything that you weren't supposed to see. (Well, perhaps a packet or two before the person reached the door, but not completely on the other side of a wall or anything.)
Their lame security via code obscurity and anti-debugger tricks just mean that anyone who uses a debugger for legit reasons gets annoyed when the game won't play, but doesn't actually help security. Even their constant updates (Q3 as an example) doesn't stop the cheats, it only makes them come out with new versions every few weeks. Perhaps they're a bit harder to find for people who haven't mastered google searching, but for anyone willing to spend an hour looking, they're just as easy to find.
The code that verifies CD-keys has two parts. There's in in-game stuff which is pretty simple. Correct number of digits, totals to an even multiple of 31337, or something. Just enough to catch people who incorrectly enter the key. Then, the real check happens on their keyservers. You try to log into a game and your client sends your encrypted key to the keyserver, which decrypts the key, checks it, and tells the server to let you play, or to kick you out.
The keyserver doesn't use a fancy algorithm, it simply looks up the key in a list of every key they've shipped. No fancy math, nothing. But, almost impossible to hack because you need to guess a real key, not simply fool an algorithm you can watch with a debugger.
But it does work. If you pick a sensible goal and stick to it, you'll get something done. If you shoot for the moon you'll be fighting your employees every step of the way and you won't get as far as if you'd adopted a more modest system.
I've seen projects where things are reasonably documented. My current company is using doxygen for turning inline comments into a web interface describing the project. It's got a very low barrier to entry (use an extra / in comments you want to show up in the web interface) and the fact that you get results immediately means it's easier to get people to continue.
I don't have to be bitten by a dog to know that it hurts. Similarly, I've never tried keeping documentation in a seperate file from the code but I know it's not an optimal solution.
Hell, it's hard enough to get people to keep comments in sync with code when it's in the same file, let alone opening another whole program and finding the appropriate area to comment. The important issue is if it's easy enough to do that it's practial, and if there's anything else that would be more helpful for the ammount of work required.
Sure, if done religiously, it may prove helpful. That's a bit of a tautology though, like saying the key to winning a fight is not getting hit. There's truth in that and it does do things like suggest judo or aikido over karate, but it's not like you can simply decide to not get hit and it won't happen. Similarly, you can't simply say "I'm going to follow this properly". People don't skip comments because they're lazy, people skip comments because there's always deadline pressure, and because they want to write the comments once when done, not at every incremental revision. It usually comes down to doing things the "right" way, as taught in comp-sci, or actually getting the project finished before bankruptcy.
Better instead to pick something that is reasonable. Especially because I've seen projects that were commented properly - new programmers could come in and use the APIs without having to ask implementation details. This was accomplished by documenting the code, not by writing design documents. It was fairly easy and because of this, never got pushed off until later. You could always take ten minutes and update the comments on a day's work, so the programmer resistance to it was very low. Being that the simple answer works, why go looking to make it more complex?
Re:"Required" email
on
Software Fashion
·
· Score: 3, Insightful
A method that proposes comments that don't live in the code is broken. It requires programmers to have a second file open, and to update two things every time they make a change. A system that requires an extra annoying step for absolutely no gain is defective.
For absolutely no gain? Yes. There are better ways, such as putting any needed documentation into the source code itself. That way not only are they more accessible when reading the code but they're easier to change and harder to forget about.
Check out doxygen (at sourceforge) for a pretty cool system.
I mostly agree with your detractors, but I'll sumarize. I think that Hungarian notation could be useful while scanning new code, but new code (not just a routine I didn't write, but a whole new project I don't know the structure of) isn't something I encounter often at work.
However, I think that it's fairly easy to read new code. Just don't bother yourself with what particular kind of string is being referenced, it's enough to know it's a string and that they're matching it for a pattern.
If you're asking yourself what type a variable is all the time, it seems that you're doing something wrong. It should be obvious because of context, because of the project (use only standard string, etc), or because you can see a declaration either at the top of the function or in the appropriate header file.
I've wasted probably less than an hour total, ever, needing to look up the type of a variable. It's just by far the least frequent problem. I'd have wasted much more time typing lpwstdstrFoo instead of Foo.
I don't think the OEM thing has any importance. Microsoft is free to sell their product to anyone they want, and only under certain conditions, but once that product ends up in your hands either by purchase or salvage, it's yours.
And yeah, their new hardware-locked installations are a real pain. My company stayed with 2k to avoid the hassle and everyone else I know simply pirates it. Their sales figures which are so impressive are based on OEM sales, their retail sales have to be almost nonexistant. (I bet they count every machine the OEM sells, even if it was shipped with a blank HD, as an XP sale.)
I think the moderation was a joke, intentionally...
Just like me saying "I've done that and it worked, but until I scrubbed the platters with steel wool they felt a bit sticky."
I actually did do a media swap before to recover data, it was a 5.25 disk for my Apple2 that someone spilled orange juice on. Open it up, wash it off in warm water, sacrifice another blank disk to be the carrier, and copy the data to a working disk immediately. Worked fine, but I wouldn't want to try it with a HD.
If you find a Win2k CD in the trash, having been used by someone who purchased WinXP, you're fine to use it. You didn't copy it in an unauthorized way and as such, it's pretty much the safe as if you bought it. Despite what people say, purchasing software from a store doesn't involve any license at all. Just you and copyright law. Upgrades may be a special area, but even if they are, it's the purchaser who didn't live up to their obligations and their copy of WinXP shouldn't be used, your copy from the trash is perfectly legit.
If you find a HD with Win2k on it, you're in a bit of a grayer area. That copy is legitimate, in that it was required to use the copyrighted work and thus was exempt from the general no-copies rule. Once a new copy was made on a new HD, the old copy (that you have) is no longer required, and thus making it wouldn't be allowed. But, you're not making the copy, only using it... I think the violation comes when the original purchaser made the second copy without destroying the first. Technically, this made a copy (one or the other) which was not required to properly use the software, and thus was a violation at this point.
I believe the law essentially says though that it's the making of the unauthorized copies (and selling, etc) that's the problem. Posession, without proof of intent to sell or anything, should be legal. Thus, using Win2k, however you got it, should be fine.
I'm not a lawyer and even if I was, this is very much up in the air thanks to megabucks corps paying for changes to the law and judges (Kaplan) to interpret it in their favor.
RPM dependency hell isn't as much a problem as using the wrong tool for the job. If you tried to install IE6, one file at a time, you'd find it to be a problem too.
There are two solutions, if you don't use either it's probably because you go to admin-targetted websites like FreshRPMs.net and use server-oriented distros, instead of using a user-oriented distro like Xandros, Mandrake, or even Redhat with Ximian(?)...
The solution is to bundle everything you may need, similar to the way Windows programs would come with VBRUNx00.DLL because not everyone had them, despite making the download bigger for the 90% of people who already had them. A reasonable solution for end-user targetted packages, and one that many packages like Mozilla do. You need the correct system libraries (libc, etc) but that's pretty much the equivalent of making sure you're using the right version on Windows.
The other solution, and probably the best one in the end, is to use tools like apt-get, emerge, the bsd-ports system, or urpmi, where you simply select the package you want and all the dependencies are resolved, downloaded, and where needed, built. This hasn't been as widely adopted as it should, but it's starting to be. And really, Grandma shouldn't be picking a Linux distro at random, like she shouldn't be running Win2k Advanced Server. If she has a user-oriented distro, she's probably never experienced RPM dependency hell, even if she tries to install something. (And some OSes like Xandros make this very easy, even offering handy lists of what you might want to install, like having shareware.com built into the OS.)
Just FYI. Sounds like you might be looking for a better solution and they do exist.
Everyone who haughtily looks down on it because it's not a real unix, or because it's new, is an idiot.
If they evaluate it and it doesn't meet their needs, or doesn't run on their hardware as well as the default OS, or they just don't like the default system font, they're free to choose not to run it. To dismiss it out of hand though, because of the association of RMS, or Communism, or "free-can't-be'good" is a stupid move which, pretty much by definition would make them an idiot.
I actually lost a job opportunity once because I put Linux experience with my other Unix experience. They thought I was unskilled because I used a "toy" system. Linux was the most popular webserver OS, FreeBSD ran CDRom.com, etc. "Toy" OSes were taking over the world and they sat in their ivory tower and declared them and everyone who used them to be playing around. They were idiots. (For the record, they were doing data processing, of the results from their big-iron number crunching. Linux could have done it fine but I'd really have stuck with IRIX which they already had instead of having them buy a PC and so on, or try to install Linux on their existing hardware. I lost the job in the initial phone interview, long before they asked for any advice.)
But, there are valid reasons for not picking Linux. Most people don't use them though.
Slashdot Mirror
Interoperability with everyone by forcing them to use the formats you use, instead of using formats that were created to be platform independent.
I wrote my resume in HTML and put it on my website. When I link to it from an email to the HR department I can see how many times they view it. It's a great way of judging the effectiveness. That'll work if I apply to a company with Windows computers, or Macs, or Sun desktops. All of which are a possibility, or even like the one animation company I applied to recently as a sysadmin, redhat 8 on every computer. The HR guy was using Open Office to view resumes but said how he liked just clicking on mine in email and loading it up in the browser instead of loading a seperate program. (I've never seen a desktop in the last year where the user didn't have the browser open.)
If nobody fights it we'll all be forced to buy a subscription to Office in order to update our resume to find a job. Or to open files we created last year, before our old copy of the program expired. Monopolies are bad for business and bad for customers.
Also, there are formats designed for platform-independent exchange of data. Word files are not, and they have a ton of annoying features (containing deleted text, being modified by simply opening them, etc) Even in the office where it's mainly Windows to Windows I tend to convert documents to HTML, or PDF where required, and put them on the internal website.
There are ways to setup defaults for new users, but neither is very handy. Either, edit the default user folder and add specific menus and such (but this doesn't do themes and stuff) or if you run a domain and "properly" create users for it, you can do some of this. I'm a little vague on the details but I watched our MS admin do it and he knew a lot of tricks. If you get an MCSE I assume they cover this, but there aren't a lot of howto pages on the net for it, or at least they're not as nicely searchable because I can't find them when I need them.
Yeah, the only times Linux crashed for me recently was when the fan on my server's second CPU stopped spinning, and when I was overclocking my desktop machine and didn't have it completely stable. (My DDR333 wouldn't do 400 for long.)
Write an HTML resume and name it .doc, Word will view it and they won't notice.
.doc and zipped up for the people who can't handle the new millenium. I almost always get hits on my website (worth doing just for that) to the resume and I've never heard a complaint. I'll show up for an interview and see a printout with the IE or Mozilla header at the top.
But nobody specifies word resumes these days, at least for anything beyond secretarial stuff. Just any email viewable resume. I always send a link to the resume on my website and a an html copy renamed to
Helps that I'm applying for skills where something other than Word are required though.
There is just such a feature. The printer you have defined changes some of the settings and makes the document appear a bit different. Potentially helpful, but in these days of primarily wanting electronic copies with the printing as a handy fallback, very annoying.
Yeah, PDF is the way to go when alignment must be right. And HTML when alignment isn't as important. PDFs are fairly portable but nowhere near as much (or as easily) as HTMl. HTML is much easier for others to work with too.
IE is only faster because it's pre-loaded and never seems to get completely swapped out. I've seen a situation where coming back from Quake3 at a lan party Notepad swapped to come up, but IE was there instantly. They can't possibly make a browser with less footprint than notepad so I assume they simply tweak swapping to keep IE available. Once IE and Mozilla are both up and fully out of cache the performance is only noticable when rendering test pages (tables in tables in tables, etc). I find that IE is faster at loading large pictures 2500x1800 or so, but that's the only noticable speed difference for me in general usage.
btw, if you're an IE user, perhaps you can tell me how to make IE create new windows full-screened. I always full-screen my browsers and it's a pain that all new IE windows open using like a ninth of my screen. No matter if I close it while full-screened or anything.
Pah, windows still breaks, and in pathetic ways. If the automatic unzip in WinXP locks up and you kill it you'll lose all your explorer (file explorer, not internet, I don't use IE so I don't know) windows and your task bar will vanish for a while. Then it'll come back without any of you tray icons.
Happened to me the same day I installed WinXP. Not great for their most stable OS yet, but lightyears ahead of Win9x I'll give you.
Windows doesn't randomly bluescreen as much but if you do more than just web browse and office work it still dies fairly often. At my work where the Windows users are running Visual Studio and developing an MMC plug-in they tend to get a week or so between crashes or some incapacitating system failure. This is on good hardware too, one of those same batch of machines has been a 24/7 Linux server for the last year.
Don't ignore the planned sales. Some of SCO's statements indicate that they were planning this pump and dump since at least the beginning of last year. Those "planned" sales were to roughly coincide with the announcements.
In fact, perhaps there's fraud in that as well...
We need real-world benchmarks then. Like AnandTech's recording of their website's DB traffic. They replay an hour of moderate traffic as fast as possible, seeing how fast easy config can handle it. It's directly relevant to their forum-heavy situation. Similar sites could probably depend on the results. And, even if not, it's a real-world test, not some synthesized benchmark that tests a task nobody'd ever do.
It's not as much that Joe Consumer is stupid, it's that Slashdot isn't popular with the non-tech crowd. How is anyone going to hear about the DRM in Intuit's product, and the fact that the other companies are better in this area, unless a news site carries a story from someone who has already been burned by it?
It takes a mainstream media carrying a "Product X" screws you before it reaches most people. Once it does, or course they'll consider switching. Nobody likes restrictions like using software only on one computer, or annoying dongles that you need to swap from machine to machine.
The news made a big deal of it, but it was the customers who didn't buy the product. As another poster said, not because they give a rat's ass about the customer, but because they like their cushy executive bonuses. I see no need to help the company in the future. Everyone else might be just as bad, but we caught these ones actively trying to sell us out, why on Earth would we want to give them another chance? They're not sorry for bothering people, they're sorry it didn't raise profits.
If every piece of software was GPLed I'd still have a job. I customize existing software so that it fits my company's needs better. One guy I worked with spent a few weeks tweaking Bugzilla, that paid his rent and kept him off welfare, yet it was GPLed software.
And yeah, it doesn't harm me if someone uses my code without paying or helping out the community but it's a freeloading thing to do and I don't like freeloaders. You wouldn't want to support me if I decided to live on welfare, why should I write software and let some company sell it if they aren't willing to give something back? I could, I just don't feel that it benefits anyone who deserves it.
I really think that even if all software was public domain (more then GPLed) with no hope of copyright, that most programmers would still have a job. People want products, Products need code, companies make and sell products. The economics would change but code doesn't write itself, someone would have to and I'm one of those someones.
Besides, if a company can't pay huge salaries to a crack team of coders and come out with something hacked together by a bunch of volunteers, they don't deserve to make money.
The ac kernel isn't a fork because it's based on the main kernel and moves with it. It's like my kernel isn't a fork just because I compile with a few patches. A fork is when you diverge with no intent to go back.
And I don't think RMS had to convince people that corporate is evil, simply that free is the best choice for their own work. I work for a company writing proprietary code by day. It doesn't "want" freedom, nor do I care if it gets it. But the code I write at night I write for people to share. Sharing code to me means that you can add to it, but your changes must remain open. It'd be like if I wrote a song, you made a few good changes to it, and then you sang a song that was 95% mine, but didn't let me sing your version.
That's what companies do with source code, when allowed, because everything they do has to go past legal and the lawyers will always choose to do nothing, when possible, because it's the road of least resistance. Thus, I believe you can't really share with companies (and expect them to share with you) unless you require it.
It's not often you can cut and paste working code from googling on the net, at best it'll use different variable names. You pretty much need to rewrite everything anyways, so not being able to paste from the net isn't much of a pain. (And anything larger than a few lines that you download is probably a library you'll want to have added to the main project.)
If they'd actually code it securely you could hack it but you wouldn't gain extra info because it wouldn't send anything that you weren't supposed to see. (Well, perhaps a packet or two before the person reached the door, but not completely on the other side of a wall or anything.)
Their lame security via code obscurity and anti-debugger tricks just mean that anyone who uses a debugger for legit reasons gets annoyed when the game won't play, but doesn't actually help security. Even their constant updates (Q3 as an example) doesn't stop the cheats, it only makes them come out with new versions every few weeks. Perhaps they're a bit harder to find for people who haven't mastered google searching, but for anyone willing to spend an hour looking, they're just as easy to find.
The code that verifies CD-keys has two parts. There's in in-game stuff which is pretty simple. Correct number of digits, totals to an even multiple of 31337, or something. Just enough to catch people who incorrectly enter the key. Then, the real check happens on their keyservers. You try to log into a game and your client sends your encrypted key to the keyserver, which decrypts the key, checks it, and tells the server to let you play, or to kick you out.
The keyserver doesn't use a fancy algorithm, it simply looks up the key in a list of every key they've shipped. No fancy math, nothing. But, almost impossible to hack because you need to guess a real key, not simply fool an algorithm you can watch with a debugger.
But it does work. If you pick a sensible goal and stick to it, you'll get something done. If you shoot for the moon you'll be fighting your employees every step of the way and you won't get as far as if you'd adopted a more modest system.
I've seen projects where things are reasonably documented. My current company is using doxygen for turning inline comments into a web interface describing the project. It's got a very low barrier to entry (use an extra / in comments you want to show up in the web interface) and the fact that you get results immediately means it's easier to get people to continue.
I don't have to be bitten by a dog to know that it hurts. Similarly, I've never tried keeping documentation in a seperate file from the code but I know it's not an optimal solution.
Hell, it's hard enough to get people to keep comments in sync with code when it's in the same file, let alone opening another whole program and finding the appropriate area to comment. The important issue is if it's easy enough to do that it's practial, and if there's anything else that would be more helpful for the ammount of work required.
Sure, if done religiously, it may prove helpful. That's a bit of a tautology though, like saying the key to winning a fight is not getting hit. There's truth in that and it does do things like suggest judo or aikido over karate, but it's not like you can simply decide to not get hit and it won't happen. Similarly, you can't simply say "I'm going to follow this properly". People don't skip comments because they're lazy, people skip comments because there's always deadline pressure, and because they want to write the comments once when done, not at every incremental revision. It usually comes down to doing things the "right" way, as taught in comp-sci, or actually getting the project finished before bankruptcy.
Better instead to pick something that is reasonable. Especially because I've seen projects that were commented properly - new programmers could come in and use the APIs without having to ask implementation details. This was accomplished by documenting the code, not by writing design documents. It was fairly easy and because of this, never got pushed off until later. You could always take ten minutes and update the comments on a day's work, so the programmer resistance to it was very low. Being that the simple answer works, why go looking to make it more complex?
A method that proposes comments that don't live in the code is broken. It requires programmers to have a second file open, and to update two things every time they make a change. A system that requires an extra annoying step for absolutely no gain is defective.
For absolutely no gain? Yes. There are better ways, such as putting any needed documentation into the source code itself. That way not only are they more accessible when reading the code but they're easier to change and harder to forget about.
Check out doxygen (at sourceforge) for a pretty cool system.
I mostly agree with your detractors, but I'll sumarize. I think that Hungarian notation could be useful while scanning new code, but new code (not just a routine I didn't write, but a whole new project I don't know the structure of) isn't something I encounter often at work.
However, I think that it's fairly easy to read new code. Just don't bother yourself with what particular kind of string is being referenced, it's enough to know it's a string and that they're matching it for a pattern.
If you're asking yourself what type a variable is all the time, it seems that you're doing something wrong. It should be obvious because of context, because of the project (use only standard string, etc), or because you can see a declaration either at the top of the function or in the appropriate header file.
I've wasted probably less than an hour total, ever, needing to look up the type of a variable. It's just by far the least frequent problem. I'd have wasted much more time typing lpwstdstrFoo instead of Foo.
I don't think the OEM thing has any importance. Microsoft is free to sell their product to anyone they want, and only under certain conditions, but once that product ends up in your hands either by purchase or salvage, it's yours.
And yeah, their new hardware-locked installations are a real pain. My company stayed with 2k to avoid the hassle and everyone else I know simply pirates it. Their sales figures which are so impressive are based on OEM sales, their retail sales have to be almost nonexistant. (I bet they count every machine the OEM sells, even if it was shipped with a blank HD, as an XP sale.)
I think the moderation was a joke, intentionally...
Just like me saying "I've done that and it worked, but until I scrubbed the platters with steel wool they felt a bit sticky."
I actually did do a media swap before to recover data, it was a 5.25 disk for my Apple2 that someone spilled orange juice on. Open it up, wash it off in warm water, sacrifice another blank disk to be the carrier, and copy the data to a working disk immediately. Worked fine, but I wouldn't want to try it with a HD.
If you find a Win2k CD in the trash, having been used by someone who purchased WinXP, you're fine to use it. You didn't copy it in an unauthorized way and as such, it's pretty much the safe as if you bought it. Despite what people say, purchasing software from a store doesn't involve any license at all. Just you and copyright law. Upgrades may be a special area, but even if they are, it's the purchaser who didn't live up to their obligations and their copy of WinXP shouldn't be used, your copy from the trash is perfectly legit.
If you find a HD with Win2k on it, you're in a bit of a grayer area. That copy is legitimate, in that it was required to use the copyrighted work and thus was exempt from the general no-copies rule. Once a new copy was made on a new HD, the old copy (that you have) is no longer required, and thus making it wouldn't be allowed. But, you're not making the copy, only using it... I think the violation comes when the original purchaser made the second copy without destroying the first. Technically, this made a copy (one or the other) which was not required to properly use the software, and thus was a violation at this point.
I believe the law essentially says though that it's the making of the unauthorized copies (and selling, etc) that's the problem. Posession, without proof of intent to sell or anything, should be legal. Thus, using Win2k, however you got it, should be fine.
I'm not a lawyer and even if I was, this is very much up in the air thanks to megabucks corps paying for changes to the law and judges (Kaplan) to interpret it in their favor.
How can there be a conspiracy to get you to upgrade when the distros are released for free?
RPM dependency hell isn't as much a problem as using the wrong tool for the job. If you tried to install IE6, one file at a time, you'd find it to be a problem too.
There are two solutions, if you don't use either it's probably because you go to admin-targetted websites like FreshRPMs.net and use server-oriented distros, instead of using a user-oriented distro like Xandros, Mandrake, or even Redhat with Ximian(?)...
The solution is to bundle everything you may need, similar to the way Windows programs would come with VBRUNx00.DLL because not everyone had them, despite making the download bigger for the 90% of people who already had them. A reasonable solution for end-user targetted packages, and one that many packages like Mozilla do. You need the correct system libraries (libc, etc) but that's pretty much the equivalent of making sure you're using the right version on Windows.
The other solution, and probably the best one in the end, is to use tools like apt-get, emerge, the bsd-ports system, or urpmi, where you simply select the package you want and all the dependencies are resolved, downloaded, and where needed, built. This hasn't been as widely adopted as it should, but it's starting to be. And really, Grandma shouldn't be picking a Linux distro at random, like she shouldn't be running Win2k Advanced Server. If she has a user-oriented distro, she's probably never experienced RPM dependency hell, even if she tries to install something. (And some OSes like Xandros make this very easy, even offering handy lists of what you might want to install, like having shareware.com built into the OS.)
Just FYI. Sounds like you might be looking for a better solution and they do exist.
Everyone who haughtily looks down on it because it's not a real unix, or because it's new, is an idiot.
If they evaluate it and it doesn't meet their needs, or doesn't run on their hardware as well as the default OS, or they just don't like the default system font, they're free to choose not to run it. To dismiss it out of hand though, because of the association of RMS, or Communism, or "free-can't-be'good" is a stupid move which, pretty much by definition would make them an idiot.
I actually lost a job opportunity once because I put Linux experience with my other Unix experience. They thought I was unskilled because I used a "toy" system. Linux was the most popular webserver OS, FreeBSD ran CDRom.com, etc. "Toy" OSes were taking over the world and they sat in their ivory tower and declared them and everyone who used them to be playing around. They were idiots. (For the record, they were doing data processing, of the results from their big-iron number crunching. Linux could have done it fine but I'd really have stuck with IRIX which they already had instead of having them buy a PC and so on, or try to install Linux on their existing hardware. I lost the job in the initial phone interview, long before they asked for any advice.)
But, there are valid reasons for not picking Linux. Most people don't use them though.