aside from bash...
on
Vim 7 Released
·
· Score: 2, Interesting
vim is the program that I use most. I used to code mostly shell stuff as a sysadmin and RAD niche programs: bash, perl, python, ruby... I always use vim for that. I also use vim for html, css, javascript, latex, and plain-ol' writing any kind of text.
These days I've been using eclipse a lot in emacs mode because my current job requires lots of java coding of a program with a huge code base, and my wrists are starting to hurt. Seriously, emacs is also good (IMHO not as fast as vim if you know both really well -- I use emacs mode in bash and know it pretty durn well too), but the weird stretches that you do in emacs will give you serious hand and wrist problems from awkward repetitive hand movements.
The other reason I prefer vim is that I REALLY LIKE having an editing mode separate from an insert mode. I like using vim to navigate through my code without worrying that I'll accidentally delete a character or make a typo. I actually do that all the time when I use emacs mode in eclipse, and then I have to undo my changes (but sometimes it takes a second to figure out which was the last intentional change).
With vim I can navigate through my file with little worry that I'll mess anything up.
And vim is FAR superior to old vi!!! Syntax-highlighting, word completion (C-n, C-p, dictionary completion, file completion...)
Emacs may be somewhat more powerful/extensible, and I believe it's quite good for coding in C/C++ and integrating with build tools -- a wonderful editor and environment. But for shear editing speed I think NOTHING can beat vim.
Re:GNOME vs KDE (not flamebait!)
on
Gnome 2.14 Review
·
· Score: 2, Insightful
I actually used to have a more polarised view. Before QT was fully gpl'ed, I didn't want to go anywhere near it and was 100% pro-gnome. Now I still like gtk and use gnome sometimes (I actually use wmii and do most of my work in aterms with an extra workspace for firefox), but since qt/kde has become more free then I've come to appreciate it too. About once or twice a year I try out kde for a few days on my debian systems. Last year I quickly got frustrated (I think konqueror is sort-of ugly as a file manager, and I like gnome's 2-panel approach rather than emulating the single MS "start-button" approach panel).
However, when I tried QT in debian unstable about a month ago, I thought the newer themes looked better, and qt apps were generally more usable. I'm glad to see them making progress. I used it for a few days, but ultimately switched back to gnome because I still prefer gtk apps (gaim, occasional gimp, evince, and I used to use gnumeric sometimes but now since 002.0 I use oocalc just because (since 2.0) overall OO is more robust than gnome-office or koffice)
Has qt been fully gpl'ed on win32 yet? This is something that I'm not clear on. I heard that it was (or is it just the new upcoming qt version that will be?) but I haven't seen any major qt apps ported to win32 yet.
I am forced to use win32 at work, and I like to have gaim available. It's just nice to be able to use the same apps (firefox, etc) on any platform. It makes it easier to get up to speed and be productive when you have to use a different platform for some reason. It's so nice to be able to use bash+cygwin, gnu make, gcc, and specialized tools like gnumeric's ssconvert on win32. So for me it's a big deal that gtk is available everywhere and qt is still *nix specific. That's probably the main reason that I stick to gtk.
I also wrote a pygtk app at my old job (small company) that many people at the office use. It worked on the couple dual-boot ubuntu boxes that I set up, and it works ok on winxp. At that time I don't think I could get FOSS pyqt on win32 (could be wrong, but I don't think so).
Having said all that, I love to see both kde/qt and gnome/gtk making progress.
I too am a self-taught coder. When you are constantly studying and learning as you go along, it's important to keep code and notes. I have a direcory called/home/myuser/notes, and I wrote a bash script that creates a file with a datestamp and opens it in vim when I type./notes (or it opens an existing one if the file for the current day's date exists). I now have > 2 years of notes saved on all kinds of stuff: my common mysqldump syntax usage, cpio usage, saslpasswd2 usage, notes on pam configuration...
In ~/notes I keep subdirs:/ruby,/perl,/c,/java,/python,/cs,... and in these I wrote a bunch of hello-world type programs that have regex usage, object creation, interface syntax, various iteration constructs... in all these languages.
I keep my/notes directory on my main home webserver in a SVN repository, and I can easily and quickly check it out from my laptop or from anywhere else. Other version control systems (monotone, darcs, cvs) are also good but svn has lots of support these days.
Then I keep larger programs in ~/code/ruby, ~/code/perl... That's where my classes and "real" code live. When I write some useful utility program that I use from the command line (lots of random scripts mostly), I keep them in ~/bin
So I keep ~/bin, ~/code, and ~/notes in svn. Occasionally I'll svn export them to a clean directory and make a tar.bz2 archive to burn to cd for backup (if I have important project code I'll back up the svn archive to get version history).
Having stuff like this will help you out a LOT! If you have to reinvent a wheel that someone else already invented because you don't know about it, that's bad in a way but at least you learn. If you never take notes or keep code, you'll find that you'll constantly be reinventing a wheel that YOU already invented -- whereas a vague recollection and some quick grepping will often result in little gems of code that you wrote a while back and would like to see for reference.
OK, maybe this is not exactly on topic, but why is it that Debian gets no respect on/. If there's an article on/. like "Ubuntu Dapper beta-1.0.652 now available for testing" or "Some guy in Australia evaluates ubuntu on his business desktop", it will get put into MAIN. If there is an article about the new graphic debian installer or Intel now supporting Debian with open source drivers, it only goes into the linux section.
I am now working as a java developer. In the past I've done development , DBA work, and sysadmin work, and whenever I want to get a server working well (or work on some development), my first option is ALWAYS Debian. Sure, right now I'm stuck working on solaris because I have to work with some commercial, binary-apps that do not run on any linux distros. Maybe I will sometimes consider using centos or something like that if I need to work with software that is only supported for RHEL.
But if I have an option, I use Debian. I have a small group of FOSS-enthusiast developer friends. We sort-of met each other arbitrarily, and we subsequently found that we all use Debian. When my cousin in Poland (I'm in the USA) was trying out different *nix distros, he eventually also settled on Debian (before he found out that I use it too).
Commercial distros appeal to newbies with their graphical installers. Ubuntu appeals to newbies who are charmed by the latest gnome apps. Many developers and sysadmins, on the other hand, who have used various distros have discovered that Debian is incredible for getting work done. It has more packages than any distro (I mean, EVERY FOSS tool is in Debian and it just keeps getting better and better -- now parrot, several FOSS java vm's, mono, Postgresql 8.1, mysql 5... are in unstable). When you want to compile from source, building (and customizing) your own deb from a source deb is VERY easy. And when you don't have time to worry about building from source and want to get some software up and running in a hurry, apt and dpkg are AMAZING.
Show me another distro right now that will let me install postgresql 8.1, parrot, mysql 5, mono, and several different FOSS java tools from binary packages in less than 5 minutes. It pisses me off when I read/. posts that say "Hey, if you want to run kernel 1.4 in 2005, try Debian." Hello, try debian unstable. Postgresql 8.1 was in debian unstable 1 day after it was released recently. If you're doing development, a bit of instability is fine (and IMHO, unstable is often more stable than ubuntu's official releases). If you want to run a production server, debian stable is hard to beat.
OK, gentoo is another great community distro and warrants respect for the hard work and quality that the community puts into it. But for me and many other people who truly love to use FOSS tools on every continent in the world in dozens of countries, whenever we consider trying out another distro for a while, we find that we're better off just using Debian.
Just use Debian (or Kubuntu if you insist on getting the bleeding edge version of KDE) or really ANY distro.
Seriously, your distro doesn't have to be KDE-Centric. Most large distros provide KDE, GNOME, and a dozen lightweight wm's. KDE will never disappear from Debian because there are hundreds of debian developers who use it. Just `apt-get install kde`.
I don't use it myself. I mostly use wmi, and I have gnome on the computer that I try to make user-friendly for my wife. KDE is a good DE and so is gnome. I prefer gnome, but it's basically just because of the look and feel, and I also like writing small apps in gtk in a few languages (although I did write a pyqt app once that I still use on my sharp zaurus and qt is totally fine).
I sort-of standardized my gui usage on gtk and gnome a couple years ago when QT was not fully GPL on every platform. If Trolltech had been fully open-source with qt, gnome probably would have never happened. As it is, they're both great and both very open now. I hope both projects continue to develop and learn from each other.
I doubt any decisions on the part of novell to standardize on gnome will be a serious setback for kde. Others will continue to use KDE, and it'll continue to develop.
Really, Gentoo is NOT the distro to be running for a webserver. What if there's a security hole in a basic library on which multiple apps depend (probably not the case here, but...). Do you recompile the whole system on your production server while everything is down?
I don't think Apache on GNU/Linux is really less secure than on OpenBSD, but use something STABLE like debian stable or centos stable. Maybe they're running gentoo stable, but still, I don't think it receives the same amount of security testing as some other distros. But I'll admit that I've only run Debian and RH in production, so maybe I'm talking out of my ass here.
They probably just did a horrible job of security with their apache or the dynamic pages that they're running on it.
Hey, I have my hot coffee mug sitting on an "1175 Hours" AOL CD right now on top of my "Programming Ruby" book on my desk. My mom always used to yell at me for putting drink cups directly on the table. With AOL CD's I protect my furniture!
(And by not installing AOL on Windows I protect my OS from bug-ridden slow death, although I don't really use windows anyway, but...).
It's basically all about the emoticons. My gaim supported all the yahooim emoticons, but then yahoo updated, and then the new emoticons weren't supported. I kept telling people at work to switch from yahoo im to gaim on windows (I was using and will continue to use gaim on GNU/linux), but they wouldn't do it because it didn't support the new emoticons.
Other people complained that gaim wasn't intuitive. Some complained that trillian looks better (I personally think that trillian resembles win32 crap nagware and think the ui is ugly, but to each her own).
Anyway, it's hard to make end users happy when you're a developer. People who can get free-as-in-beer im clients for win32 from yahoo and aol will only switch to FOSS IM clients if the latter absolutely blow the freeware versions out of the water.
It's like firefox and IE. IE already comes with windows. People aren't paying for it anyway. Sure, you can convince some people that IE will give them spyware, but for most people they don't care that firefox is better -- IE works ok. However, when the barista at my local coffee shop asked me if, as an IT guy, I could hook her up with a pirated version of MS Office and I told her to use OO 2.0 RC2 instead because it's quite good and free in every way, she got really excited and has been using it ever since.... same goes with GNU/linux and Windoze. You HAVE to buy the latter when you buy a new computer (yeah, I too build my systems from parts, but only geeks will ever do this). When hardware companies start selling pre-loaded GNU/Linux systems for less than the cost of windows, people will switch (at least businesses will first and then home users will get the same thing that they use at work and that their kids use at school).
Python is indeed a fun language. Ruby and Perl, IMO, are also fun in their own ways (ruby because it's the highest-level everything-is-an-object, everything-is-extensible language commonly used and Perl because it's madness yet incredibly efficient and effective for all its madness).
I don't know why some people are so down on Java though. I think Java is fun if you use open source implementations. For me, knowing that I'm using a community-made tool makes programming more fun. Writing Java to run on sun's jvm for some big company or writing in Windows.Forms C# for.NET to me seems like self-alienating corporate drudgery. However, I enjoy hacking with gcj/gij or writing java to run on ikvm or compiling to machine code w/gcj and playing with gtk# + mono to make little gnome apps.
I think the environment surrounding the development process and the presence or absence of a community around it affect the "funness level" of the development process more than the tools you actually used. (But still I don't like coding in some languages -- namely php... yuck, what an ugly, dull, boring mess).
Not everything in python is an object. Some builtin types like int are not objects. In ruby everything is an object and you can inherit from builtin types and override methods in the base class.
Yea, that is pretty strange. I would use Java before I'd use.NET. It's much more cross-platform then.NET is or ever will be even with Mono.
It depends what your goals are. I've been playing around with open source implementations of Java (gcj/gij mostly) and C# (mono), and the development branch of mono seems to be more stable for running gtk code (via gtk#) than javagtk (I've had some library-version compile conflicts with javagtk + gcj compiling to machine code (bytecode seems to be more reliable at least for now)). Sure, you can get java from sun and run swing apps on any platform, but what if you don't want to use swing? Gtk, IMO, is a better looking toolkit, and it has a ton more community support. It can be themed to simulate the look&feel of WinXP on win32, or have the gnome clearlooks theme to blend in nicely with gnome on *nix. Over time the use of gtk will dwarf the use of swing.
I think both the mono and the classpath communities are doing an awesome job bringing development tools to the open source community, but right now I think that mono is more robust than the open source implementations of java. And honestly, using open source tools ultimately makes cross platform work easier. Is Sun's java available for ARM or MIPS (I honestly don't know, but I'm assuming it's not)? Well, with mono or gcj/gij, when you have the source and can do anything you want with it, you can hack it to get it to compile on any platform, and if you don't have the time then other people in the community will. Over time you don't have to worry as much about support for new versions or libraries getting dropped for certain platforms when you have the source.
So basically I think open source tools without annoying licensing restrictions are, in the long term, always more cross-platform than closed-source tools, and right now I think the mono tools are just as cross-platform (and a bit more reliable) than open source java implementations.
OK, you can try to do that, but then you can't ride the subway to work. At best they just won't let you get on. At worst you'll get arrested for some BS charge (mob action, disorderly conduct, etc etc). Have you actually ever tried to argue with a cop about the law when you're in the right and they don't want to hear it?
They don't care. In 2001 NYC agreed to pay settlements of $50 million for cases of prisoner strip-search abuses
The city pays millions every year for settlements in police brutality cases. It's part of the NYPD budget. Cops do not care if you tell them "Hey, you can't do that to me. Read the Constitution." They'll kick your ass. Seriously. I once saw the cops beat the crap out of some guy at my subway station on the platform as hundreds of witnesses got off the train in rush hour. The guy was sprawled out on his stomach as 2 cops held him down with with their feet and another cop choked him while his face turned red. The guy was not moving at all, and they continued to choke him and kick him for probably a full minute as I (and at least a hundred other commuters) walked past to exit the subway station.
If you want to stop police abuses, people have to organize en masse to demand changes. Telling the cops that they're violating the law will only get you in jail or in the hospital.
120 million riders x 3 accidents per million = 360 accidents per month NOT 40 (duh).
So there you have it. 360 subway injuries per month from accidents. How many people were killed in the London subway bombings?
Even if you disregard any issues of morality or preserving freedom and liberties, simply on a cost-benefit analysis this is totally goddamn stupid -- it's just a BS pork barrel project, albeit one that screws people's lives. How many lives will this 200 million+ dollars (plus cop overtime pay) save? What is the benefit of the cost of the wars in Iraq and Afghanistan?
You'll see that the subway system sees about 120 million riders per month with 3 customer accidents and injuries per million per month. That's 40 injuries per month from accidents. Sometimes these are things like fatalities caused from someone getting bumped off of an over-crowded subway platform during rush hour onto the tracks...
So the city spends $2 million per week to "fight terrorism on the subway" and $212 million for security cameras on the subway rather than actually making a difference a difference by improving the system. Go to some G-train subway stations in brooklyn. The structural steel girders are rusting out and the stations are in dire need of maintenance.
And how much money has our government spent starting wars in the middle east (first gulf war, troops in Saudi Arabia, current Iraq occupation)... hundreds of billions of dollars
And then people over there get pissed off and want to set off subway bombs, and then we pay for it again by dealing with an army of cops checking our bags on the subway.
If they want to make subway riders safer, spend money on safety and infrastructure -- not cops -- to reduce accidents. If the government wants to eradicate terrorism, stop spending money on killing people in the Middle East. But of course getting rid of terrorism isn't the issue -- the issue is control of the dwindling global reserves of oil and new business opportunities in the middle east for American companies.
And we as taxpayers have to pay for it, and I have to let cops search my bag if I want to ride the subway to work and pay for that too.
I have like 5 of them in my closet. I'll sell you mine. Also, if you walk up and down any street where I live (NY City) on garbage day, you can pick up a couple extra ones for free. They work about 50% of the time.
I guess my home web server doesn't count in the statistics for server sales in the past year. I found the machine on the street, threw in a new hard drive from newegg, and it's running Debian GNU/Linux.
I just did some work for a company that wanted me to get rid of their spam and also upgrade their website backend. The website backend project involved setting up mysql on an old server that they had laying around, and the spam filter is postfix+spamassassin+amavis+clamav running on Debian on an extra desktop workstation computer that they had at the office. I also set up a near-identical spam filtering machine using Debian on a workstation at a law firm recently.
So, there you go, 4 new servers running Debian GNU/linux in the past year.
It's interesting that these sales figures are dollar denominated. It marginalizes GNU/Linux as a server OS (because it's free as in beer or alternately relatively cheap if it comes with support) while also informing how valuable it is (because in the hands of the skilled, it gets the job done while costing nothing).
... you want software that works
on
McVoy Strikes Back
·
· Score: 2, Insightful
"I want software that works, not software I have to hire consultants for to make it work."
You must not be a programmer or sysadmin. This is the attitude of the former COO of the company for which I used to work. A few years back a former Director of IT started building the company's web site and business on apache and php (albeit with a closed-source sybase db driving the backend). The COO thought that the company was paying too much for consultants and decided to hire someone to reimplement everything with MS SQL and IIS.
He thought that because it was easy for him to use MS (tm)(c) Windows(tm)(c) on his home computer, it would make development easier and cheaper to get rid of linux. He thought that the company would effectively be able to get the software developed and then get rid of the IT staff and then things would just continue to run with no need for maintenance b/c the company would be running "good software" for which they wouldn't have to pay someone to administer.
Well, after paying developers $50,000 to design and build part of the redesigned corporate web backend, buying a new MS Exchange and paying some totally ignorant windows admins about $10,000 to migrate to it from the old exchange 5.5 (really, it was scandalous, they couldn't accomplish this after weeks, it should take no more than 1 week -- TOPS), and buying new hardware for the new systems, the projects eventually got abandoned. They continued to go overbudget. The consultants working on them couldn't finish the job. The company spent probably $100,000 in development, software and hardware costs (and they continued to pay the COO to "manage" it all).
Then he got laid off, all the incompetent windows admins got laid off, and they hired me. I continued to develop and maintain all the linux stuff and add more open source solutions. The company spent zero for support and software costs (I ran everything on Debian. All software was free as in speech and beer). THey just had to pay the salary of one guy to manage the open source website, database, and do continued development in free languages like php, perl, python, ruby...
The argument that companies should "buy" "software that works" instead of get free software and pay someone to implement and support it is 100% BS. Companies that depend on their computer systems to work WILL ALWAYS need to pay SOMEONE to support their systems SOMEHOW (whether they hire on a full-timer, pay consultants, or enter a service agreement).
I heard some companies are paying $5,000/license for multiple BK licenses. This strikes me as being a tremendous waste of resources. Hire ONE consultant to work 5 hours a week to support everyone who needs to use the source management tools and go with a free solution like subversion, darcs, monotone, or, now, git.
I bet in 5 years BK will cease to exist because the free open source solutions will be just as good or better. The international community of open source programmers will outpace BK's innovation and develop a better solution.
So you say that the workers' jobs are to provide profits for the owners of the company. Why do you think the "profits" of the company's shareholders are more important than the livelihoods of 13,000 people who work at the company?
If you trace your argument to its roots, it boils down to the idea that people in society should work for the profits of the small percentage of people who "own" companies. I agree with you that companies in our society exist to make money and nothing else, but maybe people who have worked hard to make the company run feel some sort of entitlement afterward and think that their needs are more important than the future profitability of the company.
Maybe you're among the group of owners. But if you're not, then your point of view to me seems pathological. I think anyone who works at a company and then just passively accepts that they can be discarded inconsequentially is crazy. And it strikes me that our society seems pretty screwed up when programmers posting on/. espouse the view that there is something WRONG with their fellow programmers organizing to protect their own interests.
If I were in Europe now, I'd go to the picket line and show my support.
Perl has a lot going for it. mod_perl and html::mason (used by/. and apache) together are a much more maintainable solution to large web apps than PHP, IMO. I had to maintain some really ugly old PHP code at my old job. Any language can get ugly if the coding style is ugly, but properly written html::mason does a lot more to separate Model, View, and Controller than any PHP solution I've seen.
I did a lot of perl programming back in the day. Although I still like it a lot (I think the DBI is one of the best database libraries out there), I do find myself working more and more in both python and ruby for rapid development.
If you love perl but sometimes get a headache from staring at long lines of code with lots of sigils, braces, and parentheses, take a look at ruby. From what I've seen of Perl6 so far, it will also clean up some syntax while retaining neat perlish tricks and adding a lot of advanced programming language features.
I have high hopes for the parrot project too. Maybe one day we'll all be able to write parts of our apps in ruby, python, perl6, ponie, and integrate them together on a common VM platform. I think this would be an incredible step for open source software development! Still, I'm not going to hold my breath waiting for this... it looks like this scripting nirvana is still pretty far off.
It's interesting to hear that so many people see no future in IT in the USA because I've been fighting tooth and nail to get into the industry in the past 2 years. I graduated college with a degree in US History a few years ago, saw that many of my peers were ending up in dead-end paper-sorting office jobs, and decided to teach myself some programming languages. I learned perl, got a contract to build a perl CMS, got another short gig, and then I landed a job as the Director of IT at a small, failing company (just quit that job and I'll be looking for other work).
If you think there's no future in IT in the USA, look at all the people graduating from liberal arts colleges with degrees in psychology, philosophy, anthropology, history, english...
I agree 100% that IT isn't worth it. In order to be competitive, I've spent ALL my free time in the past 2 years studying maybe 7 or 8 different computer languages and practicing *nix sysadmin and DBA stuff on old throwaway hardware. Now I'm finally starting to work on a couple free software projects, and I like it -- otherwise I'd be out.
And while I think it's definitely not worth it economically because for the same effort I could get a law degree, earn more money, and have more job stability, aside from getting a law degree going the MBA path, where are the opportunities in the USA? I have an engineer friend who has been laid off maybe 4 or 5 times in 3 years working as a process engineer in auto plants (finally landed a job for a Japanese auto maker in the USA and now he has SOME job stability).
The USA trade and budget deficits continue to grow. The 3rd world is developing infrastructure and human resources, and increasing productivity in the USA economy mean that companies can generate the same or more revenue with fewer workers.
Sure, IBM, Microsoft and other companies making these pitches about a shortage of workers in IT are just trying to get cheaper labor by increasing labor supply. However, if IT is a dead-end path as I'm hearing from a lot of you, WHERE is there opportunity in the mid-term future of the American market economy?
I use shorewall on my LEAF/Bering router on an old Pentium 1. It's been routing and protecting my home cable network and a couple internal servers for over a year now (current uptime is probably 5 months or so).
I also set it up on an x86 machine on Debian at my old job when their POS proprietary firewall/router fried itself.
I've told a few people who I've worked with that I think that Shorewall is the BEST DOCUMENTED open source app I've ever used. I learned much of what I know about proxy arping, arp caches, how DMZ's actually work, CIDR, and lots of other stuff like that from the Shorewall documentation. Even if you don't intend to USE Shorewall, if you want to learn more about networking, take a look at the Shorewall docs. It's probably the best concise explanation of many network concepts that I've come across (including text books, other online docs...)
So, Thanks Tom Eastep. I've learned a LOT from your work, and you've made an incredible contribution to free open source software!!!
Slashdot Mirror
vim is the program that I use most. I used to code mostly shell stuff as a sysadmin and RAD niche programs: bash, perl, python, ruby... I always use vim for that. I also use vim for html, css, javascript, latex, and plain-ol' writing any kind of text.
These days I've been using eclipse a lot in emacs mode because my current job requires lots of java coding of a program with a huge code base, and my wrists are starting to hurt. Seriously, emacs is also good (IMHO not as fast as vim if you know both really well -- I use emacs mode in bash and know it pretty durn well too), but the weird stretches that you do in emacs will give you serious hand and wrist problems from awkward repetitive hand movements.
The other reason I prefer vim is that I REALLY LIKE having an editing mode separate from an insert mode. I like using vim to navigate through my code without worrying that I'll accidentally delete a character or make a typo. I actually do that all the time when I use emacs mode in eclipse, and then I have to undo my changes (but sometimes it takes a second to figure out which was the last intentional change).
With vim I can navigate through my file with little worry that I'll mess anything up.
And vim is FAR superior to old vi!!! Syntax-highlighting, word completion (C-n, C-p, dictionary completion, file completion...)
Emacs may be somewhat more powerful/extensible, and I believe it's quite good for coding in C/C++ and integrating with build tools -- a wonderful editor and environment. But for shear editing speed I think NOTHING can beat vim.
I actually used to have a more polarised view. Before QT was fully gpl'ed, I didn't want to go anywhere near it and was 100% pro-gnome. Now I still like gtk and use gnome sometimes (I actually use wmii and do most of my work in aterms with an extra workspace for firefox), but since qt/kde has become more free then I've come to appreciate it too. About once or twice a year I try out kde for a few days on my debian systems. Last year I quickly got frustrated (I think konqueror is sort-of ugly as a file manager, and I like gnome's 2-panel approach rather than emulating the single MS "start-button" approach panel).
However, when I tried QT in debian unstable about a month ago, I thought the newer themes looked better, and qt apps were generally more usable. I'm glad to see them making progress. I used it for a few days, but ultimately switched back to gnome because I still prefer gtk apps (gaim, occasional gimp, evince, and I used to use gnumeric sometimes but now since 002.0 I use oocalc just because (since 2.0) overall OO is more robust than gnome-office or koffice)
Has qt been fully gpl'ed on win32 yet? This is something that I'm not clear on. I heard that it was (or is it just the new upcoming qt version that will be?) but I haven't seen any major qt apps ported to win32 yet.
I am forced to use win32 at work, and I like to have gaim available. It's just nice to be able to use the same apps (firefox, etc) on any platform. It makes it easier to get up to speed and be productive when you have to use a different platform for some reason. It's so nice to be able to use bash+cygwin, gnu make, gcc, and specialized tools like gnumeric's ssconvert on win32. So for me it's a big deal that gtk is available everywhere and qt is still *nix specific. That's probably the main reason that I stick to gtk.
I also wrote a pygtk app at my old job (small company) that many people at the office use. It worked on the couple dual-boot ubuntu boxes that I set up, and it works ok on winxp. At that time I don't think I could get FOSS pyqt on win32 (could be wrong, but I don't think so).
Having said all that, I love to see both kde/qt and gnome/gtk making progress.
I too am a self-taught coder. When you are constantly studying and learning as you go along, it's important to keep code and notes. I have a direcory called /home/myuser/notes, and I wrote a bash script that creates a file with a datestamp and opens it in vim when I type ./notes (or it opens an existing one if the file for the current day's date exists). I now have > 2 years of notes saved on all kinds of stuff: my common mysqldump syntax usage, cpio usage, saslpasswd2 usage, notes on pam configuration...
/ruby, /perl, /c, /java, /python, /cs, ... and in these I wrote a bunch of hello-world type programs that have regex usage, object creation, interface syntax, various iteration constructs... in all these languages.
/notes directory on my main home webserver in a SVN repository, and I can easily and quickly check it out from my laptop or from anywhere else. Other version control systems (monotone, darcs, cvs) are also good but svn has lots of support these days.
In ~/notes I keep subdirs:
I keep my
Then I keep larger programs in ~/code/ruby, ~/code/perl... That's where my classes and "real" code live. When I write some useful utility program that I use from the command line (lots of random scripts mostly), I keep them in ~/bin
So I keep ~/bin, ~/code, and ~/notes in svn. Occasionally I'll svn export them to a clean directory and make a tar.bz2 archive to burn to cd for backup (if I have important project code I'll back up the svn archive to get version history).
Having stuff like this will help you out a LOT! If you have to reinvent a wheel that someone else already invented because you don't know about it, that's bad in a way but at least you learn. If you never take notes or keep code, you'll find that you'll constantly be reinventing a wheel that YOU already invented -- whereas a vague recollection and some quick grepping will often result in little gems of code that you wrote a while back and would like to see for reference.
OK, maybe this is not exactly on topic, but why is it that Debian gets no respect on /. If there's an article on /. like "Ubuntu Dapper beta-1.0.652 now available for testing" or "Some guy in Australia evaluates ubuntu on his business desktop", it will get put into MAIN. If there is an article about the new graphic debian installer or Intel now supporting Debian with open source drivers, it only goes into the linux section.
/. posts that say "Hey, if you want to run kernel 1.4 in 2005, try Debian." Hello, try debian unstable. Postgresql 8.1 was in debian unstable 1 day after it was released recently. If you're doing development, a bit of instability is fine (and IMHO, unstable is often more stable than ubuntu's official releases). If you want to run a production server, debian stable is hard to beat.
/.!!!
I am now working as a java developer. In the past I've done development , DBA work, and sysadmin work, and whenever I want to get a server working well (or work on some development), my first option is ALWAYS Debian. Sure, right now I'm stuck working on solaris because I have to work with some commercial, binary-apps that do not run on any linux distros. Maybe I will sometimes consider using centos or something like that if I need to work with software that is only supported for RHEL.
But if I have an option, I use Debian. I have a small group of FOSS-enthusiast developer friends. We sort-of met each other arbitrarily, and we subsequently found that we all use Debian. When my cousin in Poland (I'm in the USA) was trying out different *nix distros, he eventually also settled on Debian (before he found out that I use it too).
Commercial distros appeal to newbies with their graphical installers. Ubuntu appeals to newbies who are charmed by the latest gnome apps. Many developers and sysadmins, on the other hand, who have used various distros have discovered that Debian is incredible for getting work done. It has more packages than any distro (I mean, EVERY FOSS tool is in Debian and it just keeps getting better and better -- now parrot, several FOSS java vm's, mono, Postgresql 8.1, mysql 5... are in unstable). When you want to compile from source, building (and customizing) your own deb from a source deb is VERY easy. And when you don't have time to worry about building from source and want to get some software up and running in a hurry, apt and dpkg are AMAZING.
Show me another distro right now that will let me install postgresql 8.1, parrot, mysql 5, mono, and several different FOSS java tools from binary packages in less than 5 minutes. It pisses me off when I read
OK, gentoo is another great community distro and warrants respect for the hard work and quality that the community puts into it. But for me and many other people who truly love to use FOSS tools on every continent in the world in dozens of countries, whenever we consider trying out another distro for a while, we find that we're better off just using Debian.
Debian deserves more respect on
What is the best KDE-Centric distro?
Just use Debian (or Kubuntu if you insist on getting the bleeding edge version of KDE) or really ANY distro.
Seriously, your distro doesn't have to be KDE-Centric. Most large distros provide KDE, GNOME, and a dozen lightweight wm's. KDE will never disappear from Debian because there are hundreds of debian developers who use it. Just `apt-get install kde`.
I don't use it myself. I mostly use wmi, and I have gnome on the computer that I try to make user-friendly for my wife. KDE is a good DE and so is gnome. I prefer gnome, but it's basically just because of the look and feel, and I also like writing small apps in gtk in a few languages (although I did write a pyqt app once that I still use on my sharp zaurus and qt is totally fine).
I sort-of standardized my gui usage on gtk and gnome a couple years ago when QT was not fully GPL on every platform. If Trolltech had been fully open-source with qt, gnome probably would have never happened. As it is, they're both great and both very open now. I hope both projects continue to develop and learn from each other.
I doubt any decisions on the part of novell to standardize on gnome will be a serious setback for kde. Others will continue to use KDE, and it'll continue to develop.
Here's what they're running:
Date: Tue, 01 Nov 2005 20:01:34 GMT
Server: Apache/1.3.31 (Unix) (Gentoo/Linux)
Really, Gentoo is NOT the distro to be running for a webserver. What if there's a security hole in a basic library on which multiple apps depend (probably not the case here, but...). Do you recompile the whole system on your production server while everything is down?
I don't think Apache on GNU/Linux is really less secure than on OpenBSD, but use something STABLE like debian stable or centos stable. Maybe they're running gentoo stable, but still, I don't think it receives the same amount of security testing as some other distros. But I'll admit that I've only run Debian and RH in production, so maybe I'm talking out of my ass here.
They probably just did a horrible job of security with their apache or the dynamic pages that they're running on it.
Hey, I have my hot coffee mug sitting on an "1175 Hours" AOL CD right now on top of my "Programming Ruby" book on my desk. My mom always used to yell at me for putting drink cups directly on the table. With AOL CD's I protect my furniture!
(And by not installing AOL on Windows I protect my OS from bug-ridden slow death, although I don't really use windows anyway, but...).
It's basically all about the emoticons. My gaim supported all the yahooim emoticons, but then yahoo updated, and then the new emoticons weren't supported. I kept telling people at work to switch from yahoo im to gaim on windows (I was using and will continue to use gaim on GNU/linux), but they wouldn't do it because it didn't support the new emoticons.
... same goes with GNU/linux and Windoze. You HAVE to buy the latter when you buy a new computer (yeah, I too build my systems from parts, but only geeks will ever do this). When hardware companies start selling pre-loaded GNU/Linux systems for less than the cost of windows, people will switch (at least businesses will first and then home users will get the same thing that they use at work and that their kids use at school).
Other people complained that gaim wasn't intuitive. Some complained that trillian looks better (I personally think that trillian resembles win32 crap nagware and think the ui is ugly, but to each her own).
Anyway, it's hard to make end users happy when you're a developer. People who can get free-as-in-beer im clients for win32 from yahoo and aol will only switch to FOSS IM clients if the latter absolutely blow the freeware versions out of the water.
It's like firefox and IE. IE already comes with windows. People aren't paying for it anyway. Sure, you can convince some people that IE will give them spyware, but for most people they don't care that firefox is better -- IE works ok. However, when the barista at my local coffee shop asked me if, as an IT guy, I could hook her up with a pirated version of MS Office and I told her to use OO 2.0 RC2 instead because it's quite good and free in every way, she got really excited and has been using it ever since.
Python is indeed a fun language. Ruby and Perl, IMO, are also fun in their own ways (ruby because it's the highest-level everything-is-an-object, everything-is-extensible language commonly used and Perl because it's madness yet incredibly efficient and effective for all its madness).
.NET to me seems like self-alienating corporate drudgery. However, I enjoy hacking with gcj/gij or writing java to run on ikvm or compiling to machine code w/gcj and playing with gtk# + mono to make little gnome apps.
I don't know why some people are so down on Java though. I think Java is fun if you use open source implementations. For me, knowing that I'm using a community-made tool makes programming more fun. Writing Java to run on sun's jvm for some big company or writing in Windows.Forms C# for
I think the environment surrounding the development process and the presence or absence of a community around it affect the "funness level" of the development process more than the tools you actually used. (But still I don't like coding in some languages -- namely php... yuck, what an ugly, dull, boring mess).
Not everything in python is an object. Some builtin types like int are not objects. In ruby everything is an object and you can inherit from builtin types and override methods in the base class.
It depends what your goals are. I've been playing around with open source implementations of Java (gcj/gij mostly) and C# (mono), and the development branch of mono seems to be more stable for running gtk code (via gtk#) than javagtk (I've had some library-version compile conflicts with javagtk + gcj compiling to machine code (bytecode seems to be more reliable at least for now)). Sure, you can get java from sun and run swing apps on any platform, but what if you don't want to use swing? Gtk, IMO, is a better looking toolkit, and it has a ton more community support. It can be themed to simulate the look&feel of WinXP on win32, or have the gnome clearlooks theme to blend in nicely with gnome on *nix. Over time the use of gtk will dwarf the use of swing.
I think both the mono and the classpath communities are doing an awesome job bringing development tools to the open source community, but right now I think that mono is more robust than the open source implementations of java. And honestly, using open source tools ultimately makes cross platform work easier. Is Sun's java available for ARM or MIPS (I honestly don't know, but I'm assuming it's not)? Well, with mono or gcj/gij, when you have the source and can do anything you want with it, you can hack it to get it to compile on any platform, and if you don't have the time then other people in the community will. Over time you don't have to worry as much about support for new versions or libraries getting dropped for certain platforms when you have the source.
So basically I think open source tools without annoying licensing restrictions are, in the long term, always more cross-platform than closed-source tools, and right now I think the mono tools are just as cross-platform (and a bit more reliable) than open source java implementations.
OK, you can try to do that, but then you can't ride the subway to work. At best they just won't let you get on. At worst you'll get arrested for some BS charge (mob action, disorderly conduct, etc etc). Have you actually ever tried to argue with a cop about the law when you're in the right and they don't want to hear it?
0 61EFC3F540C738FDDA90994DB404482&incamp=archive:sea rch
They don't care. In 2001 NYC agreed to pay settlements of $50 million for cases of prisoner strip-search abuses
http://query.nytimes.com/gst/abstract.html?res=F4
The city pays millions every year for settlements in police brutality cases. It's part of the NYPD budget. Cops do not care if you tell them "Hey, you can't do that to me. Read the Constitution." They'll kick your ass. Seriously. I once saw the cops beat the crap out of some guy at my subway station on the platform as hundreds of witnesses got off the train in rush hour. The guy was sprawled out on his stomach as 2 cops held him down with with their feet and another cop choked him while his face turned red. The guy was not moving at all, and they continued to choke him and kick him for probably a full minute as I (and at least a hundred other commuters) walked past to exit the subway station.
If you want to stop police abuses, people have to organize en masse to demand changes. Telling the cops that they're violating the law will only get you in jail or in the hospital.
btw, I shouldn't be posting late at night...
120 million riders x 3 accidents per million = 360 accidents per month NOT 40 (duh).
So there you have it. 360 subway injuries per month from accidents. How many people were killed in the London subway bombings?
Even if you disregard any issues of morality or preserving freedom and liberties, simply on a cost-benefit analysis this is totally goddamn stupid -- it's just a BS pork barrel project, albeit one that screws people's lives. How many lives will this 200 million+ dollars (plus cop overtime pay) save? What is the benefit of the cost of the wars in Iraq and Afghanistan?
I live in NY City, and for the past few weeks there have been cops in many subway stations doing random searches. According to this article:
. htm
http://www.msnbc.msn.com/id/8660152/
This is costing the city $2 million per week.
If you look at this page (New York MTA):
http://www.mta.nyc.ny.us/mta/ind-perform/per-nyct
You'll see that the subway system sees about 120 million riders per month with 3 customer accidents and injuries per million per month. That's 40 injuries per month from accidents. Sometimes these are things like fatalities caused from someone getting bumped off of an over-crowded subway platform during rush hour onto the tracks...
So the city spends $2 million per week to "fight terrorism on the subway" and $212 million for security cameras on the subway rather than actually making a difference a difference by improving the system. Go to some G-train subway stations in brooklyn. The structural steel girders are rusting out and the stations are in dire need of maintenance.
And how much money has our government spent starting wars in the middle east (first gulf war, troops in Saudi Arabia, current Iraq occupation)... hundreds of billions of dollars
And then people over there get pissed off and want to set off subway bombs, and then we pay for it again by dealing with an army of cops checking our bags on the subway.
If they want to make subway riders safer, spend money on safety and infrastructure -- not cops -- to reduce accidents. If the government wants to eradicate terrorism, stop spending money on killing people in the Middle East. But of course getting rid of terrorism isn't the issue -- the issue is control of the dwindling global reserves of oil and new business opportunities in the middle east for American companies.
And we as taxpayers have to pay for it, and I have to let cops search my bag if I want to ride the subway to work and pay for that too.
I have like 5 of them in my closet. I'll sell you mine. Also, if you walk up and down any street where I live (NY City) on garbage day, you can pick up a couple extra ones for free. They work about 50% of the time.
I guess my home web server doesn't count in the statistics for server sales in the past year. I found the machine on the street, threw in a new hard drive from newegg, and it's running Debian GNU/Linux. I just did some work for a company that wanted me to get rid of their spam and also upgrade their website backend. The website backend project involved setting up mysql on an old server that they had laying around, and the spam filter is postfix+spamassassin+amavis+clamav running on Debian on an extra desktop workstation computer that they had at the office. I also set up a near-identical spam filtering machine using Debian on a workstation at a law firm recently. So, there you go, 4 new servers running Debian GNU/linux in the past year. It's interesting that these sales figures are dollar denominated. It marginalizes GNU/Linux as a server OS (because it's free as in beer or alternately relatively cheap if it comes with support) while also informing how valuable it is (because in the hands of the skilled, it gets the job done while costing nothing).
"I want software that works, not software I have to hire consultants for to make it work."
You must not be a programmer or sysadmin. This is the attitude of the former COO of the company for which I used to work. A few years back a former Director of IT started building the company's web site and business on apache and php (albeit with a closed-source sybase db driving the backend). The COO thought that the company was paying too much for consultants and decided to hire someone to reimplement everything with MS SQL and IIS.
He thought that because it was easy for him to use MS (tm)(c) Windows(tm)(c) on his home computer, it would make development easier and cheaper to get rid of linux. He thought that the company would effectively be able to get the software developed and then get rid of the IT staff and then things would just continue to run with no need for maintenance b/c the company would be running "good software" for which they wouldn't have to pay someone to administer.
Well, after paying developers $50,000 to design and build part of the redesigned corporate web backend, buying a new MS Exchange and paying some totally ignorant windows admins about $10,000 to migrate to it from the old exchange 5.5 (really, it was scandalous, they couldn't accomplish this after weeks, it should take no more than 1 week -- TOPS), and buying new hardware for the new systems, the projects eventually got abandoned. They continued to go overbudget. The consultants working on them couldn't finish the job. The company spent probably $100,000 in development, software and hardware costs (and they continued to pay the COO to "manage" it all).
Then he got laid off, all the incompetent windows admins got laid off, and they hired me. I continued to develop and maintain all the linux stuff and add more open source solutions. The company spent zero for support and software costs (I ran everything on Debian. All software was free as in speech and beer). THey just had to pay the salary of one guy to manage the open source website, database, and do continued development in free languages like php, perl, python, ruby...
The argument that companies should "buy" "software that works" instead of get free software and pay someone to implement and support it is 100% BS. Companies that depend on their computer systems to work WILL ALWAYS need to pay SOMEONE to support their systems SOMEHOW (whether they hire on a full-timer, pay consultants, or enter a service agreement).
I heard some companies are paying $5,000/license for multiple BK licenses. This strikes me as being a tremendous waste of resources. Hire ONE consultant to work 5 hours a week to support everyone who needs to use the source management tools and go with a free solution like subversion, darcs, monotone, or, now, git.
I bet in 5 years BK will cease to exist because the free open source solutions will be just as good or better. The international community of open source programmers will outpace BK's innovation and develop a better solution.
So you say that the workers' jobs are to provide profits for the owners of the company. Why do you think the "profits" of the company's shareholders are more important than the livelihoods of 13,000 people who work at the company?
/. espouse the view that there is something WRONG with their fellow programmers organizing to protect their own interests.
If you trace your argument to its roots, it boils down to the idea that people in society should work for the profits of the small percentage of people who "own" companies. I agree with you that companies in our society exist to make money and nothing else, but maybe people who have worked hard to make the company run feel some sort of entitlement afterward and think that their needs are more important than the future profitability of the company.
Maybe you're among the group of owners. But if you're not, then your point of view to me seems pathological. I think anyone who works at a company and then just passively accepts that they can be discarded inconsequentially is crazy. And it strikes me that our society seems pretty screwed up when programmers posting on
If I were in Europe now, I'd go to the picket line and show my support.
Perl has a lot going for it. mod_perl and html::mason (used by /. and apache) together are a much more maintainable solution to large web apps than PHP, IMO. I had to maintain some really ugly old PHP code at my old job. Any language can get ugly if the coding style is ugly, but properly written html::mason does a lot more to separate Model, View, and Controller than any PHP solution I've seen.
I did a lot of perl programming back in the day. Although I still like it a lot (I think the DBI is one of the best database libraries out there), I do find myself working more and more in both python and ruby for rapid development.
If you love perl but sometimes get a headache from staring at long lines of code with lots of sigils, braces, and parentheses, take a look at ruby. From what I've seen of Perl6 so far, it will also clean up some syntax while retaining neat perlish tricks and adding a lot of advanced programming language features.
I have high hopes for the parrot project too. Maybe one day we'll all be able to write parts of our apps in ruby, python, perl6, ponie, and integrate them together on a common VM platform. I think this would be an incredible step for open source software development! Still, I'm not going to hold my breath waiting for this... it looks like this scripting nirvana is still pretty far off.
It's interesting to hear that so many people see no future in IT in the USA because I've been fighting tooth and nail to get into the industry in the past 2 years. I graduated college with a degree in US History a few years ago, saw that many of my peers were ending up in dead-end paper-sorting office jobs, and decided to teach myself some programming languages. I learned perl, got a contract to build a perl CMS, got another short gig, and then I landed a job as the Director of IT at a small, failing company (just quit that job and I'll be looking for other work).
If you think there's no future in IT in the USA, look at all the people graduating from liberal arts colleges with degrees in psychology, philosophy, anthropology, history, english...
I agree 100% that IT isn't worth it. In order to be competitive, I've spent ALL my free time in the past 2 years studying maybe 7 or 8 different computer languages and practicing *nix sysadmin and DBA stuff on old throwaway hardware. Now I'm finally starting to work on a couple free software projects, and I like it -- otherwise I'd be out.
And while I think it's definitely not worth it economically because for the same effort I could get a law degree, earn more money, and have more job stability, aside from getting a law degree going the MBA path, where are the opportunities in the USA? I have an engineer friend who has been laid off maybe 4 or 5 times in 3 years working as a process engineer in auto plants (finally landed a job for a Japanese auto maker in the USA and now he has SOME job stability).
The USA trade and budget deficits continue to grow. The 3rd world is developing infrastructure and human resources, and increasing productivity in the USA economy mean that companies can generate the same or more revenue with fewer workers.
Sure, IBM, Microsoft and other companies making these pitches about a shortage of workers in IT are just trying to get cheaper labor by increasing labor supply. However, if IT is a dead-end path as I'm hearing from a lot of you, WHERE is there opportunity in the mid-term future of the American market economy?
I use shorewall on my LEAF/Bering router on an old Pentium 1. It's been routing and protecting my home cable network and a couple internal servers for over a year now (current uptime is probably 5 months or so). I also set it up on an x86 machine on Debian at my old job when their POS proprietary firewall/router fried itself. I've told a few people who I've worked with that I think that Shorewall is the BEST DOCUMENTED open source app I've ever used. I learned much of what I know about proxy arping, arp caches, how DMZ's actually work, CIDR, and lots of other stuff like that from the Shorewall documentation. Even if you don't intend to USE Shorewall, if you want to learn more about networking, take a look at the Shorewall docs. It's probably the best concise explanation of many network concepts that I've come across (including text books, other online docs...) So, Thanks Tom Eastep. I've learned a LOT from your work, and you've made an incredible contribution to free open source software!!!