• commandlinefu.com lets you post your command line commands of choice, and lets others vote on them. Good for if you’re looking for something specific, or if you want some idea of what other people use the command line for, and what they find most useful.
• Pixelbeat‘s Linux Commands – A practical reference. Also, try More Linux Commands – A practical reference for even more CLI goodness.
• For what looks like a pretty newb-friendly introduction to command line, try LinuxCommand.org. Includes a section on shell scripting, which is where a lot of the power of the command line comes to get even more pumped up.

A quick search at the SourceForge open source developer resource returns over 900 programs that match the search term “text editor”.

Consider that almost every time you need to enter text into a program or web form, you are using an editor, and you can imagine just how often text editing is implemented.

Yet, with all those editors to choose from, for most who want to edit text, the choice boils down to two: Vi (pronounced “vee eye”) and Emacs. Each are actually families of functionally equivalent editors, but the Linux default implementations are VIM (Vi, IMproved) and GNU Emacs.

Both are designed for text editing, with differences. Emacs is modeless, meaning you can enter text or commands at any time; Vi uses two basic modes, one for entering text, the other for executing commands. Being modeless, Emacs is more familiar to users accustomed to more common modeless word processors like Microsoft Word; Vi’s modes can be confusing at first, but permit faster touch typing because commands don’t involve numerous multi-keystroke commands that use Alt, Ctrl, ESC and other keys (which is what you use in Emacs).

Free Software Foundation founder Richard Stallman wrote the first version of Emacs in 1976; the current stable version is 23.2 (as of 2010). Written in and incorporating Lisp, Emacs is endlessly extensible and customizable.

As a smaller program, Vi loads much faster than Emacs and is built from the ground up as a programmer’s editor, so many programmers prefer Vi for utilitarian source code editing. Emacs, on the other hand, is designed to edit any kind of text file, and combined with its built-in Lisp implementation that can mean coding source, outlining, editing documents, to-do lists, contact files, calendars and scheduling, reading email and news, even websurfing. Emacs also comes with a selection of text-based games including Tetris, Snake, even a version of the old Eliza “psychotherapy” program.

Real old-school *NIX gurus know and use both Vi and Emacs: vi for fast, simple jobs and Emacs for more complicated tasks; but for others, Emacs works better as an all-around text editor.

You can bypass the issue with one of the many GUI text editors available like Gedit, the official text editor of the GNOME desktop environment (gedit.org), Kate, the KDE Advanced Text Editor (kate.kde.org), or Java-based programmer’s editor jEdit (jedit.org).

It’s just something like three and a half years since this was written, but these things just keep staying as important as ever. I’ll add a few more editorial comments in-line here, but slowly AND surely open source is taking over the key systems we use: things like telephones and web browsers and web servers and all the software that makes the web/Internet run are all increasingly being designed with open source software at their cores.

Not long ago, while leading my lavish CPU columnist lifestyle (up at noon, surfing the ‘net for a few hours before getting ready to party, etc etc), I read “Things I can do in Linux that I can’t do on Windows” on Dan Martin’s blog, as well as Prakash Advani’s list of “101 reasons why Linux is better than Windows”, even if he lists only 36 reasons so far.

[Actually, Prakash now has 80 items in his list, as of the beginning of October 2010, and Dan's piece is still as compelling and succinct today as it was when I first read it.]

I’ve written before about things you can’t do with Windows but can with Linux, but it’s been a while. And the fact that I haven’t done anything with Windows for the past seven years or so means that there’s a whole universe of things I can’t do with Windows because I can’t do anything with Windows.

If I were really lazy (as opposed to just regular-lazy) I’d rehash items from Prakash and Dan’s lists for this month’s column and go back to watching unreleased Ren and Stimpy episodes on my 102 inch plasma TV [I did not own any such a thing as that; this sentence was pure sarcasm]. But because I’m truly conscientious I’ll add value for loyal CPU readers with my own top four list of what Linux gives me: resilience, maintainability, usability, and no/low-cost solutions.

Dan and Prakash go into the specifics, and give some good examples of applications that you can’t get on Windows. For example, Dan, a Web developer, likes having the ability to run different versions of Internet Explorer simultaneously using the IEs4Linux project. But for me, it all boils down to these four things:

Resilience: I include here what used to be meant by the now-meaningless marketing-speak “robustness”, as well as reliability and even security. Most software works fine running on systems with lots of resources and perfectly formed data. But in the real world I need an OS that gracefully handles too many processes running on a system with an old CPU and insufficient RAM. Remote system-pwning exploits of Linux are rare, email-borne viruses unheard of, and apps can fail without requiring a reboot.

Manageability: Can I easily back out software upgrades? Automate patch installation? Change configuration without rebooting? Administer the system free of activation codes? Backup personal configurations and data? I get all that with Linux. Can I run a supported version of Linux on otherwise obsolete hardware? Can Linux run on modern computer platforms from palmtop to mainframe? Yup. Can I examine and modify the source code? Yes. Do I worry about vendor lock-in, or getting stuck with proprietary products orphaned by bankruptcy or other corporate hi-jinks? Not at all.

Usability: I needn’t relearn everything every time some vendor decides it’s time to “upgrade” the UI. Configurability falls under this heading, too; I can configure the way almost everything works with KDE, GNOME or one of dozens of other desktop managers. Did I mention portability? Not just across platforms, but across time and space. I can carry system settings/desktop configuration/key data on a USB thumb drive and boot into my desktop from any PC. And I don’t need to re-learn (or re-configure) all my applications and utilities when I update system software, nor do I need to update scripts or applications every time I upgrade my OS.

No/low-cost: Microsoft’s deal last year with Novell signaled Linux is a legitimate OS, so why use the high-priced spread? You don’t really pay much extra for Windows when they buy a new PC (the already low-cost license is subsidized by shovelware publishers who pay to be included in the standard OS pre-install). Microsoft’s real cash flow comes from corporate sales, same as for Linux vendors like Red Hat and Novell; key developers of Linux components get funded by companies like IBM that are increasingly coming to rely on it. Personally, most of the “cost” I’ve recovered has been in the form of reducing the amount of time I spend fiddling around with computers running Linux—most of the things I need/want to do are pretty easy (or at least uncomplicated) and pretty well-documented online.

Sure, applications exist that aren’t available (yet) on Linux (for example, tax prep); and yes, there are times when proprietary software works better than open source. But for me, the benefits of open source simply overpower any disadvantages I’ve yet encountered.

I’m still using the wireless keyboard I bought back in early 2005, but to be honest is feels downright cheezy when I compare it to the Apple keyboards I’ve been using lately. Not so much that the Microsoft ergonomic keyboard isn’t well-designed, but it feels very plastic-y compared to the Cupertino keyboards.

Microsoft rocks!

What? Something nice to say about Microsoft? Yes. I needn’t like or use Windows or Office or Outlook or Internet Exploder (er, Explorer) or any other of their proprietary offerings, but if you’re looking to buy a quality ergonomic keyboard at the local big box or technology store, Microsoft is the only game in town.

When repetitive stress injuries like carpal tunnel syndrome loom, you’ve got to nip them in the bud with a properly adjusted chair, a keyboard tray to properly position your keyboard, a good mouse (I prefer trackballs) and an ergonomic keyboard. I went through it a few years ago, and rather than gamble on undoubtedly better but expensive mail-order keyboards, my choice at retail was between ultra-cheap no-names and Microsoft’s “Natural”s, with the Natural Keyboard Pro being the pick of the litter (avoid the cheesy, cheaper but oddly-laid out Natural Keyboard Elite at all costs).

The corded Microsoft keyboard I bought back then served me well and would still be on my desk if my son Zoom (a.k.a. “The Terminator”) hadn’t destroyed the “&” key one snowy afternoon. I’d hoped to use that keyboard until I retired, especially since I rarely see the wired version “Pro” keyboard in stores anymore. You’ve got to have the
ampersand, so I knew it was time to face my fear of wireless keyboards and find a Linux-compatible one.

I scooped a Microsoft Wireless Optical Desktop Pro v2.0 package (including a mouse with “Tilt-wheel for side-to-side scrolling”) and rushed home to try it.

The “Getting Started” booklet opened ominously: “Important! Install the Software”; then, I’ll have to restart the computer “if prompted to do so” (twice, maybe, for a Mac). Only then, I’m told, can I “Insert the Batteries”, plug it in and test it out.

“How could a Microsoft wireless product possibly work with Linux?” I can hear you asking. The answer is “Nanny nanny boo boo. It works as well as the wired version, maybe better.” I plugged it in and it worked; SUSE Linux even automatically and correctly auto-detected it.

Granted, I can’t use any of the “special” keys Microsoft grafts onto their keyboards, but it’s easy enough to see what signals those keys emit when pressed (using the “showkey -s” from a console session). Remapping them to do something useful within Linux would be more complicated, but it’s not something I’m even remotely interested in spending time on since there are plenty of keys that actually do work.

Linux rocks at automatic hardware detection and installation. I’ve collected dozens of hardware driver CDs for everything from video and audio cards to hard drives to trackballs, joysticks, and yes, keyboards, and never once have I suffered from not installing the software. It all works just great, and I don’t even even have to restart the system to get things going.

Maybe I don’t upgrade or change hardware on my system every day or even every week, but I do it often enough for me to be really happy about it all happening automatically, quickly, and correctly, without annoying system restarts.

Why don’t hardware manufacturers include Linux software with their products? Sometimes they do, but they usually get it wrong one way or another. For example, by making their drivers binary-only, closed source affairs, or by not making the drivers portable across CPU platforms. Invariably, the vendors just aren’t as good or as fast at writing reliable drivers as the open source community itself.

As for keyboards, I’m sure there are better ones than Microsoft’s, but they can’t seem to get their product into the retail channel. I’m not going to gamble $300 on a keyboard that I can’t at touch before I buy, no matter how much it promises to change my life. Microsoft delivers a quality product into half a dozen different major retailers within a 5-mile radius of my home, and I’m delighted to use their ergonomically correct keyboard.

Text is literally the lingua franca of computing. ASCII uses seven bits to encode 95 printing characters—a through z, A through Z, punctuation and symbols—and 33 non-printing characters—things like backspace and carriage return—that all computers can understand. Although most text editors default to the American English character set, they can also usually be reconfigured to use standard ISO character sets for virtually any language.

Text edtors differ from word processors and desktop publishing programs in that they produce raw text files rather than formatted documents with different fonts, font sizes, graphical elements, or other visual flourishes. But text editors add value in the way they are used: they can format and display text files in useful ways during the editing process, for one thing. Programming editors display source code formatted properly for the programming language being used, and can do things like highlight places where there are missing syntax elements.

Text editors can be handy for writing any kind of document that does not need to be formatted or that uses ASCII-based formatting like HTML tagging; being simple text files, writers and editors can easily compare different versions of the same document, and writers can concentrate on content rather than formatting issues.

Text-only applications abound in computing, and especially so in the open source world where so many tools were first developed when there was no such thing as a GUI. I’ll be posting more over the coming days about text editors, the Big Two text editors, and more about how to use them and what your other open source options can be.

The new year is always a good time to be thankful. In the US, we’ve got our political liberties, even if it’s easy to ignore the political aspects of free software and just assume that someone will always be willing to write it.

The biggest free software story this year was SCO’s legal crusade against Linux. The company formerly known as Caldera (as in “Caldera Linux”) now claims that somewhere within the Linux kernel is code improperly taken from the original UNIX kernel, which SCO owns. Which code? They won’t say. SCO is using the legal system to spread FUD and extort license fees from nervous Linux users.

I was surprised to discover that the SCO Group was still maintaining a website, because I’d thought they’d been ground into dust after their disastrous string of losing legal battles. Here’s the Wikipedia SCO Group entry, where you can read all the gory details; the bottom line is, Linux won, SCO lost.

Microsoft sure is an interesting company. How can software selling for hundreds of dollars in the US be sold profitably elsewhere for $40? That’s what happened when Thailand announced a $250 “people’s PC” running Linux. Windows and Office could have tripled the price tag, so what a pleasant surprise when Microsoft announced it would deliver Thai versions $40! Golly, I can’t wait for those discounts to make it here to the States!

Which reminds me of the recent “we’ll show you ours” sessions that Microsoft has been holding for Chinese experts to examine Windows source code. China is the world’s biggest undeveloped market, a hotbed of intellectual property piracy–and a growing center of free software use and development. Is China checking out Microsoft’s goodies to make sure there are no CIA backdoors, or to make sure they can put their own backdoors in?

This is so quaint: no one cares any more about Windows backdoors, but the government still would like to get a backdoor into the Internet. In Government Seeks Back Door Into All Our Communications, Seth Schoen writes about last week’s NY Times story about “a Federal government plan to put government-mandated back doors in all communications systems, including all encryption software.” at CommonDreams.org

According to the CIA World Factbook, Chinese per capita GDP in 2002 was about $4,400. I predict that in 2004 Microsoft will discover it can sell Chinese versions of Windows and Office for under $40 too.

In a widely reported story from 2007, Bill Gates was reported to have announced that a Windows/Office bundle for students would be sold in China for approximately $3; here’s one report: Microsoft Takes On Digital Divide with $3 Windows-Office Bundle. Those deals have yet to make it to the US, though.

Security and proprietary code may mix, but not well. Speaking out on how Windows-only networks represent a serious security threat cost respected security expert Daniel Geer his job when he co-authored the report “CyberInsecurity: The Cost of Monopoly“. Read the report; download and print it, just in case it disappears. Here’s the URL: http://www.ccianet.org/papers/cyberinsecurity.pdf. [NOTE: as predicted, that page is 404 and long gone, but the paper can be seen at Cryptome.org, at the link supplied].

To oversimplify, Geer et al state that a monoculture (e.g., the global network of Windows nodes) is susceptible to massive failure whenever a successful attack (e.g., NIMDA worm) is launched. In other words, if you’ve got nothing but sheep on a ranch the size of Texas, when anthrax breaks out you’re going to be neck-deep in a sea of dead sheep, which not only isn’t good for the sheep it’s bad for anyone within smelling distance of the ranch.

Microsoft is a big customer of Geer’s employer, @stake; Geer got the boot. Nice example, right? Geer is a star so he won’t be out of work long. But what if you’re Joe Sysadmin, are you going to risk it all to tell your boss that Windows-only is not just bad policy, it’s a huge security risk?

@stake was bought out by Symantec in 2004.

Finally, electronic voting. Since 2000, we discovered that it’s not just a matter of going to the polls on Election Day and pulling a lever, punching a card or marking a circle to have your vote count. Now the digital voting machine industry wants to sell us billions of dollars’ worth of “direct recording electronic” (DRV) voting systems (running Windows, of course!) that leave no auditable records, that use secret software to tally votes, and that can be tampered with undetectably.

This is a pretty sticky topic. Just Google “electronic voting problems” and you’ll see that these machines continue to cause problems all over the country and the world.

Surely it makes sense to give voters a printed receipt that they can check to verify their votes were cast–and recorded–properly. How much extra could it cost, and how much is it worth to us as citizens to know that our votes are tallied correctly? It makes business sense to keep track of every nickel, dime and dollar spent in a donut shop or drug store, which is why we get those nice printed receipts every time we buy anything. So why not when we vote? Whatever the cost, I’d consider it well spent, far more so than .

In fact, there has been ongoing support for open source software development of open source digital voting systems. You can find out more at The Open Source Digital Voting Foundation (OSDV).

For an article with so many links, I was surprised in a good way by just how many of them are still active and up to date (more or less) after more than five years. Great stuff if you’re interested in learning to program–or learning to teach young people how to program.

It’s more than just turtles these days with heavy-duty Logo

If French is the language of love, Logo is the language of learning–and not just learning to program, either. Elegant and simple, Logo is endlessly extensible and a powerful tool for study at almost any educational level, from elementary to post-doctoral.

If you think Logo is kid stuff, think again. Seymour Papert, mathematician, educator and co-founder of the MIT Artificial Intelligence Laboratory, designed Logo as a programming language easy enough for kids to pick up fast, but sufficiently robust to serve as a platform for computing projects of almost any complexity.

Programming in Logo means using the “turtle”: an entity that lives in the Logo environment and responds to commands you give it, either interactively at a command line, or through interpreted programs. Tell the turtle to go forward 50 units onscreen with the command “fd 50″; turn to the right with the command “rt 90″. Real programming comes quickly with loops (“repeat”), conditionals (“if”), and procedures.

Logo works well for teaching geometry: create a line by telling the turtle to draw as it moves forward; create a square by telling it, four times, to draw a line and turn 90 degrees. Robotics is just as well-suited to Logo: just turn the virtual turtle into a real robot. Research on “programmable bricks” at MIT during the 1990s produced the Lego Mindstorms robotics kit, powered by Logo.

My own first Logo was from Brian Harvey, creator of Berkeley Logo (a.k.a. UCBLogo, see Brian’s page), bundled with the book The Great Logo Adventure: Discovering Logo on and Off the Computer by Jim Muller.

You can find a downloadable version of Jim’s book if you dig into the Softronix website, as well as other Logo-related links.

Or you can try Brian’s three-volume The Great Logo Adventure: Discovering Logo on and Off the Computer for a rigorous introduction to serious Logo programming, at the Berkeley Logo site.

SUSE Linux (my [former] favorite) often chokes when compiling apps, like Berkeley Logo, from source, so I sought out a new Logo experience.

Only two–NetLogo, from Uri Wilensky, Center for Connected Learning and Computer-Based Modeling, Northwestern University and StarLogo, from the MIT Media Lab–met all my needs. Number 1: my Logo must run under Linux. Other Number 1: my Logo be free. And my final Number 1: an active project with a community of users and developers.

Where UCBLogo offers a basic interface to control and watch the turtle, NetLogo and StarLogo both make it trivial to create and control many turtles, making it dead simple to simulate large population behaviors. Both come with Logo code for modeling bird flocking, prey/predator population variations, gas molecule diffusion, creating fractal images, demonstrating mathematical and statistical theorems and much more.

The biggest hurdle: understanding that to program turtles, screen “patches”, and overall (“observer”) behaviors, you issue those commands in different modes. But once sorted out, it becomes almost trivial to tweak programs by adding, changing, or removing behaviors, er, I mean procedures. As Java apps, NetLogo and StarLogo both run everywhere, so choice is as much a matter of taste as function. I’ll continue using both as I climb the programming learning curve.

I could have bought a commercial Logo–and there are some very impressive looking commercial implementations available, mostly for Windows, sadly–but that’s not my way.  Whenever I buy software, I read the box or other promotional material, and I buy into the product: it’s going to change my life, so I’m happy to spend my hard-earned bucks for it.

But after the install comes the disappointment. My computer runs no faster, nor can it do much more than before. I still must study the new program to make it work; the license precludes returning it if I’m unsatisfied, or selling it to someone else for whom it may work better. And the revision that solves all my problems is invariably due just six months hence.

Free software never disappoints. If it helps at all, I’m happy. I’ve gotten more than my money’s worth, and I never suffer buyer’s remorse. Perhaps a commercial product could do so much more, but it’s still a gamble. I’d rather go for the sure thing.

Logo Resources

Free Logo Implementations

• Elica is a Windows-only, 3D Logo; apparently free but I couldn’t verify. It looks extremely cool and capable of doing some very excellent things like gaming, all in a format accessible to kids.
• StarLogo is an excellent version of Logo from MIT capable of programming multiple turtles as well as “patches” (bits of space within which the turtles act) that allow the turtles to interact with their environment. Released under a “free for research and education” license.
• NetLogo is another excellent research version of Logo, from Uri Wilensky, Center for Connected Learning and Computer-Based Modeling at Northwestern University; “free for research and education” license.
• MSW Logo is a Windows version of Brian Harvey’s UCBLogo.
• UCBLogo (a.k.a. Berkeley Logo) Brian Harvey, a lecturer at University of California, Berkeley, wrote and maintains UCBLogo; runs under Mac, Windows and *NIX.
• Logo++ is a simple, free (GNU license) version of Logo for Windows/X11.
• XLogo is a Mac OSX version of Logo released under BSD license.
• rLogo is Logo for the web; sparsely supported (?) and released under the GNU license.

Resources

• The Logo Foundation

Online/Free Logo Books/Courses

• The Great Logo Adventure [ (text), (companion CD)
• Computer Science Logo Style (vol 1, vol 2, vol 3)
• An Introduction to MSW Logo
• Learning LOGO

• Springpad looks to be an interesting concept of grabbing stuff off the web, and then Springpad goes ahead and does things like grabbing more information/links about what you’ve grabbed, and then give you the option to add/remove/sort through all the data.
• Verb is for lists, mostly task lists, with bells and whistles, and sharing options.
• MangoTask is part of a whole collaborative workgroup productivity suite. The whole thing looks very complicated, and it reminds me too much of things like Lotus Notes, circa 1996, for some reason.
• Orgu is another “simplest list service on the Internets” type of list service. It’s free, and it does lists, which you can share. And you can use their “ready made” lists for special occasions, if they have the right one for you.

EDIT: GQueue is NOT a Google application, and there is no connection between Google and GQueue other than that GQueue uses your Google account for authentication.

It is definitely designed as a “to do” list app, with check boxes on every item to indicate whether or not you’ve “completed” the task, as well as options for adding tags and tasks and sub-tasks and notes and so on. It’s also got an option for publishing your list, like so:

And that actually worked pretty well. I think this may work reasonably well. It certainly gives me something to think about in terms of how I’d like to be able to work with lists.

One might question whether Microsoft’s recent use of their portfolio of patents is a proper use of the United States Patent and Trademark Office, whose mission, “to promote the progress of science and the useful arts by securing for limited times to inventors the exclusive right to their respective discoveries” is defined in Article 1, Section 8 of the United States Constitution.

Microsoft’s bombshell about all the patents that Free/Open Source Software (FOSS) allegedly violates hit in mid-May, amid much hair-tearing, whining, gloom and doom. Because the copyright angle didn’t work out so well for SCO’s legal misadventures against Linux, it’s time to roll out Microsoft’s secret weapon: patents. Here we are in August, and (I predict) Red Hat and all the other big open source vendors are still doing business as usual, and I’ll tell you why:

Microsoft wants a piece of the pie, since every CPU running Linux is a CPU that isn’t running Windows, which means money not rolling into Redmond. Royalties are the name of the game, and the less muss and fuss customers put up to get indemnified against possible legal action from Microsoft, the better off for everyone involved, but most especially Microsoft. So, on May 13 Microsoft announced that FOSS infringes patents: Linux kernel (42), desktop environment and GUI programs (65), OpenOffice.org (45), email (15), and “assorted” (68).

Here’s where the “no-fuss, no-muss” comes in, and it’s a bit tricky. Microsoft won’t go on the record about which patents are infringed. They don’t have to, since they’re not suing anyone (yet). Going public would simplify things for open source developers to challenge—or work around—all those patents, and thus minimize Microsoft’s leverage over everyone who uses FOSS.

So Microsoft is discussing the patents privately with their big customers, by which I understand them to be saying “revealing the infringed patents under non-disclosure so Microsoft won’t lose their leverage”, and working out royalty deals. This works because according to the USPTO database Microsoft owns almost 7,000 patents; you’d have a hard time figuring out which patents they claim are violated by which open source software.

Now, there are a lot of questionable aspects to the whole situation. Does Microsoft’s deal with Novell to distribute vouchers for Linux equate to actually distributing Linux? Because if so, it means Microsoft would have to abide by the GPL. Does the lack of an expiration date on those vouchers mean that when GPL version 3 licensed software starts coming out, Microsoft will have to abide by the clause in GPLv3 that eliminates the cross-licensing loophole they exploited in their Novell deal?

How many of Microsoft’s patents will even hold up to courtroom scrutiny? Google “Supreme Court KSR patents” for details on the recent SCOTUS ruling that says the USPTO has been way too generous in granting patents for the past 25 years or so. The bad news for Microsoft is not just that new patents will be held to a higher standard but that current patents are going to be easier to defend against.

Microsoft could litigate Red Hat into dust, but not IBM, which depends heavily on Linux for its business—and has its own huge portfolio of patents (46,197) with which to counter-sue. Wanna bet Microsoft violates some of IBM’s software patents?

Microsoft’s brinksmanship is stunning: try to scare their biggest customers, who also happen to be some of the biggest users of FOSS, into paying royalties for Microsoft’s possibly questionable patents. Or else… what? The implied threat is that Microsoft will go ahead and sue the pants off their best customers, at great expense to all parties. Talk about going nuclear.

Very risk averse companies with deep pockets and little patience for surprises may be willing to pay off to avoid courtroom action. But as soon as Microsoft goes nuclear with a lawsuit they’ll have to lay down their patent cards, which means FOSS developers can challenge, work around or innovate over any potentially infringing code.

Could patents actually kill the FOSS movement? Remember, the whole purpose of the patent system is “to promote the progress of science and the useful arts”; if Microsoft tries using patents to kill off an entire industry, they might not be happy with the resulting backlash.