User: Password:
Subscribe / Log in / New account

Leading items

The Grumpy Editor's 2008 retrospective

By Jonathan Corbet
December 24, 2008
Holidays are an exercise in tradition. One of the more charming holiday traditions around LWN is to look at the predictions made at the beginning of the year and measure them against reality. There is, after all, great value in things which make us laugh. This year's predictions were featured in the January 3, 2008 edition. As might be expected, some of them were better than others.

What was predicted

Your editor's first prediction was that support for Flash playback would mature in 2008. In some sense, that may be true. Your editor's desktop system, running the Rawhide build of Gnash, can now faithfully display a wide variety of Flash ads, web site "intros," and various other thoroughly useless bits of media. A Flash-based "interactive tour" offered by LWN's bank worked nicely. But support for many other Flash features, including audio and simple playback from online sites, still is not especially solid, and other interactive Flash applications do not work at all. This problem, it seems, is still not solved.

The prediction of the KDE 4.0 release required little in the way of foresight, as did the prediction that users would be unhappy. That stage was well set before the beginning of the year. A continued focus on power management was also an easy thing to foresee; there will be great value in making our systems more power-efficient into the indefinite future.

Flush from those two obvious successes, your editor went off and stated that the bulk of the realtime tree would be merged into the mainline kernel by the end of the year. Oh well. Your editor should know by now that expecting deterministic merge times for realtime patches is a sure path to disappointment; latencies in this area are always higher than one would like. In this case, the realtime developers got stuck in a high-priority interrupt (taking over the x86 architecture) with the result that realtime work got preempted and suffered from severe starvation.

As predicted, debate over Microsoft's OOXML format continued, and Microsoft succeeded in obtaining standard status for that format anyway. Things have since gotten quieter, though, perhaps because people see it as a done deal and no longer worth fighting about.

The GPL was the subject of two predictions this year. One was that more projects, perhaps even glibc, would move to GPLv3. There is a steady stream of analyst verbiage to the effect that GPLv3 is quickly growing in popularity (example), but the truth of the matter is that the number of conversions in projects which really matter appears to be low. Projects with significant numbers of developers and users continue to approach GPLv3 with caution.

The other prediction was that GPL enforcement actions would continue, and perhaps grow. The recent FSF lawsuit against Cisco makes it clear that the GPL enforcers are serious about what they are doing. Your editor cannot help but wonder, though, whether the increasingly litigious actions by the Software Freedom Law Center might not eventually lead to a serious backlash within the community. We are about freedom, not punitive damages. Enforcement of the GPL is necessary if we expect our licenses to be taken seriously, but overly zealous - or greedy - litigation could encourage those who say that use of free software exposes companies to an unacceptable level of risk.

Your editor included a rosy prediction about the One Laptop Per Child project and where it would go over the course of the year. In fact, OLPC has continued to work toward its goal of putting laptops into the hands of children around the world. But your editor completely missed the way internal divisions would rise to the surface and distract OLPC developers from what they are trying to do. OLPC seems to have moved beyond the worst of that, and much-needed development on the Sugar software continues. But the project seems far from its original goals, and the increasing popularity of ultra-mobile systems, while vindicating the original vision behind the OLPC hardware, threatens to render the XO hardware obsolete and irrelevant.

Ever the optimist, your editor said that the days of hardware hassles would be over. We are closer. Finding an off-the-shelf system - server, desktop, laptop, or palmtop - which is fully supported by Linux is now easily done. OK, maybe the modem is not supported, but few people will be inconvenienced by that omission anymore. That said, there will probably never be a shortage of uncooperative hardware manufacturers; if we value our free operating system, we must continue to support manufacturers who work with our community, and avoid those which do not.

The prediction that the intensity of competition between distributors would increase was reasonably well satisfied. One need only look at Novell's "migrate from Red Hat" offering or the continued attacks on Ubuntu, not all of which have to do with its community participation.

Finally, the three "community" predictions at the end of last January's article were all satisfied reasonably well. None of them were especially daring, so that should not be surprising.

What was not predicted

One commenter in January asked about the lack of predictions about SCO. In December, it is hard to say that SCO deserved a place there. The company still exists in some form, but it no longer has much to warrant the attention of the Linux community. Your editor predicts that there will be no SCO predictions in 2009 either.

So what else did your editor miss? Perhaps at the top of the list is the evolution of the Linux platform as it is used in mobile devices, and in cellular telephones in particular. Google's (unpredicted by your editor) Android platform has made a splash, regardless of what one might think of its openness. The first Android phone has been reasonably well received, and it would appear that more are on the way. The merger of the LiPS and LIMO consortia shows that some consolidation is happening in this area. The announced plans to open Symbian were also an interesting development. In the near future, the handset business seems likely to be firmly dominated by free software - though, alas, the bulk of those handsets will not be designed to pass the benefits of that freedom on to their owners.

Your editor has often predicted software patent troubles, though he did not do so in 2008. What was completely unforeseen, though, was Red Hat's resolution with Firestar Software. The company got itself out of a patent bind, and, in the process, removed the patent as a threat to the wider development and user community too. We may see this sort of solution repeated for patent problems in the future - if we are lucky.

Finally, unpredicted - and unpredictable - was the series of "infrastructure issues" which shut down much of the Fedora project for a good month. That episode showed us a number of things: how much some of us depend on our distributors' infrastructure, how vulnerable we can be to intrusions, and how the interests of the companies behind some distributions can interfere with the availability of useful information. Months after the fact, we still have no idea what happened with the Fedora project; it is not unreasonable to wonder if we will ever know.

Despite problems like that, and other small distractions (the total meltdown of the global financial system, for example), Linux has only grown stronger over the last year. Our community has grown, our software has gotten better, and the economy around free software has gotten stronger. Your editor predicted that, too, but not even he is so arrogant as to claim credit for having foreseen something nearly as obvious as the sunrise.

Comments (17 posted)

PDF-based presentations with 3-D effects

December 24, 2008

This article was contributed by Bruce Byfield

At first, the idea of adding 3-D transitions to command line presentation software may give you a kind of cognitive dissonance. Just as you would if someone had added a GPS tracking system to a one-horse cart plodding along at two kilometers an hour, you have to wonder why anyone would bother. But, the dissonance disappears as you start to explore the control and precision you have in command-line programs like PDFCube and Impressive (formerly KeyJNote). Both are small and efficient programs that allow you to add transitions and other special effects to PDF-based presentations, although the range of options varies considerably between the two programs.

Before using either PDFCube or Impressive, you need to have to have support for 3-D graphics installed. PDFCube works well with OpenGL, as well as with the drivers and video cards listed on its hardware compatibility page. By contrast, Impressive is somewhat more erratic under OpenGL, with some transitions displaying slowly, especially when you have less than two gigabytes of RAM available. However, by picking and choosing effects, you can still test drive Impressive without resorting to proprietary drivers.

Both applications are available as source code from their project sites. However, you will also need to install dependencies for PDF support, such as Poppler for PDFCube, and Xpdf Reader or Ghostscript for Impressive. Impressive also requires Perl and Python. For convenience, you may prefer to use the Debian packages for both programs, or, in the case of PDFCube, the packages available in the Fedora and Ubuntu repositories. Impressive is also available for OS X and Windows.



With version 0.0.3 just released, PDFCube is more a proof of concept than a finished application. In fact, it currently has only one transition effect — a spinning cube. However, a day after the latest release, maintainer Mirko Maischberger has already posted a brief announcement on the project home page that he has already started work on "an abstraction layer for 3D effects (cube, fading, cover flow) to be done in C++ and OpenGL)."

What you currently have in PDFCube is the basic engine. No options are available, so all you need to type to try PDFCube is pdfcube filename.pdf.

However, before trying PDFCube, take the time to read its man page to learn how to navigate within the program. Unlike full office applications like Impress or KPresenter, PDFCube is driven completely by keyboard commands, and — so far, at least — does not work with the mouse at all.

Fortunately, the basic commands are few. You press the 'c' or space key to move to the next page of a presentation using an effect, or the PageUp key to move to the next page without any effect or the PageDown key to move to the previous page without effect. You can also use the 'h','j','k', and 'l' keys to zero in on one of the corners of the current page, or the 'z' key to zoom in on the center. Pressing any of these keys zooms out again, while Esc stops the presentation. These are all the controls that you are likely to need.

As Maischberger suggests on the project home site, the spinning cube is easy to overdo, so you might want to limit its use to major transitions. You can impose this limit by adding the page numbers before the places you want the transition. For instance, if you entered pdfcube filename.pdf 0 3, you would have the spinning cube between pages 1 and 2 and pages 4 and 5 only. Other transitions would lack the effect.

Another point to be aware of with PDFCube is that is designed for landscape oriented pages. You can display PDF files with a portrait orientation, but the application currently gives you no way of scrolling up or down the page. But, this limit aside, PDFCube shows a simplicity and performance that you don't often see in its desktop equivalents.


At version 0.10.2, Impressive is already much more complete than PDFCube. It not only runs slideshows from directories with BMP, JPEG, PNG, and TIFF graphics as well from PDFs, but also includes a complete set of controls for fine-tuning how its presentations run — to say nothing of several unique controls for running a presentation.

You can view a complete list of options with impressive --help, or from the project documentation page. They include options to set up an automatic slideshow, complete with a loop from the end back to the beginning, to set the size of the presentation window, and just about every other aspect of the running and appearance of a presentation that you can imagine. Two especially noteworthy options are -d, which allows you to set a time for the entire presentation, then pace yourself by an unobtrusive bar along the bottom of the screen, and -u, which polls original files periodically to see if they are updated.

If you want to use slide transitions, you will need to enter impressive --listtrans to see a list of over 20 possible transitions. All the transitions have names like SlideUp or WipeDownRight that are clear enough to be self-explanatory, although the help screen does include a slightly longer description. You can use a transition by adding its name with the -t option. However, unlike PDFCube, Impressive currently limits you to a single transition for the entire slide show — a limitation that might frustrate some users, but also prevents the aesthetic disaster of anyone using too many.

In addition, Impressive includes several handy controls. Pressing the Tab key opens a view of all the slides in the presentation, while pressing the Enter key enables a spotlight that follows the mouse and can be used as a built-in pointer.

[Impressive highlighting]

Still another option is to draw an enclosed shape with the mouse, which results in the rest of the screen darkening and blurring, so that the audience's attention is focused on the area you defined. You can add multiple highlighted areas, each of which you can close with a right mouse-click. The screen returns to normal when you close the last highlighted area.

Impressive's view of all Slides is reminiscent of the slide view in many programs, or the Sun Presenter Console for, but its highlight boxes and spotlight are both features that I haven't seen in desktop-oriented programs. These features alone make Impressive worth a look, but more experienced users might also appreciate the wealth of available options — even if they don't often use many of them.


Both PDFCube and Impressive are works in progress, with some ways — and, at the current rate of development, perhaps some years — to go before their 1.0 releases. However, in the current versions, PDFCube has the superior basic engine, while Impressive allows users the greater control. Despite PDFCube's lack of options and Impressive's mediocre OpenGL support, both are worth keeping at least an occasional eye on.

In their separate ways, both demonstrate that, contrary to what many desktop users seem to assume, command line applications are not just archaic remnants. You need time to enter all the options in a command line application, but, if you take the trouble to familiarize yourself with the applications, you may find their controls easier to use than the cluttered editing windows of a desktop application like Impress. Far from being outdated, applications like PDFCube and impressive are practical demonstrations that command line applications can be both modern and innovative.

Comments (5 posted)

The 2008 Linux and free software timeline

Here is LWN's eleventh annual timeline of significant events in the Linux and free software world for the year.

As always, 2008 proved to be an interesting year, with great progress in useful software that made our systems better. Of course, there were some of the usual conflicts—patent woes, project politics, and arguments over freedom—but overall, the pace of free software progress stayed on its upwardly increasing trend. 2008 was a year that saw the end of SCO—or not—the rise of Linux-based "netbooks", multiple excellent distribution releases, more phones and embedded devices based on Linux, as well as major releases of software we will be using for years (, Python, KDE, ...). We look forward to seeing what 2009—and beyond!—will bring.

This is version 0.85 of the 2008 timeline. There are certainly errors and omissions; if you find any, please send them to rather than posting them as comments.

  • January: SCO delisted, Sun buys MySQL, KDE 4, 2.6.24, ...
  • February: Mozilla Messaging, LSB 3.2, vmsplice(), ...
  • March: OpenOffice, GCC, ...
  • April: OOXML approved, 2.6.25, Ubuntu 8.04, ...
  • May: Fedora 9, Sugar Labs, Debian OpenSSL bug, ...
  • June: Wine 1.0, openSUSE 11.0, Firefox 3, ...
  • July: Kaminsky DNS flaw, 2.6.26, Stormy Peters, ...
  • August: Fedora infrastructure, JMRI, Debian, ...
  • September: Kernel Summit, Linux Plumbers Conference, Firefox EULA, ...
  • October: GIMP 2.6, Python 2.6, 2.6.27, Ubuntu 8.10, ...
  • November: Theora, iPhone Linux, Fedora 10, MySQL 5.1, ...
  • December: Python 3.0, Debian woes, FSF vs. Cisco, Slackware 12.2, openSUSE 11.1, ...

For previous years' timelines, head over to our timeline index.

Comments (none posted)

No Weekly Edition on New Year's Day

A longstanding holiday tradition at LWN is to not publish our usual Weekly Edition during the last week of the year. It's a good time to catch up with friends and family, and there is usually not a whole lot of news to report during that time anyway. This year, that break lands on what would otherwise be the January 1 edition. We'll post occasional articles, but the next Edition is due on January 8. Thanks to all of LWN's readers for another great year, and best holiday wishes to all of you.

Comments (4 posted)

Page editor: Jake Edge
Next page: Security>>

Copyright © 2008, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds