Development

Kolab 3.1 adds WebDAV

Test builds for version 3.1 of the Kolab groupware suite have been released. As before, the system provides a solid mail, address book, and calendaring solution with both web-based and desktop clients available. But this release departs from the project's traditional reliance on IMAP and adds support for a number of new services built on top of WebDAV. The change has the potential to simplify deploying Kolab for many users and offices.

Kolab is a client-server groupware system. The data is stored and managed on the server and there are web front-ends available (such as the Roundcube webmail client), but the emphasis has long been on supporting outside desktop clients as well. Historically, Kolab used special IMAP folders to store contacts and calendar entries on the server, in addition to the email messages for which IMAP was designed. The project defined its own XML-based storage format that encapsulates other data types into MIME messages on the server side. It covers iCalendar events, to-do tasks, journal entries, and free/busy statuses, plus text notes, vCard contacts, and shared "resources" (e.g., conference rooms).

The Roundcube client that ships with Kolab includes plugins to utilize the non-email IMAP storage directly, and KDE's Kontact has long been the primary desktop application supported by Kolab. Support for the Kolab storage format also made its way into other groupware suites like Citadel. Many other client programs, however, required an add-on—such as SyncKolab for Mozilla Thunderbird or several proprietary connectors for Microsoft's Outlook. There are also several mobile clients for which Kolab connectors are available, although not all of them support the most recent version of the storage format, 3.0.

As one might guess, version 3.0 of the storage format was rolled out in conjunction with version 3.0 of Kolab itself, which was released in January 2013. Thus, the addition of an entirely new data access protocol for the 3.1 is all the more remarkable.

The project announced the plan in March, calling it both "the biggest and most exciting thing on our roadmap" as well as "the one people have asked us for most." Using the SabreDAV WebDAV library (in order to not "reinvent the wheel", according to the announcement), the project would add CalDAV and CardDAV access to scheduling and contact data. As the 3.1 alpha announcement explains, the old IMAP-based access protocols are not going away. The CalDAV and CardDAV functionality is implemented as a separate data access layer dubbed "iRony" by the project. The most obvious gain from the move is immediate access to Kolab server data from the sizeable assortment of existing CalDAV- and CardDAV-aware applications. The built-in Mac OS X calendar and address book fall into this category, but so do quite a few open source programs, including web-based, mobile, and desktop clients.

Kolab 3.1 is also adding support for general-purpose file storage, which will also be accessible over WebDAV. But the file storage functionality is not limited to WebDAV: the project has also implemented an IMAP-based file store, and it has worked with the ownCloud project to make an ownCloud server accessible as storage, too. The ownCloud-Kolab integration is not yet ready, but the Kolab project describes it as "trivial", as is integration with Dropbox or other cloud storage systems.

The 3.1 alphas also include the ability for multiple users to share mail folders. That is, not only can multiple user accounts read mail from a shared inbox (such as contact@example.com), but those accounts can also send messages as contact@example.com. Other changes incorporated include new administration features, like the ability to define a policy for how user account UIDs are generated, and support for managing multiple email domains within a single Kolab server.

The alpha release showcasing all of these new features is available from the project's Git repository or in packages for CentOS. Debian packages have been assembled in the past, but so far none are available for the 3.1 code.

Historically the term "groupware" tended to apply to information systems designed to support office environments, but it recent years that has changed. The public scandals about privacy breaches, surveillance, and user tracking that seem to plague every major web application service provider have served to raise the profile of open source groupware suites—increasingly, they are seen as a self-hosted alternative to relying on Google, Microsoft, and Apple for valuable web services, even if deployed only for a single user. The Kolab team is clearly aware of this shift, and cites the system's ability to serve as a private service beyond the reach of the NSA PRISM program.

Kolab has been in active development for just over ten years, and in that time it has earned a reputation for quality production in more ways than one. There are quite a few open source groupware products on the market, but many of them suffer from a lack of client compatibility, either working only with their bundled web front-end, or focusing on a small set of third-party client applications. Kolab has long offered support for a wide assortment of client endpoints, and implementing CalDAV and CardDAV access is likely to extend the reach of that assortment in a big way.

Support for these protocols is also a positive sign for CardDAV and CalDAV themselves, which in the past have been criticized as being overly complex (which, for example, gave rise to the now-abandoned GroupDAV protocol). Their slow adoption can be seen as part of a march toward compatibility among competing groupware projects, however. When all is said and done, it will be interesting to see whether this WebDAV support signals the start of a slow decline in support for Kolab's IMAP-based data protocols. But perhaps that would not be a tragic loss in the long run, as it might allow the project to expend less energy maintaining private protocols. The project does not like reinventing the wheel, after all, and as of Kolab 3.1 alpha, it is now responsible for maintaining two separate access protocols.

Comments (1 posted)

Quotes of the week

— Ryan Paul

— Anil Dash

Comments (none posted)

Wine 1.6 released

Version 1.6 of the Wine Windows emulation system is out. "This release represents 16 months of development effort and around 10,000 individual changes. The main highlights are the new Mac driver, the full support for window transparency, and the new Mono package for .NET applications support."

Comments (15 posted)

Razor and LXDE-Qt merging

Jerome Leclanche has posted an announcement that the Razor and LXDE-Qt (LXDE ported from GTK to Qt) lightweight desktops will be merging under the LXDE-Qt banner. "So what happens now? Our two teams have met up and discussed the issues and we have decided that the best course of action for both projects is to focus on a single desktop environment, instead of two. There have been talks of "merging" ever since LXDE-Qt was announced. Having taken the decision to collaborate, we've all had the pleasure of working together already. Our plan is to cherry-pick the best parts of Razor and LXDE and include or port those to LXDE-Qt. Other components will be ported straight from GTK code or rewritten from scratch. In the end, we want to offer the best possible experience while reusing as much code as possible. It will not be an easy process and as always, we welcome anyone who wishes to help, be it with development, translations, or general feedback." In addition, Leclanche mentioned an offer to bring Razor (and now, perhaps, LXDE-Qt) under the KDE umbrella.

Comments (37 posted)

Apache OpenOffice 4.0 released

The Apache OpenOffice 4.0 release is now available. "OpenOffice 4.0 features a new, more modern user interface, improvements to Microsoft Office interoperability, enhanced graphics support and many other improvements." See the release notes for lots of details.

Comments (51 posted)

U-Boot 2013.07 available

Version 2013.07 of the U-Boot bootloader has been released. This version is notable for adding support for doing cryptographically verified boot. The announcement adds: "There's still work to be done to make it easier to tie things to a TPM (it's just scriptable today), but hey, there it is. One might even say this is one of those technical features that makes companies chose something other than U-Boot.."

Full Story (comments: none)

Development newsletters from the past week

• Caml Weekly News (July 23)
• What's cooking in git.git (July 18)
• What's cooking in git.git (July 21)
• What's cooking in git.git (July 22)
• Haskell Weekly News (July 17)
• OpenStack Community Weekly Newsletter (July 19)
• Perl Weekly (July 22)
• PostgreSQL Weekly News (July 22)
• R Journal (June)
• Ruby Weekly (July 17)
• Tor Weekly News (July 24)

Comments (none posted)

Airlie: Introducing Virgil - 3D virtual GPU for qemu

Dave Airlie has posted an introduction to Virgil, a new 3D virtual GPU project for the QEMU emulator, for which he is seeking developer input. The goal, he says, is "a 3D capable virtual GPU for qemu that can be used by Linux and eventually Windows guests to provide OpenGL/Direct3D support inside the guest. It uses an interface based on Gallium/TGSI along with virtio to communicate between guest and host, and it goal is to provided an OpenGL renderer along with a complete Linux driver stack for the guest." Airlie also goes out of his way to be clear that he is recruiting developers, not users, noting "I can't stress this strongly enough, this isn't end user ready, not even close".

Comments (24 posted)

Xiph.org: Introducing Daala, part 2

Xiph.org has published part 2 of its introduction to the new Daala video codec. This edition covers frequency domain intra-prediction. "Daala predicts frequency coefficients in the frequency domain, rather than pixels in the spatial domain. Frequency coefficients are not shared between blocks, allowing us to keep the classic definition of 'neighboring' as well as avoiding any circular decoding dependencies. In addition, Daala's frequency-domain prediction draws from a larger input data set consisting of up to four complete blocks. In theory, the additional degrees of freedom could give us a more efficient predictor."

Comments (none posted)