Archive for July, 2007

Content reuse from flickr

At the OSS Watch conference in April last year, I took photos of some of the speakers and activities and uploaded them to flickr, under a creative commons licence. Yesterday my photo of Glyn Moody taken at the event was taken from there to illustrate his Wikipedia entry.

It’s not the best portrait in the world. The composition, lighting and resolution could all be improved. But it was good enough, and when I uploaded it to flickr, I tagged it with who was it in, so it was easy enough to find.

These two factors, good enough and easy enough to find; are key to content reuse. They’re also key to open source: software has to be good enough to convey a distinguishing characteristic and clear enough in describing the distinguishing characteristic for those who need that characteristic to find in via their favourite search engine.

In this case the distinguishing characteristic required was a picture that was clearly of Glyn and which represented him fairly and honestly, and the photo, for all its faults, is clearly of him and is a fair and honest representation. My choice of creative commons for flickr photos is was a no-brainer for me, but not for all flickr users: of more than half a billion photos, only tens of millions are licensed under the creative commons.

Licence choice and communities: R and GRASS

Several months after it was first published, I recently stumbled across an interesting post by Michael Tiemann of the OSI about the way in which two projects R (a statistical package) and GRASS (a geographical information system) are co-evolving:

GRASS and R both offer integration modules so that a GRASS-centric user can use R, or an R-centric user can fully exploit GRASS. Both GRASS and R also implement integration modules for the PostgreSQL database, and PostgreSQL also returns the favor so that a PostgreSQL-centric user can reach into both GRASS and R (and vice-versa).

This to me is an example of the kind of peaceful co-existence which open source encourages. Each of these communities has a very different view of the world, leading to different ways of representing, accessing, querying and storing data. Three-way interoperability means that all three communities have the best of all three worlds. PostgreSQL is, of course, a much larger player than R and GRASS, but using it as their back end R and GRASS have access to a much larger pool of optimisation effort than they otherwise would.

Key to this, of course, is the use of compatible open source licenses by both R and GRASS.

Debian Art: licencing choices

The Debian community has grown a new limb in the form of Debian Art, a project to collect and coordinate wallpapers, splash screens, sounds, icons, logo, banners and similar art work for Debian releases.

Interesting, all the artwork is licensed under the GPL2 or released into the public domain. The GPL family of licences are specifically written for use with the source code of executable programs and fit poorly with textual or image content. Public domain is tricky, particularly in Europe where there are stronger moral rights than in some other areas.

I was surprised to discover that creative commons are not offered as options. It turns out the problem centres on long-standing incompatibilities between the creative commons licences and the Debian Free Software Guidelines (DFSG). The DFSG is the key rallying point for the Debian community, so Debian art (like every project in the Debian community) is constrained by it. Still it is disappointing to see that the project is not reaching out to other open communities and encouraging interoperability by allowing dual-licensed content.

New Emacs version gets things done

A new version of the fabled Emacs text editor has been released, with a range of new features including a “Getting things done” mode.

Emacs and vi are the traditional text editors used on UNIX and Linux platforms. vi has a minimalist philosophy, including a rich, powerful set of features which have remained largely unchanged for 30 years. Emacs has a core basic of functionality on which users are encouraged to build “modes” for different activities. Thus there are modes for reading email, browsing the web, directly editing compressed files and even those only accessible over ftp, as well as more mainstream modes to enabling the editing and correction of a wide variety of textual file formats.

The software design philosophy choice has interesting effects on the resulting user community: whereas Emacs has a large community of contributors still building the system, vi is essentially frozen with a much smaller contributor community focused on bug fixes, portability and optimisations, even though both appear to have a similar number of users. The smaller size and well-defined feature set led to vi to be selected over Emacs for standardisation in the POSIX standard.

The new version will take several months to filter through the distribution channels before appearing in the next upgrade of the major operating systems.

Personally I’m disappointed that “better Unicode support” is still on the Emacs TODO list rather than on the new features list.