May 1 – 6 , 2006, Dagstuhl Seminar 06181
Ralph Johnson (University of Illinois – Urbana-Champaign, US)
Shriram Krishnamurthi (Brown University – Providence, US)
Thomas Kühne (Victoria University of Wellington, NZ)
Michael Sperber (Tübingen, DE)
For support, please contact
Even as research into typed programming languages continues apace, languages with no prescribed type system, which we shall refer to as latently-typed languages—such as Erlang, Lua, Python, Ruby, Smalltalk, Scheme, Self—continue to be a fertile terrain for innovative research. This research spans a broad range of subject areas such as language design, programming environments, programming methodology, education, cross-language integration, and application frameworks. It is not entirely coincidental that many of these innovations have emerged in languages without a single, fixed type system.
Innovative work on latently-typed languages, however, has been done in diverse communities that have tended to not publish in the same conferences or attend the same venues, thereby losing valuable cross-pollination. These barriers have been erected partly because of an unfortunate segregation into historical language paradigms (primarily “object-oriented” versus “functional,”, and because of differences in emphasis (“development” versus “semantics,” “industrial” versus “research,” and so on).
The goal of the seminar was to unite these disparate communities to exchange ideas and identify key areas for future research, and to lay the groundwork for future cooperation. Thus, the purpose of the workshop was acquainting the participants with work that has been happening in other communities, rather than being a forum for presenting novel research ideas.
Prior to the workshop, a mailing list with all participants was established. Its purpose was to solicit input on organization and content. The mailing list established areas of interest for the discussion, and collected many other suggestions that helped the organizers shape and prepare the workshop itself. The organizers consolidated the areas of interest, and used the mailing list to form groups of speakers to prepare survey talks on these subjects. Each of these groups then collaborated on their talks, which formed the backbone of the workshop schedule.
The presentation of novel research was left to a “soap-box session” of ten-minute talks, as communication of novel research was not a primary aim of the workshop. Moreover, a distributed demo session allowed implementors to show their systems.
As the workshop progressed, the lively after-talk discussions and evening conversations soon made it clear that a large number of subjects could not be accomodated in the survey talks, but was nevertheless interesting to many people. A “wish list” of additional talk subjects, paired with offers from the participants to give talks, led to yet another open session with longer talks. An open discussion session focussed on controversial statements, collected in a “Controversy Corner” during the workshop.
More discussion after the workshop produced a list of major “must-read” publications related to latently-typed languages. (The list is being maintained at http://www. deinprogramm.de/dagstuhl-06181/papers.html.)
- Programming Languages / Compiler
- Latent typing
- Object-oriented programming
- Functional programming
- Introductory teaching
- Domain-specific languages
- Programming environments