No longer impenetrable

Oracle closes Fortress language down for good

Chris Mayer

After a decade of development, Oracle decides to axe the long-running high performance language project.

Oracle is to cease all production on the long-running Fortress language project, seeking to cast aside any language that isn’t cutting the mustard financially.

Guy Steele, creator of Fortress and also involved in Java’s development under Sun jurisdiction, wrote on his blog: “After working nearly a decade on the design, development, and implementation of the Fortress programming language, the Oracle Labs Programming Language Research Group is now winding down the Fortress project.“

He added: “Ten years is a remarkably long run for an industrial research project (one to three years is much more typical), but we feel that our extended effort has been worthwhile.”

Originally coming out of Sun Microsystems Labs and influenced by Fortran, Fortress is a rich parallel JVM-running language that drew heavily from Java but pushed things on by adding in the ‘for’ loop for example. Its goal was to become the premier choice for advanced languages, and make the whole process of creating high performance software easier.

Its syntax bears resemblance to the likes of Scala and Haskell, and herein lies its failure; it just couldn’t shift those two in terms of developer mindshare and quickly fell down the pecking order, certainly after it became part of Oracle’s wider portfolio after purchasing Sun. With the last stable release 10 months ago, the writing was clearly on the wall.

The team behind Fortress have met their fair share of teething issues in targeting the JVM with the Fortress type system. Steele notes ‘severe technical challenges’ aside from that mismatch, such as the implications of implementing a symmetric multimethod dispatch. But that isn’t a slight on the team at all – their work paved the way for others to be daring and break conventions in language development. Both Clojure and Scala have dabbled along the way with revolutionary features like Fortress tried.

Steele went on further to say that the team were “now unlikely to learn more (in a research sense) from completing the implementation of Fortress for JVM” so its the end of the road for their part. But Fortress will still remain available as an open source language like it has since 2007’s olive branch to the community, just before a 1.0 version appeared for the JVM.

Whilst production won’t end straight away, we can expect a lengthy winding-down period in order to, in Steele’s words, get “the code and language specification into the best shape that we can.” Plus there will be some research papers provided to perhaps stoke the fires of some wanting to use Fortress’s ideals in their own project.

It’s just another victim in Oracle’s cost-cutting measures (see OpenOffice), and you begin to wonder who is next in the firing line in this initiative. Fortress was really just a research project and nothing more. We shouldn’t underestimate its impact though in inspiring the design of other languages.

Fortress had a good run for a research project, but considering it only ever had five blog entries to its name, it’s surprising it even lasted three years at Oracle. We wish the team all the best in their other future endeavours.

Inline Feedbacks
View all comments