Lucid Inc.

Lucid Incorporated was a software development company founded by Richard P. Gabriel in 1984, which went bankrupt in 1994.

Beginnings

Gabriel had been working for Lawrence Livermore National Labs on a computer hardware project called "S1", the first incarnation of which used a CISC processor. The compiler technology necessary to take full advantage of the instruction set proved to be infeasible to develop, and the second incarnation was instead a RISC processor. The secret ingredient was laser pantography, a process which used a focussed laser to etch the semiconductors of the chip rather than the usual photographic mask. The team working on this project began writing a business plan to produce supercomputers, including all its basic software, with this spun-off technology. During the process of fleshing out the business plan and seeking venture capital, the goal changed from producing supercomputers to producing commercial implementations of, and development environments for, the recently finalized programming language Common Lisp, which Gabriel expected to become the standard AI language. Lucid's prospects were enhanced by the fact that five of the ten initial founders (Bill Scherlis, Scott Fahlman, Eric Benson, Rodney Brooks, and Gabriel) were on the committee that had written the Common Lisp standard; moreover, Gabriel was the initiator of the Common Lisp design effort, Fahlman was its "de facto" leader, and both Gabriel and Fahlman were part of the standard's five-person core group of authors (known as the "Quinquevirate"; these were Guy L. Steele Jr., Fahlman, David A. Moon, Daniel L. Weinreb, and Gabriel). The first CEO was Tony Slocum, formerly of IntelliCorp; and Gabriel was Lucid's Chief Technical Officer (CTO) and first president. The interpreter and the environment for Common Lisp they intended to market was not for the then-dominant Lisp machines, however. Regular workstations had become fast enough to reasonably run Lisp languages, and it was this, much larger market, that Lucid targeted.

Six months after getting $600,000 in seed money, Lucid had a bare-bones implementation running. On the strength of this, they received a further $3,500,000 in venture capital, and had OEM agreements with Sun Microsystems, Apollo Computer, and Prime, and they had an agreement with Symbolics that they would put a Lucid Common Lisp cross-development environment on their Lisp machines, enabling programmers to develop software on Symbolics Lisp machines that could run on Unix workstations under Lucid Common Lisp. (Symbolics apparently agreed to this because they had drastically underestimated the speed of Lisp on Unix workstations to be somewhere around 17 times as slow as on a Symbolics 3600.)

Initial success

The product the company ultimately shipped was an integrated Lisp IDE for Sun Microsystems' RISC hardware architecture—this sidestepped the principal failure of Lisp machines by in essence rewriting a lesser version of the Lisp machine IDE for use on a more cost-effective and less moribund architecture.

Despite its success in shipping its environment with many computer manufacturers like Hewlett-Packard, IBM, Intel, Groupe Bull, DEC etc. Lucid suffered several setbacks during this period, including a new CEO, Bob Kadarauch (the old one, while extremely competent and effective, suffered from depression when the company failed to take off despite his doing everything right), and a massive investment trying to duplicate Symbolic's Lisp machine OS for an IBM workstation (the RT-PC) which flopped. The fact that Lucid, to focus on producing quality programs, had no contact with actual customers meant that the systemic problems with the whole Lisp industry were exacerbated by a lack of feedback from the users.

In 1987, Gabriel resigned as President, but remained its CTO.

Decline

Eventually Lucid's focus shifted (during the AI Winter) from the Lisp market (which was still growing at this time) to an Object-oriented IDE for C++ called "Energize". A core component of the IDE was Richard Stallman's version of Emacs, GNU Emacs. GNU Emacs was not suitable for Lucid's needs, however, and several Lucid programmers (including Jamie W. Zawinski) were assigned to help develop GNU Emacs to meet those needs. Friction arose between the programmers and Stallman over how to handle GUI issues, and Lucid forked the software—thus they were primarily responsible for the birth of XEmacs.

By 1994, Lucid's attempts to reinvent itself as a C++ company, and its neglect of its still profitable Lisp sideline had ended in failure, and the company's revenues fell to levels which could not sustain it. Lucid Incorporated went bankrupt. The rights to Lucid Common Lisp were sold to Harlequin Ltd. which was bought by Global Graphics; Global Graphics then sold the right to Xanalys Corporation, which spun off LispWorks, the current rights holder which sells Lucid Common Lisp under the "Liquid Common Lisp" [http://www.lispworks.com/products/lcl.html] label. The rights to Energize apparently were bought by Tartan, Inc.

External links

* [http://dreamsongs.com/NewFiles/PatternsOfSoftware.pdf "Patterns of Software"] - a collection of essays by Gabriel, including some memoirs about Lucid (pdf)
* [http://www.dreamsongs.com/DiBona-OReillyLetter.html "Letter to Chris DiBona and Tim O'Reilly"] —(Open letter by Gabriel written about the XEmacs fork in response to the description of the fork in [http://www.oreilly.com/catalog/opensources/book/toc.html] )