KARL: a Victim of Europe’s Missing Innovation Culture?

(updated version: December 1, 2013)

Reiner Hartenstein, IEEE fellow, TU Kaiserslautern, http://hartenstein.de

Abstract. EDA (Electronic Design Automation) massively stimulates not only the growth of the microelectronics industry, the computer industry and other high-tech industry branches, but also massively stimulates the growth of the entire economy. Also Hardware Description Languages are very important for the efficiency of EDA tools and systems used for the design of computer systems and other microelectronic components. However, it was typic for the (missing) European innovation culture that as soon as an innovative area is becoming a trend, unerringly at that time research funding will be finished. The EDA industry, the powerful key to massive economic growth, disappeared from Europe. This EU innovation phobic syndrome also destroyed the success story of KARL.

EDA (Electronic Design Automation) massively stimulates not only the growth of the micro-electronics industry, the computer industry and other high-tech industry branches, but also massively stimulates the growth of the entire economy (see fig. 1, prepared with data from 1997 [1, 2]). Also Hardware Description Languages (HDL) are very important for the efficiency of EDA tools and systems used for the design of computer systems and other microelectronic components. During the 80ies when a massive commercial break-through of HDLs did not yet happen, KARL has been the most successful language at the time with 93 site licenses worldwide [3] and has been, together with its interactive graphic interface language ABL [4 ], the backbone of he first complete microelectronics EDA framework worldwide [ 5 ] (fig. 2– for acronyms see fig. 7). KARL has been interfaced to a lot of other software [ 7, 8], created a lot of literature [9, 10] and has a good quotation index [11]. We offered a lot of support to our users outlined by material about its subsystem implementations [12] and organized user group meetings [ 13, 14]. Within the projects CVT [ 15 ] (involved industry: British Telecom, Daimler-Benz, Italtel, MATRA-HMS, Thomson) and CVS [ 16] (involved AEG-Telefunken, ALCATEL, DECNET, Honeywell-Bull, Italtel, Olivetti, SEL, SGS, Thomson) this effort has been massively funded by the ESPRIT program of the European Union.

KARL is still the only fully calculus-capable HDL [29]. This has been demonstrated already in 1978 by the design of a multiplier floorplan fully automatically derived from the math formula [ 28]. Still in the year 2001 this was worldwide the only example of top-down EDA usage of Term Rewriting Systems (TRS) as has been confirmed by a leading TRS expert [ 30].

However, at that time it was typical of the (missing) European innovation culture that as soon as an innovative area is becoming a trend, unerringly at that time research funding is finished (also see fig. 3 [ 1 ]). The EDA industry, the powerful key to massive economic growth (fig. 1), almost completely disappeared from Europe (fig. 6 [25]).  This EU innovation phobic syndrome also contributed to destroying the success story of KARL. In a meeting at the 13th ESPRIT CAVE Workshop at May 1989 in Nice, France, officers of the commission of the European Union decided, after having spent for CVT and CVS more than 85 million ECU (~ €), to stop research funding on hardware description languages: "since now there is standard: VHDL". However, in 1989 there was published only a preliminary draft dealing with a partial implementation [ 6 ]. In fact, there was not any standard. standardiz-ation was just proposed by a DoD-supported aggressive VHDL lobby. Did this lobby influence ESPRIT policy via secret NATO channels?

Organized by IFIP working group 10.5 this was a highly successful conference series rather effectively covering research and application on all kinds of hardware description languages and their applications: "The International Conference on Computer Hardware Description Languages and their Applications (CHDL) [ 22 ]. The variety of languages described visible also here has been commented by Richard Newton as the "language of the year effect" (fig. 4). Here Richard Newton asked, what will be the next language after VHDL, as indicated by the question marks at the rightmost curve. After CHDL’97 [ 23 ] this very important CHDL international conference series was killed by the DoD-supported VHDL lobby (also see [ 27 ]). The IFIP working group had assigned the organization of the next CHDL to colleagues in the US: to be held in June 1999 at the University of Virginia, Charlottesville, VA, USA, to be organized by Ron Waxman, Phil Wilsey, P. Subrahmanyam, and Hillel Ofek. However, the minutes of an IFIP WG meeting tell us that "several members of the IFIP WG 10.5 indicated that although CHDL’99 and VLSI’99 are scheduled to be held in the US, there are currently no volunteers who have stepped forward to organize these events in the US" [ 24]. It is quite incredible, that a major US university is unable to organize an international conference. Has this been arranged by the VHDL lobby to remove competitor languages away from the window? Killing the CHDL conference series also removed the platform for the next "language of the year" (fig. 5). All current and future competitors of VHDL lost a very important platform. Prof. Graham Hellestrand from the University of new South Wales in Australia failed by proposing to organize CHDL-2001 in Sydney, Australia. The IFIP WG 10.5 did not support this proposal [ 24] nor start another effort. Why not? This happened probably because of a conflict of interest by the IFIP WG 10.5 chair at that time, who had just founded ECSI for running another HDL-related conference series (FDL). The hardare description language KARL, having been the dominant design language during the 80ies, is a Pascal dialect, what contributed to its popularity. One reason of its later decline has been the fading away of the programming language Pascal, being part of the alarming development as criticised by Fred P. Brooks (see figure 8).

VHDL has been criticized a lot by several VIPs. For instance, Jo Costello, at that time CEO of the EDA giant CADENCE, said in a keynote, that the introduction of VHDL has caused damage by several hundreds of millions of Dollars to industry. Also in a keynote Richard Newton called VHDL „the nroff of EDA". (a very inconvenient historic old word processing language, used a long time before graphical user interfaces came up.) In contrast to KARL, VHDL is not calculus-capable and its abstraction levels do not reach down to floor plan layout level. There are indications, that the very, very weak FPGA programmer productivity is caused by VHDL use. Maybe the real funding shut down motivation has been, that the EU commission officers have been afraid since VHDL was going to be massively pushed by the DoD in the USA. At least in the EDA scene it seemed to be a typical European behavior, to throw away the tools, as soon as competition shows up (fig. 3 [ 3]). Meanwhile the EDA industry is almost completely located in the USA, and Europe just provides some of its users.

A few years ago the commission of the European Union started again looking at innovation [17]. Not only in Europe, but also worldwide we urgently need an efficient innovation culture to avoid the breakdown of our massively computer-controlled world. Because of the von Neumann syndrome [ 18, 19] the electricity bill to run all these computers may become unaffordable, maybe, already within a decade. The 25 year old technology to reduce the energy needed for computing by several orders of magnitude is available. However, a huge effort will be required to implement it almost everywhere. We have to reinvent computing [ 20, 26]. A massively deep going challenge has to be mastered here, so that a widely spread innovation awareness has to be present via support by the mass media (see [ 21], sorry, in German language).


References

[1] Erich Barke: Brauchen wir eine europäische EDA-Industrie ? 1997, GMM-Studie; see [ 2]

[2] http://www.vde.com/de/InfoCenter/Archiv/Fachbeitraege/GMM/Documents/MCMS/edastudie1.pdf

[3] http://www.fpl.uni-kl.de/staff/hartenstein/KARLUsers.html#karl_lic

[4] http://www.opengrey.eu/item/display/10068/149439

[5] http://xputers.informatik.uni-kl.de/karl/karl_history_fbi.html#fig33

[6] http://helios.informatik.uni-kl.de/karl/KARLindex.html#karl_lic

[7] http://helios.informatik.uni-kl.de/karl/KARLindex.html#rtcode_interfaced

[8] http://helios.informatik.uni-kl.de/karl/KARLindex.html#implmented

[9] http://helios.informatik.uni-kl.de/karl/KARLindex.html#studies

[10] http://helios.informatik.uni-kl.de/karl/KARLindex.html#related

[11] http://helios.informatik.uni-kl.de/karl/KARLindex.html#karl_quot

[12] http://hartenstein.de/KARL/

[13] http://hartenstein.de/KARL/ABAKUS-Passau.html

[14] http://hartenstein.de/KARL/ABAKUS-Igls.html

[15] http://hartenstein.de/KARL/CORDIS-Search-CVT.htm

[16] http://hartenstein.de/KARL/CORDIS-Search-CVS.htm

[17] http://ec.europa.eu/enterprise/articles/innovation/

[18] http://helios.informatik.uni-kl.de/staff/hartenstein/Hartenstein-Delft-Sep2007.pdf

[19] http://hartenstein.de/CV/vNsydrome.htm

[20] http://www.inf.pucminas.br/sbc2010/anais/pdf/semish/st03_02.pdf

[21] http://hartenstein.de/innovation/

[22] http://hartenstein.de/CHDL/

[23] http://www.it.uc3m.es/~ifip/chdl97/

[24] http://tima.imag.fr/ifip/wg10-5/Minutes/minutes11-96.pdf

[25] CAD-Info 1-2/1999, Hanser Verlag 1999

[26] http://hartenstein.de/EIS2/

[27] http://hartenstein.de/CHDL/#1999

[28] http://www.fpl.uni-kl.de/staff/hartenstein/lot/TR-Structured-Design.pdf

[29] http://www.fpl.uni-kl.de/staff/hartenstein/Design_Algebra.pdf

[30] http://hartenstein.de/TR/