From kh@athens.cs.waikato.ac.nz Tue Apr 27 21:27:02 1993 Received: from grace.waikato.ac.nz by dkuug.dk with SMTP id AA21951 (5.65c8/IDA-1.4.4j for ); Mon, 26 Apr 1993 23:24:59 +0200 Message-Id: <199304262124.AA21951@dkuug.dk> Received: from athens.cs.waikato.ac.nz by waikato.ac.nz; Tue, 27 Apr 93 09:24 +1200 Received: by athens.cs.waikato.ac.nz (16.6/16.2) id AA01569; Tue, 27 Apr 93 09:27:02 +1200 Date: Tue, 27 Apr 93 09:27:02 +1200 From: Keith Hopper Subject: The importance of LIS To: sc22wg15@dkuug.dk X-Charset: ASCII X-Char-Esc: 29 Greetings, I hope that I express the concern of many other countries, members of which are represented in WG15 when I say that any attempt to cancel or seriously postpone the LIS work will be seen as catastrophic! Yes! I have chosen the word with care. Let me explain the NZ reasoning behind this very strong statement since lack of funds will prevent me keeping up with John 'after hours' in Heidelberg! From a practical point of view software production in NZ is based around small firms and what might be called 'niche' markets so that we are not in direct opposition to the mighty weight of Oracle, Microsoft or Fred Bloggs! Several of these niche markets are specialist software which is to be made usable on a variety of platforms and, far more important is being written in a quite wide variety of languages -- Ada, Modula-2, Pascal, Lisp, Prolog, some private ones, together with the low-level stuff (C, C++ and assembler). A significant proportion of this material is written for real-time and embedded applications with the need to port not only entire applications, but also many functional libraries. It is this latter one which gives rise to one of our main concerns. Without any intention to 'lecture', the concern expressed is that there needs to be a traceable starting point for the POSIX specification which is not limited in its expressibility by any one binding (whether C, Ada, M-2 or ...) nor in its correctness or completeness by any implementation dependencies which have been necessarily introduced into the specification of the particular language concerned. The principles of traceability and derivability, so important in the testing of standards is seen also to be a necessary feature of thinking when defining such a fundamentally important matter as the environmental interface for the majority of the world's future software. In other words, we believe that the LIS is the foundation stone upon which all bindings and application implementations should be built so that we can port a library written using POSIX facilities in Ada and make use of it in conjunction with some other library written making use of POSIX facilities in Modula-2 (say) and put them together with other code written in Extended Pascal into a program which will work as intended! The only way we can see to doing this sensible is that the underlying services are defined in the LIS and all language bindings are provably (yes, provably!) derivable from that single specification. Just to show you that this is not intended to be a 'pie-in-the-sky' approach to the current resource problem, I have consulted with colleagues here and, depending upon how working is set up we estimate that a handful of workers here in NZ could and should contribute to an international effort at drafting LIS material, although for financial reasons the majority of the work and communication would have to be done using electronic mail. This suggestion is made as a subsidiary to the following proposal -- WG15 should offer to take over the technical work of LIS drafting and spread it throughout the world, operating on an e-mail basis to reduce costs and enable thinking to be done in those odd moments between rising and the phone ringing (or however you find time for extra work). We hope to be able to put actual names in the hat as a result of the next meeting of Uniforum NZ which will be held towards the end of May, when I am presenting to everyone the current state of the POSIX standards art. May I ask you to consider this plea for the continuance of LIS, perhaps using the revised drafting mechanism I suggest in listening to the latest state of the art from IEEE and considering the US position vis-a-vis the international requirements. Regards, Keith PS On a per capita basis three or four people offering from NZ is a bit like 60 or 70 from UK or two to three hundred from the US. The little figures do add up!