RCONN@SIMTEL20.ARPA (Rick Conn) (03/08/85)
I received a flyer from and made a phone call to General Transformation Corporation, PO Box 10083, Berkeley, CA 94709 (415)644-0702 today re their new Ada compiler. The data: Validation -- planned to take place as follows: pre-validation in 2nd quarter of 1985, validation in 3rd quarter of 1985 Host System -- IBM PC/XT (later will shoot for the IBM PC/AT) Target System -- Native object code for the Intel 8086/8088 micro with 8087 numeric coprocessor running PC-DOS 2.0 or later Availability -- soon after pre-validation, during 3rd quarter of 1985; cost is $1000 Host System Requirements -- IBM PC/XT with at least 5M bytes of disk, monitor, 360K byte floppy, 512K bytes RAM, 8087 numeric coprocessor My comments -- (1) They claim that current tests show that compilation speeds of at least 1000 lines per minute (elapsed time) will be achieved; I will have to see that to believe it, but, for a fully validated compiler, I wouldn't mind settling for 100 lines per minute (2) They will include rep specs! The marketing manager hinted that they plan to include ALL of Chapter 13 of the Ada Language Reference Manual! They will also include the pragma interface for interfacing with other languages. (3) I got the general impression that they were quite competent (I had a good discussion with the sales manager). They were also quite responsive to my call. (4) Will be looking forward to thee announcement of their validation. Will then probably push for acquisition of a copy for evaluation here. If this is all that it claims to be, it will be a marvelous step forward for the Ada effort overall. Imagine offloading compilations to a PC, and then uploading the debugged code to a mainframe for compilation into the main system. Rick The opinions expressed herein are my own and not necessarily those of my employer. I am not aware of any affiliation that I may have with General Transformation Corporation. -------
Stachour@HI-MULTICS.ARPA (Paul D. Stachour) (03/08/85)
Rick (and others):
--> In reguard to offloading to a PC or elsewhere <--
Remember that offloading compilations to another machine will not be
that easy in Ada. It's the separate compilation and library access that
will have to be 'kept-in-sync' also.
Ada is "more than a language" ("Ada, more than a language" is one of
the talks that I give). The library managers available on most systems
are inadequate to support good development, only suitable for
single-person or very-small-team efforts. To do this "right" your
off-loaded-compiler will need access to the "real-library", and most
compilers will not, and cannot, due to their implementation, support
same. The DDC (Dansk Datamatic Center) family of compilers is one
family that will be able, because of their design, to support same; I
certainly hope that there will be others. The DDC family does not yet
support this function, nor has DDC promised that they well, but the
potential is there.
I have a biased interest in the DDC family because the company I
work for, Honeywell, is doing its Ada compilers jointly with DDC.
...PaulRCONN@SIMTEL20.ARPA (Rick Conn) (03/08/85)
Paul, Your point is well-taken. For very large projects, especially when a good CM system or library management system is not available to keep the libraries in sync, offloading may not be feasible. However, I continue to feel that there are many situations in which offloading (at least during unit coding and testing) is quite feasible. I am just now completing involvement in a spelling checker program, which is a small- to medium-sized effort. The design was oriented around two major objects: the DOCUMENT to be checked and the DICTIONARY (there can be more than one) to check against. There are two packages which deal with these objects, DOCUMENT_HANDLER and DICTIONARY_MANAGER, and each package was assigned to one person to code and unit test. The specifications, which, of course, were subject to change, were planned during the design phase, and the development (including coding and unit testing) of the bodies was done in complete independence. Such development could easily have been offloaded to a PC if an Ada compiler existed for one. The only requirement for this would be that the MACHINE_DEPENDENCIES packages be duplicated on each PC. Each package was in complete control of its object; the DOCUMENT_HANDLER was given the name of a file containing a document and provided words and context information to its user. The user was never concerned with details of document structure or manipulation. To illustrate, MICRO:<ADA.EDUCATION> contains the spec to DOCUMENT_HANDLER as the file DH.ADA. Integration of these and other independently-written packages is now taking place with a high degree of success. I feel that this object-oriented mindset is key to this success, and Ada is instrumental in supporting it. Rick -------