rfg@ics.uci.edu (Ronald Guilmette) (10/03/89)
In article <6661@hubcap.clemson.edu> billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu writes: > > The US Government is actively planning to maximize the use of > automatic translation technology during the transition from Ada > 83 to Ada 9X. > Holy smokes!!!! Is it just me folks, or does the statement above imply that (a) Ada 9X has already been designed, and (b) it *does not* provide upward compatibility for Ada 83 programs? Obviously, just following this newsgroup is not enough to keep ones self "in the know" regarding the current state of the Ada 9X deliberations! Somebody please reduce my ignorance level and tell me what parts of Ada 83 have already been declared to be obsolete. Tasking? Fixed-point? Could somebody maybe post a list? // rfg
billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu (William Thomas Wolfe, 2847 ) (10/04/89)
From rfg@ics.uci.edu (Ronald Guilmette): >> The US Government is actively planning to maximize the use of >> automatic translation technology during the transition from Ada >> 83 to Ada 9X. > > Holy smokes!!!! Is it just me folks, or does the statement above imply that > (a) Ada 9X has already been designed, and (b) it *does not* provide upward > compatibility for Ada 83 programs? It means only what it says: whatever 9X's final form, the use of automatic translation technology will probably be maximized. This probably will take the form of an automatic translator being built to Government specs and placed in the Ada Software Repository. I would think that there would be enough similarity to make the use of automatic translation reasonable; this technology has proven not to be exceptionally useful where remarkably dissimilar languages are involved, such as in a Fortran-to-Ada conversion. Bill Wolfe, wtwolfe@hubcap.clemson.edu
rfg@ics.uci.edu (Ronald Guilmette) (10/04/89)
In article <6667@hubcap.clemson.edu> billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu writes: >> >> Holy smokes!!!! Is it just me folks, or does the statement above imply that >> (a) Ada 9X has already been designed, and (b) it *does not* provide upward >> compatibility for Ada 83 programs? > > It means only what it says: whatever 9X's final form, the use > of automatic translation technology will probably be maximized. > This probably will take the form of an automatic translator being > built to Government specs and placed in the Ada Software Repository. > > I would think that there would be enough similarity to make the > use of automatic translation reasonable; ... I think that you missed my point entirely. I have to assume that there is a large base of Ada 83 users out there who hope and pray that "the use of automatic translation" would *not* be "reasonable", but would instead be TOTALLY UNNECESSARY when moving to Ada 9X. I will ask one more time and hope for a more direct answer. Has it already been decided that Ada 83 and Ada 9X will be sufficiently incompatible so as to *require* translation? (Hint: this is a yes-or-no question.) If the answer is no, then why is the government planning on building/using a translator (or translators) when the need for such tools & processes is not yet even established? Has Samuel Pierce moved over to DoD from HUD, or is this just the 1990's version of the $600 screwdriver? If the answer is yes, then it *must* logically follow that *somebody* knows what the incompatibilities are. Otherwise, how could anyone know that automatic translation will be required (or even useful). If so, that person (or persons) are doing a disservice to the Ada community by not comming forward to warn Ada 83 users about features to avoid from now on. Could it perhaps be the case that the individuals who know what the incompatibilities are (ahem, I mean what they "will be") are keeping it to themselves in the hope of later capitalizing on this "insider information"? Either way, something here smells like three-day-old fish. // rfg
ryer@inmet (10/05/89)
RE: Ada 83/9X Compatibility Right now there is a "Requirements Team" working out of IDA including some government and some industry people. They are analyzing the change requests that have been submitted, the results of various public workshops, the history of waivers granted by DOD to not use Ada, etc, and will come out with a set of *requirements* for new Ada. For example, they may and may not make the statement "Full support for object-oriented programming, at least at the functionality level of C++ must be added". They are not expected to say do (or do not) add "classes" as a new type, or change type derivation as follows. There is a contract to be competed this winter called the Mapping/Revision contract. The winner will take the 9X *requirements* and design a language (set of language changes) to meet them. As far as I know, "upward compatibility" has not been established as a firm requirement for that contract, but no competent language designer is going to make incompatible changes without an excellent reason and a lot of squirming around to avoid it. It wouldn't surprise me if someone is already building an automatic translator. It is a time-honored custom to begin implementing the solution before the requirements have been defined. Is the translator to be old Ada to new (perhaps it deduces and inserts classes to make programs better), or new Ada to old (to take advantage of un-upgraded compilers)? Mike Ryer Intermetrics
billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu (William Thomas Wolfe, 2847 ) (10/05/89)
From rfg@ics.uci.edu (Ronald Guilmette): > I will ask one more time and hope for a more direct answer. Has it already > been decided that Ada 83 and Ada 9X will be sufficiently incompatible so > as to *require* translation? (Hint: this is a yes-or-no question.) I do not know as a fact that this is the case, but I do recall reading about plans for automatic 83 => 9X translation. Perhaps someone having more direct knowledge of the situation could comment in greater detail. I would expect that there *would* be sufficient incompatibility, just on the basis of the experiences with Ada 83 which have been documented in Ada Letters, etc., which indicate that certain changes are necessary. I would also expect that if a free 83 => 9X translator is provided via the Ada Software Repository, the conversion process would be dominated by the time required to upgrade programmers, rather than the time required to automatically upgrade existing Ada 83 software. If you want to get a good idea of what the likely changes will be: o ACM SIGADA Ada Letters for the last 5 years or so o Ada 9X revision requests o proceedings of Tri-Ada and other conferences o the last chapter of Paul Hilfinger's ACM Distinguished Dissertation, "Abstraction Mechanisms and Language Design" are all excellent sources. Bill Wolfe, wtwolfe@hubcap.clemson.edu
stt@inmet (10/05/89)
Here is the wording from the Ada9x project plan: A revision requirement that does not meet the "upward compatibility" criteria will only be considered if it affects a very large portion of the Ada community and its absence in the revised standard has a serious negative impact on application development. -- Ada 9X Project Requirements Development Plan, August 89, -- footnote on page 2 (and 3). S. Tucker Taft Intermetrics, Inc. Cambridge, MA 02138
jbg@sei.cmu.edu (John Goodenough) (10/06/89)
In article Re: Translating 83 => 9X (Was: Re: Ada 9X objectives) of 5 Oct 89 01:56:29 GMT billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu writes: > If you want to get a good idea of what the likely changes will be: ^^^^^^ > o ACM SIGADA Ada Letters for the last 5 years or so > o Ada 9X revision requests > o proceedings of Tri-Ada and other conferences > o the last chapter of Paul Hilfinger's ACM Distinguished > Dissertation, "Abstraction Mechanisms and Language Design" > > are all excellent sources. These sources BY NO MEANS specify the _likely_ changes in Ada 9X. These are sources of suggestions being made, but no one yet knows what the likely changes might be, because the change requests have not yet been analyzed against the stated goals of the 9X process, which are: "to revise ANSI/MIL-STD-1815A to reflect current essential requirements with minimum negative impact and maximum positive impact to the Ada community. The Ada 9X process is a revision and not a redesign of the language and should be viewed as a natural part of the language maturation process." [From the Ada 9X Project Plan, January 1989] In my opinion, this means the goal is to improve the usability of Ada, and this means fixing problems while not destabilizing the adoption process or the quality of Ada implementations. How this can be done will be the subject of much discussion in the next few years, but major and widespread incompatibilities certainly are not consistent with the stated goals. It is certainly premature to suggest that some kind of automatic translator will be required. In fact, I think most people involved with the effort would be appalled if something called an "automatic translator" from Ada 83 to Ada 9X was required. John B. Goodenough Goodenough@sei.cmu.edu Software Engineering Institute 412-268-6391 -- John B. Goodenough Goodenough@sei.cmu.edu Software Engineering Institute 412-268-6391
rfg@ics.uci.edu (Ronald Guilmette) (10/07/89)
In article <6678@hubcap.clemson.edu> billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu writes: > > I do not know as a fact that this is the case, but I do recall > reading about plans for automatic 83 => 9X translation. Perhaps > someone having more direct knowledge of the situation could comment > in greater detail. If you can find the reference, I'm sure that many here would be interested to read that same publication. // rfg
billwolf%hazel.cs.clemson.edu@hubcap.clemson.edu (William Thomas Wolfe, 2847 ) (10/08/89)
From rfg@ics.uci.edu (Ronald Guilmette): >> I do not know as a fact that this is the case, but I do recall >> reading about plans for automatic 83 => 9X translation. Perhaps >> someone having more direct knowledge of the situation could comment >> in greater detail. > > If you can find the reference, I'm sure that many here would be interested > to read that same publication. Unfortunately, at the time I considered it an interesting tidbit of information, but did not anticipate ever having to cite it. At any rate, we must also realize that translation is only one of the two major avenues of transition; the other is to exploit Ada's ability to call code written in other languages. If the manufacturers of Ada 83 compilers decided to sell an upgrade which would permit the use of pragma Interface to Ada 9X, and the 9X compilers also had the ability to interface to Ada 83, then the primary purpose of automatic translation would be to enable the use of the more powerful Ada 9X by the system's maintainers. Bill Wolfe, wtwolfe@hubcap.clemson.edu