From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: * X-Spam-Status: No, score=1.3 required=5.0 tests=BAYES_00,INVALID_MSGID, MSGID_RANDY autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 103376,db9b06af44b7c91c X-Google-Attributes: gid103376,public From: Robert Dewar Subject: Re: ACE functionality in Ada Date: 1999/05/16 Message-ID: <7hmapp$j3h$1@nnrp1.deja.com>#1/1 X-Deja-AN: 478429799 References: <7heqc9$bj8$1@nnrp1.deja.com> <7heuhk$f1m$1@nnrp1.deja.com> <7hf84v$nqj$1@nnrp1.deja.com> <373D89CE.C09BCDB2@home.com> X-Http-Proxy: 1.0 x34.deja.com:80 (Squid/1.1.22) for client 205.232.38.14 Organization: Deja.com - Share what you know. Learn what you don't. X-Article-Creation-Date: Sun May 16 11:39:37 1999 GMT Newsgroups: comp.lang.ada X-Http-User-Agent: Mozilla/4.04 [en] (OS/2; I) Date: 1999-05-16T00:00:00+00:00 List-Id: In article <373D89CE.C09BCDB2@home.com>, bbardin@home.com wrote: > The main thing that is required for distributed use in a > heterogeneous network is to make the message format > architecturally neutral (and the same for every node). One > way to achieve that is to use a textual form, which is what > GLADE does. The compiler must generate the same format > for the messages regardless of the architecture of the target, > which is achieved by using the same compiler (gnat) for each > target. Several points. First, yes, the GLADE/GNAT combination fully supports heterogenous distribution. No, this feature is not required by the RM. It is fair to say that the RM was designed primarily with the intention of supporting homogenous distribution, but in practice we have found no respect in which the definition inteferes with full heterogenous distribution. There are some slight difficulties that appear if multiple highly incompatible versions of Standard are around (e.g. an Integer in one implementation may be out of range of an Integer in another implementation), but by following the (in any case desirable) approach of avoiding implementation dependent types in Standard, this problem is eliminated. As for "using a text format", that's a bit misleading. The only requirement is that an architecture neutral format be used for streams. In the case of GNAT/GLADE, for heterogenous distribution, the format used is XDR (which is an international standard for architecture neutral representation of internal computer data). As for "using the same compiler (gnat) for each target", it is not at all clear that this is necessary in theory, since it is quite possible in practice for another compiler to be compatible with the XDR formats used by GNAT. Of course in real life, so far, the only implementation that supports Annex E is GNAT, so using any compiler that supports Annex E and using GNAT mean the same thing at the moment! Robert Dewar Ada Core Technologies --== Sent via Deja.com http://www.deja.com/ ==-- ---Share what you know. Learn what you don't.---