In copycat82, the section "3.2.5. System Structure." starts with a list of the tuple, the letters in the formal net-model, as copycat82 plagiarizes E-nets. I provide the letter-for-letter mappings to E-nets, as links to those pages, that I prove the plagiarism, about each of those letters. Please visit those pages, to witness the close-fit.
|copycat82/83||Proof of plagiarism|
|L||3.2.1.||S||location ==> "control state variable"|
|xi||3.2.2.||D [,T]||(token-attributes, environment variables) ==> (data-items[, and "data-types"])|
|Fi, in||peripheral locations ==> "data-independent control-transfer specification"|
|Fi, out = Fi||resolution procedure ==> "data-dependent control transfer specification"|
|R||3.2.4.||R||resolution location ==> "interconnection relation"|
|A||3.2.6.||C||[macro] transition ==> "software component"|
The last (the third) paragraph of the copycat82 section is after Da80 section "V. Global Model." I point out that, the two, or three, sentences at the start of Da80 section, define the "longer" paragraph in copycat82, but when lengthening the paragraph, copycat82 only commits absurdity, and limits itself to the Da80 range, although Da80 excludes verification aspects, as stated in its third sentence (in section V). Why would copycat82 observe the same limits, although its title includes the term "... analysis of distributed software systems ..." ?
The rest of the Da80 section V, and related content about hybrid models, is trackable, in copycat82, at the end of, section 3.2.3 (page 57), where copycat82 includes brief commentary about transition procedures, gi,j (renamed as "data-transfer specification" - whether any data gets "transferred" in it, or not), again a bit verbosely, in a way that refers to the possibilities of using if statements, while loops, etc., to compose program-modules. (This is at the end of the section that presents resolution-procedures.) In other words, "nothing is lost." Only shuffled. Strictly plagiarized.
That third paragraph in copycat82, section 3.2.5, starts with a cliam that the tuple, as listed on this page, called "SYSTEM," is "from the point of view of a global observer." (p.59) But it ends, with the opposite point, that is, "... will represent the internal structure of components rather than the global system view." (page 60).
In between, on pages 59-to-60, in its third paragraph, copycat82, presents a false-copy of what Da80 concisely discusses. (Keep in mind that, copycat82, never cites Da80, anywhere, though.) Da80 had written, in section "V. Global Model" (p.641):
If we have the model of the two peer entities of Fig.11, we still need a model of the transmission medium to be able to consider the complete end-to-end protocol. The nondeterministic aspet of the transmission medium model does not allow the use of the same modeling techniques and this raises additional difficulties. As this is related to the verification aspects, it will not be treated in this paper.
The copycat82 (page 59), starting with the "However ..." is the part. Where Da80 refers to the nondeterminism, copycat82 guesses/ignores it to be only something about the timings, around the global system. But, there are a lot more problems in a distributed system. It does not have to be a "network-layer" protocol. For example, availability issues, present nondeterminism, too. In a lot of applications, this is an important issue. If the network is not fully-connected, whether intermittently, or permanently, this is certainly to influence the software, too.
Also, for example, the issues about data-migration, and/or load-balancing, especially if distributed-data is not managed transparently to that application, would present other problems, too. These present difficulties, at the application, and operating systems, levels, and introduce confusion, if you would strictly concentrate on "token-passing to node x2, for event e2."
As such, the insufficient-replacement in copycat82 that refers to the only problem of, inconsistent-global-view, by any node, "since it reflects the status of the system considerably before the current time" (pages 59-60), is obviously not an improvement in content. It loses meaning, when it is abridging the term, nondeterminism, that Da80 uses.
Not to mention that the word "considerably" is rather vague, and not appropriate, because, elapsed-time is a relative issue. We cannot tell whether five seconds is "considerable time" for a module, more considerably than five days is for another module. Not to mention that, if the protocol is designed to ensure correct operation, without assumptions about time, then time is never an issue. Even in a single-task software, data-access semantics could introduce multiple versions of a data-item, if that is the issue. That is not about time. It is about multiple copies. For example, if the called function, receives the parameter value in, and it restores out, at exit.