Talk:Computer program

Recommendations

(Note: I recently reviewed this article at GAN.) @Timhowardriley: Thought I should put some more comments in addition to the ones I already made, considering you're renominating it immediately, also for a reviewer to consider:

• "The interpreted language Basic (1964) has had the most impact on programming" – Really? I don't have the source but... it's strange to state it so unequivocally. Why not C, essentially the lingua franca of the programming world?

Yes, according to Dr. Stair. I added his quote to the citation. Dr. Stair's assertion resonates with me because without interpreted Basic, I wouldn't have become a programmer. I'd rank C at #2. Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

But that is one opinion, and probably not an opinion shared by most programmers. The history of programming languages is sufficiently complex that it's probably impossible to single out one as the most influential. What about Lisp, C++, even Java? Ovinus (talk) 14:54, 9 June 2022 (UTC)

Most importantly, it's sourced. All unsourced arguments are original research. Another published author may state that language X has had the most impact. Well, then we have something to resolve. Timhowardriley (talk) 18:52, 9 June 2022 (UTC)

Common sense, fair presentation, and editorial discretion are also important. This book on the history of languages says for FORTRAN, "The importance of FORTRAN is sufficiently obvious not to need further justification here", while, of BASIC, "This language is widely used, and in many ways is the simplest of all the major languages in use today. It has wide educational use, and is a 'first language' for many students." (page xix). To put it bluntly: No, we should not state in wiki voice that BASIC is the most influential language. That is an opinion and definitely undue weight on one author's opinion. Ovinus (talk) 19:58, 9 June 2022 (UTC)

Fair enough. I'll research a better thesis statement. Timhowardriley (talk) 20:13, 9 June 2022 (UTC)

• The given BASIC program is not a Hello World program.

It is the "Hello World" program of this article. Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

But it is not a "Hello World" program... Ovinus (talk) 14:54, 9 June 2022 (UTC)

• Just in case you don't know, MOS:PSEUDOCODE is an important guideline. For articles like these it's reasonable to have exceptions, including code samples, but that does not mean that they should be indiscriminate collections of code (e.g., the contents of "Object-oriented programming").
• Microcode is... I suppose it's technically "programming", but really it's much closer to architecture design. I don't know much about this area, but I imagine that (say, Intel) CPU designers work on microcode concurrently with implementing the underlying uops manifest in hardware units. In any case, it deserves less attention.
• Also, if mentioned at all, microcode needs a better description. In particular, it's important to know that it's much slower than non-microcoded instructions. Perhaps an illustrative example is, say, 64-bit integer multiplication vs integer division (the latter is microcoded and potentially a whopping 99 cycles on, say, Haswell [1], vs. probably only a handful of cycles for multiplication).
• "At the hardware level, a microcode program controls the circuits throughout the motherboard." and "Microcode instructions move data between a CPU's registers and throughout the motherboard." Not true; the motherboard and its interactions with the CPU are controlled by firmware. Microcode is specific to the CPU.

Thank you. I fixed one assertion and removed the other sentence from the article. Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

• I just want to say that the BASIC program is much, much better for readers than the C program and its associated baubles. Even newbie programmers will not understand the C code, but the BASIC program is highly readable. I encourage you to use snippets of that type.
• The C code is... idiosyncratic. Why should the Person's grade be heap-allocated and wrapped in a class rather than simply a char, or a struct containing const char*, etc? Same with the Person struct, too. C is not Java, and there are better simple examples of encapsulation and/or manual memory management that actually makes sense design-wise. As an aside, neither the person nor their grade is freed.

The Person's grade is heap-allocated and wrapped in a class because that's the essence of object-oriented programming. Maximizing the CPU's efficiency isn't the objective. CPU cycles are relatively cheap compared with the human labor of developing and maintaining the computer programs. Regarding freeing the memory: yes, it wouldn't be distracting to add free(). But I assume the operating system performs garbage collection. Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

C is simply not an object-oriented programming language, and it does not have classes. I'm making no claim regarding efficiency, although you're right, heap allocation is slow. My point is that it is an anti-pattern to default to heap allocation. I can't speak for C programmers, but many new C++ programmers have a habit of allocating everything using new, which is annoying to maintain. Ovinus (talk) 14:54, 9 June 2022 (UTC)

Object-oriented programming is a method of programming. The method is to write operations on objects. Objects in an OOP language are called classes; objects in a non-OOP language are called abstract data-types. It doesn't require an object-oriented language to implement OOP. True, a major programming annoyance is managing memory. That's why it important to clearly understand which area of memory each variable is located. The areas are text (where constants are stored), register, static, stack, and heap. Object-oriented objects are always stored in the heap. When using non-OOP languages, the programmer is forced to know this. When using a non-OOP language, the compiler hides from the programmer that heap memory is used. Anyway, the computer program article forks to the example sections in two other articles -- Imperative_programming#Examples and Declarative_programming#Examples. This section describes memory management and is intended to be a part of computer program: Imperative_programming#C. Timhowardriley (talk) 19:13, 9 June 2022 (UTC)

But then why use C as an example of object-oriented programming, instead of an object-oriented language? It is misleading at best. Ovinus (talk) 19:58, 9 June 2022 (UTC)

Because the thesis statement is "Object-oriented programming need not be confined to an object-oriented language." Timhowardriley (talk) 20:12, 9 June 2022 (UTC)

Er... that should not be the thesis statement. (Indeed, Wikipedia articles don't really have a "thesis statement".) That's like saying, in an article about screws, bolts and nails, "Screws need not be used with a screwdriver; a hammer may also be used." The C code you are writing is an example of bad practice. Ovinus (talk) 16:48, 10 June 2022 (UTC)

Object-oriented programming is a programming method, not a language. The C code in this example is a parallel algorithm with the Imperative_programming#C++ example. The C++ example is object-oriented programming in an object-oriented language. The C example is object-oriented programming in a language that predates object-oriented languages. Independent of my use of the words "thesis statement", it's rational to illustrate an example of an object-oriented program using C. If you juxtapose the two examples, you will glean insight into the mechanics of an object-oriented compiler. Timhowardriley (talk) 00:45, 11 June 2022 (UTC)

Hm. There's a reasonable discussion to be had on how much the article should emphasize the broad categories of 1. program execution; 2. programs as essentially abstract devices (detached from a particular computer or compiler); and 3. what average programming looks like. I think this part you've written is quite heavy on (1), when people who have never heard of object-oriented programming need (3) and maybe (2). Anyway, the fact remains that there's no need to have C anti-patterns in the article Computer program, along with its lovely include guards and object files and linking procedures, when we could have, say, a straightforward Python object (perhaps a Person and Grade object, just as you've done). Ovinus (talk) 04:42, 11 June 2022 (UTC)

• "Application software is the key to unlocking the potential of the computer system." – This isn't true for, say, little microcontrollers in a microwave

I added Dr. Stair's quote to the citation. A microwave isn't a computer system? Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

Hm. Even if you deny that the chip in a microwave is a computer (personally I'd classify it as one), it's still reasonably common to run programs entirely in kernelspace. Ovinus (talk) 14:54, 9 June 2022 (UTC)

I was snarky. Sorry. Timhowardriley (talk) 19:14, 9 June 2022 (UTC)

You're fine? We're having a good discussion. Ovinus (talk) 19:58, 9 June 2022 (UTC)

Thank you for contributing your critical thinking. I also appreciate your open mind. Since 1989, I worked in the metaphorical trenches programming land mines with hairpin triggers. Researching for this article has raised my environment to the clouds programming art. Timhowardriley (talk) 22:49, 9 June 2022 (UTC)

• "Labels allow the programmer to work with variable names." – Sure, but labels are also places to jump to

Ovinus (talk) 12:20, 9 June 2022 (UTC)

Thank you for your time and feedback. Timhowardriley (talk) 14:45, 9 June 2022 (UTC)

• "Functional programming is executing functions just-in-time (JIT)" – What does this mean? Functional programming, among other things, is about treating functions as first class citizens, which the snippet of code doesn't do. Also, if you're not executing a function just in time then... are you evaluating it lazily? I'm confused. JIT compilation is a totally different beast than anything that can be reasonably shown in C. Ovinus (talk) 16:24, 11 June 2022 (UTC)

An operator isn't a statement

This edit qualified this sentence with "may": "Unlike assembly, these languages may[c] generate many machine language instructions for each statement." I added note [c] to say "C programming language statements like x++ will compile to a single instruction" because that was the claim in the edit reason. However, ++ are two symbols that together form an Operation (mathematics), not a statement. "Principles of Information Systems" (page 159) says, "[A] programming language [is] a set of keywords, symbols, and a system of rules for constructing statements[.]" Statements are keywords plus symbols. The x++ instruction is a postfix, increment operator. See Operators_in_C_and_C++#Arithmetic_operators. Therefore, the qualifier isn't warranted. Timhowardriley (talk) 15:39, 11 June 2022 (UTC)

This is pedantic in the wrong direction. Indeed, the C language spec considers various things to be "statements", the most important one here being an expression statement. x++; is a statement that usually compiles to one instruction. y = x * 2 + 3; is a statement that usually compiles to one instruction. Conversely, statements may compile to no instructions at all; x += strlen("cow") - 3; will get optimized out. An unreachable return statement will lead to a ret being omitted. The "may" is important; languages and compilers are complex. Ovinus (talk) 16:21, 11 June 2022 (UTC)

Dr. Stair is not being pedantic by saying statements are keywords. Operators in C and C++ is not being pedantic by saying x++ is an operator. Furthermore, "Comparative Programming Languages" says (page 82), "Statements are the commands in a language which perform actions and change the state." Keywords and commands are the same thing. Most importantly, Dr. Stair says without qualification (page 160), "With third-generation and higher level programming languages, each statement in the language translates into several instructions in machine language." The article is paraphrasing this quote. Your assertion that x++ is a statement and it usually compiles to one instruction is original research until you can source it. Why did you qualify your assertion with "usually"? It's easier for the article's reader to go with Dr. Stair than to try to understand your usually exception. Timhowardriley (talk) 20:09, 11 June 2022 (UTC)

I changed the article's note to "C programming language operators like x++ will usually compile to a single instruction." The change reflects the new information from this thread. The article now conflates statement with operator. This is a mistake. Statements are keywords and operators are symbols. Timhowardriley (talk) 20:29, 11 June 2022 (UTC)

Taken as an absolute (which I don't think he intended), Dr. Stair is wrong. [2] has 12 statements and only a handful of machine instructions out. The text is misleading without the qualifier; the takeaway is that there is a weaker correspondence between high-level statements and what's actually executed. Often, yes, that means that a single statement will generate "several" or even "many" asm instructions. But often, that means a statement will generate a single instruction, or get optimized out entirely. That is critical information. Ex: [3] ("When you specify a higher level of optimization for a compiler, the compiler will often move code around in the assembly output file, eliminate code entirely, and do other code transformations that obfuscate the correspondence between the high-level code and the assembly output"). It is categorically false that every C statement generates many machine code instructions after optimization; if compilers did that, no one would use them. Ovinus (talk) 21:44, 11 June 2022 (UTC)

Regarding Taken as an absolute (which I don't think he intended): True, Dr. Stair wasn't pedantic.

Regarding Dr. Stair is wrong.: Whew! Timhowardriley (talk) 22:15, 11 June 2022 (UTC)

Heh! Indeed those phrases shouldn't be divorced: I'm sure his writing is excellent and I haven't a hundredth his expertise. I also don't know the context for that sentence. It's just that the idea that "xyz languages turn my statements into many machine language instructions each" is not as accurate or informative (or even understandable, tbh, beyond a basic level) as "xyz languages have a weaker correspondence with the outputted machine code; high-level statements may be turned into several instructions, combined, or even entirely optimized out if doing so would not change the program's behavior". But maybe that's just me; I'd love to get a third opinion. Ovinus (talk) 22:57, 11 June 2022 (UTC)

I hope this edit fills in all the gaps. Timhowardriley (talk) 20:16, 12 June 2022 (UTC)

GA Review

This review is transcluded from Talk:Computer program/GA3. The edit link for this section can be used to add comments to the review.

Reviewer: Caleb Stanford (talk · contribs) 21:59, 25 June 2022 (UTC)

Starting review. Caleb Stanford (talk) 21:59, 25 June 2022 (UTC)

GA review (see here for what the criteria are, and here for what they are not)

---

The prose is well-written and the article surveys a lot of useful information. I haven't read this article before. This is also my first GA review so corrections/feedback appreciated. The main thing that would improve the article in my opinion is better organization, and clarity with respect to the different dimensions of computer programs considered. I will give some examples and details below.

1. It is reasonably well written.

   a (prose, spelling, and grammar): b (MoS for lead, layout, word choice, fiction, and lists):

2. It is factually accurate and verifiable.

   a (reference section): b (citations to reliable sources): c (OR): d (copyvio and plagiarism):

   The article is well-sourced throughout but the chosen reference style repeats the same references many times. For an article with this pattern of references, I would personally appreciate a different citation style that lists each reference only once: either the page number style (like [1]:25-32) or the short form style (Wilson, p. 394 with the full citation included only once in a separate list: Wilson, Leslie B. (2001). Comparative Programming Languages, Third Edition.). The actual references listed contain a relatively small number of books but that is not easily apparent from reading the reference list.

3. It is broad in its coverage.

   a (major aspects): b (focused):

   Some key concepts are not covered, such as the semantics of computer programs. Also, coverage is difficult to assess due to article organization (more comments on organization below)

4. It follows the neutral point of view policy.

   Fair representation without bias:

   The article seems to reflect an engineering and systems viewpoint and does not include a perspective from programming languages and programming languages theory. Viewpoints are generally listed fairly but I am not sure fair/due wait is given to each sub-concept and many concepts are missing. As with coverage, this is a bit hard to tell due to the organization of the article (more comments below). One particular example that I doubt is neutral is the delineation of a "5th generation programming language".

5. It is stable.

   No edit wars, etc.:

   No problem here.

6. It is illustrated by images and other media, where possible and appropriate.

   a (images are tagged and non-free content have non-free use rationales): b (appropriate use with suitable captions):

   Media/illustrations look great!

7. Overall:

   Pass/Fail:

   See comments below.

Although it isn't directly one of the GA criteria listed above, my biggest concern with the article overall is its organization. Reading through the article, I don't get a clear sense of which sections should contain which information, and I don't know that the current section choices reflect a clarity of thought with respect to different aspects or features of computer programs.

Specific comments, by section. These thoughts are mostly unedited, please let me know if any comments are unclear:

• Hello world program: missing some context on what a hello world program is (maybe a brief history) and how it relates to the key concept of a computer program. I'm not sure why the section is discussing concepts of variable declaration and use; that's specific PL syntax stuff, what we want to explain here is what a computer program is and how a Hello World program is used as an example of one to illustrate program syntax.
• The History section is generally very good, but the last subsection "Programming environment" seems out of place as written (also, lacks references). Is this intended to discuss a development in the history of programs to more advanced GUIs and IDEs? Or does this belong in a different place? Also what is the criteria for what programs are going to be listed in the History section, e.g. why are we not include a COBOL program, a C program, a C++ program, etc?
• The programming language section includes a view of PLs by "generation", which in my experience is not a common way to understand programming languages today and reads as obsolete. In particular, first through fourth generation programming languages seem like perfectly reasonable delineations, but the Fifth-generation programming language is a dubious concept that I don't think makes sense from a programming languages standpoint. This delineation needs some qualification. For example, who proposed this classification and when was it proposed, and how widely was it accepted?
• Why does the PL section include "Imperative Languages" and "Declarative languages" but "Object-oriented programming" is in a different place?

• I would suggest a different organization with a section on "programming paradigms", including imperative, declarative, OOP, Functional, statically typed, dynamically typed, and many other paradigms that are missing or not discussed currently.
• It could make sense for PL to be a separate section from programming paradigms.

• The article is missing a discussion of programming language syntax and semantics. The word ["semantics"](Semantics (computer science)) does not appear in the article.
• The section "Computer programming" also lacks some clarity on what should be included here. Why is the "waterfall model" discussed but no other systems development paradigms or engineering methodologies? Is this section intended to be an overview of the Software development process? Rather than try to rehash specific paradigms/processes like Waterfall, perhaps it should summarize what computer programming consists of in relation to computer programs, and what are key concepts such as process models, development/code organization strategy, software lifecycle, etc.
• Program modules: I am not sure why this is a top-level section. The word "module" can refer to sections of a program, or namespaces, or files. It does not necessarily have to do with software aging, but can be simply about software size and separation of concerns. Is this section about code organization more generally?
• Functional categories: I like this section.
• Other inclusions: The article mentions only a few programming languages and should at least mention, in a few sentences, different programming languages that exist and how computer programs differ in different languages. For example:

• the article doesn't contain the words "static" or "dynamic"
• the article never mentions "Python", "Java" except for a sentence in the intro, "Javascript", or many other popular languages.

• The article lacks a discussion of what counts as a computer program, not an easy concept. E.g. HTML? LaTeX? Have computer scientists discussed what counts as a program and how have different scientists agreed/disagreed on basic inclusions?

Minor comments:

• First sentence: I disagree that the difference between imperative and declarative programming is whether the instructions are a list or a set. This would only be true in certain languages, not in others. I would suggest: "In computer programming, a computer program is a set or sequence of instructions in a programming language that a computer can execute or interpret."
• Is "The Free Dictionary: computer program" a reliable source? Didn't find it listed at WP:Reliable sources/Perennial sources

• For a topic with a controversial definition like this, more work may be required to source the definition in the lead.

• Program execution box at the top: Why do notable compilers only include GCC, LLVM, and CLang
• "For examples of declarative languages, visit Declarative_programming#Examples." -- would be great to have a broad survey of a few examples, this can be pointed to with "see also" or another similar template rather than ad hoc text
• "Example packages include accounts payable and payroll" -- This is a non sequitur. Examples of fifth generation languages include specific applications for accounts payable and payroll? I am not sure what the writer had in mind. (But I don't think the entire concept of a 5GL is accepted today, see comments above.)
• Programming-in-the-small/programming-in-the-large: seems out of place, why is this discussed here and not e.g. under modules
• "Utility programs detect computer viruses." Comment may be out of place / unreferenced / unexplained

Note that I didn't look at the article history or if there were any previous reviews so these are basically raw/unedited thoughts.

Good luck! Best, Caleb Stanford (talk) 23:06, 25 June 2022 (UTC)

Regarding Why does the PL section include "Imperative Languages" and "Declarative languages" but "Object-oriented programming" is in a different place?: My research categorizes Object-oriented programming as a programming method, not a programming language. Timhowardriley (talk) 07:12, 26 June 2022 (UTC)

Regarding Is "The Free Dictionary: computer program" a reliable source? No, it isn't a reliable source. However, the definition does properly introduce the rest of the lead section. This article has a history of contentious definitions of computer program. The current definition and its source are the result of collaboration. Timhowardriley (talk) 08:07, 26 June 2022 (UTC)

Regarding Some key concepts are not covered, such as the semantics of computer programs.: See User:Timhowardriley/sandbox. Timhowardriley (talk) 08:43, 26 June 2022 (UTC)

Thanks for the edits so far! Several major improvements IMO. If you want help editing the article, let me know. Caleb Stanford (talk) 13:32, 26 June 2022 (UTC)

The article could use a section titled, "Syntax and semantics". The guiding definition I plan to follow is from Comparative Programming Languages by Wilson and Clark. It says, "The syntax (or grammar) of a programming language describes the correct form in which programs may be written while the semantics denotes the meaning that may be attached to the various syntactic constructs." Timhowardriley (talk) 16:21, 26 June 2022 (UTC)

Thanks! I think that's a good idea, basing on Wilson and Clark seems fine, and I can help with those sections once you put in a placeholder. Other than Wilson/Clark, I think we may want 1-2 other good textbook sources from a programming languages/compilers perspective to complement the already excellent existing references. I'm not going to edit the article without your direction because I'd rather not contest informal "ownership" over the article content but there are plenty of places I see room for improvement and I'll continue to monitor and post feedback here on the talk page. Caleb Stanford (talk) 16:49, 26 June 2022 (UTC)

Status query

Caleb Stanford, Timhowardriley, what is the status of this nomination/review? It looks like this review was never formally failed by Caleb Stanford, and that although Timhowardriley added a failure to the Article history template and immediately renominated the article, the review seems to be proceeding above. If there is work to be done, then continuing it in the auspices of a GA review on this page, where work to improve the article is a regular part of the reviewing process (unless there's too much wrong with the article to set right in a reasonable number of days). Note that if Caleb Stanford is planning to do significant work writing/editing sections of the article, a new reviewer will ultimately be needed for those sections. BlueMoonset (talk) 23:43, 26 June 2022 (UTC)

@BlueMoonset: my assessment is fail with suggested revisions, but I am new to the reviewing process so I guess I did not follow the right protocol to formally resolve it above? In the talk page, User:Timhowardriley updated it but that caused to list as "nominated" for a 4th round on the nominations list, which was incorrect so I tried to roll that back to list the 3rd round target instead of 4. (Re writing/editing -- I haven't done significant edits yet to the article but will keep that in mind.) Caleb Stanford (talk) 23:57, 26 June 2022 (UTC)

Caleb Stanford, the big question when you do a review and decide that the article doesn't meet the criteria, is whether the places it falls short can be rectified during the review process. It is expected that the majority of nominations will have something wrong, and sometimes a number of things. If the ways in which it falls short can reasonably be fixed within about seven days, the usual thing is to place the nomination on hold rather than to fail it outright, specifying on the review page what needs to be done to improve the article to the point where it meets the criteria. However, if the article's shortcomings are such that too much is wrong for it to be fixed in that period, failing it is the proper thing to do. Note that the last two GA nominations have been failed immediately because of significant issues (GA1 and GA2), so if that isn't the case here, then reopening this is a consideration. However, if this time significant issues remain, then it would be appropriate to fail the article as you did; what would not be appropriate in that case is an immediate renomination by Timhowardriley without addressing said significant issues first. BlueMoonset (talk) 00:16, 27 June 2022 (UTC)

Thanks for clarifying! And especially for the overview of how the process works. Despite my generally positive review above (and positive view of the article's potential), the more I look into it the more problems and quibbles I find, so I expect that revisions will take more than 7 days. For that reason, I will list as failed for now and remove the GAN template. Caleb Stanford (talk) 00:46, 27 June 2022 (UTC)

This is okay. I'll describe BNF in terms that even I can understand it. Then I'll renominate it. Timhowardriley (talk) 18:08, 27 June 2022 (UTC)

If possible, please let me know before re-nominating. I'd like to do some work on the article organization first and the article may not currently be stable. Lots of things out of place. OOP shouldn't be its own section, place it under "Programming Paradigms". The list of PLs with details and syntax is probably out of place; this isn't list of programming languages by type, it should instead provide an overview of how different languages relate to the key concept of a program. The article's strength right now is the History section (aside from the last part) and the section on "functional categories". Caleb Stanford (talk) 19:15, 27 June 2022 (UTC)

Hack at it. I'll only be in my sandbox. Thanks for your help. Timhowardriley (talk) 20:03, 27 June 2022 (UTC)

Definition

I think the first paragraph puts undue weight on the distinction between imperative and declarative. It's not relevant for the definition of a program, which is just some commands written in a language for a computer to execute. What is fundamental to a computer program is syntax and semantics, not the format that the commands are written in (which can be a list, a syntax tree, a set, or something else, as long as it is (at least in almost all cases) conforming to some Formal grammar.

For example, Datalog is a classic declarative language in which programs are written as sequences of commands. If I were describing the syntax of datalog, it would be a sequence. It's true that the semantics of Datalog doesn't depend on the order, but the distinction in the first paragraph is confusing. The definitions at Declarative programming and Imperative programming are much better quality and I would suggest those are used here, but I think the distinction should be removed entirely. Caleb Stanford (talk) 13:32, 26 June 2022 (UTC)

Forming a definition for this article has been as fun as chasing my own tail. Regarding Datalog is a classic declarative language in which programs are written as sequences of commands.: Datalog#Features,_limitations_and_extensions says, "Unlike in Prolog, statements of a Datalog program can be stated in any order." Timhowardriley (talk) 16:15, 26 June 2022 (UTC)

Thanks -- but I am aware of that (I referenced that in my post: "doesn't depend on the order") and my point is that it's not a set, it's a sequence that can be stated in any order. In other words, the syntax is a sequence, the semantics is a set. Also, the exact sentence you quote mentions Prolog as being order-dependent, which is another declarative language.

My contention is that the definition as currently stated is controversial, and it doesn't agree with definitions at either of the linked articles imperative programming and declarative programming. More importantly, the distinction between imperative and declarative is irrelevant for what a computer program is, violating WP:Undue and WP:Relevance. (Edit: though I think mentioning it somewhere in the intro is good, it shouldn't be part of the definition/scope.) By the way, programming languages is my area of research expertise, I'm not just making stuff up out of a hat. :) Caleb Stanford (talk) 16:26, 26 June 2022 (UTC)

Feel free to correct the definition. Also, that edit won't constitute the significant contribution necessary to preclude rating the article GA. Timhowardriley (talk) 16:55, 26 June 2022 (UTC)

OK, sounds good! Hope my message didn't come across as too strongly worded. Edited to add, I think mentioning declarative/imperative somewhere in the intro is good, it shouldn't be part of the definition/scope of the concept in the first sentences. I may make an edit at some point soon. Thanks Caleb Stanford (talk) 17:01, 26 June 2022 (UTC)

class=C

I think class=C is more appropriate for the article's present state. @Timhowardriley: I'd like to request that you stop re-nominating the article in its current form (as you have now done twice) without a careful pass by someone else. I am happy to do a pass at some point but the article needs, in my estimation, more than 7 days of work. Caleb Stanford (talk) 15:18, 28 June 2022 (UTC)

For completeness sake, here are the B-class criteria and why I think it fails currently.

1. The article is suitably referenced, with inline citations.

Agree

2. The article reasonably covers the topic, and does not contain obvious omissions or inaccuracies.

 Disagree The article coverage overlaps with, but does not exactly align with, what an overview of "computer program" should cover. Some ideas of what's not covered is in the suggested organization below.

3. The article has a defined structure.

 Disagree The current organization of the article is a bit of a mess, currently unstable, and lacks focus. An idea for how to improve it is below.

4. The article is reasonably well-written.

? Maybe Parts are well-written, others need editing and cleanup.

5. The article contains supporting materials where appropriate.

Agree For the most part.

6. The article presents its content in an appropriately understandable way.

? Maybe It's very hard to present a concept of such basic importance in an understandable way to a general audience -- I believe that the article is currently too technical, covering a technical overview of many important concepts, but should explain to a very basic level (e.g., to someone who has never written a computer program or never heard of what it is) what a computer program is.

Caleb Stanford (talk) 19:02, 28 June 2022 (UTC)

Request for comments on downgrade to Class=C

Was it appropriate to downgrade this article to Class=C on this edit? Timhowardriley (talk) 18:33, 28 June 2022 (UTC)

Why do you need a full-blown thirty-day formal WP:RFC for this? --Redrose64 🌹 (talk) 19:05, 28 June 2022 (UTC)

Article-class assessments aren't particularly formal, and the assignment of a letter which isn't even shown to the reader isn't important enough to warrant a site-wide RfC. Timhowardriley: The criticisms of the article are well-founded. Presenting such a broad topic in a broadly satisfactory way is a Herculean task, sure, but it can definitely be done better than how it currently is. Ovinus (talk) 19:31, 28 June 2022 (UTC)

• Classing it as "C" isn't unreasonable, a full RFC to discuss it probably is. -Kj cheetham (talk) 19:46, 28 June 2022 (UTC)
• Pinged into this discussion by yapperbot. After a quick skim, C-class seems reasonable to me. Suriname0 (talk) 02:38, 29 June 2022 (UTC)
• Invited by bot. Would recommend Timhowardriley to ask Caleb Stanford for the required improvements to this article, which i believe will be a lot more productive than this rfc. Cheers, Captain Jack Sparrow (talk) 05:15, 29 June 2022 (UTC)
• The change to C would have deserved a thread on the talk pageBefore the change, after the change, there is no need to have an RfC, you could have just made a thread. I got notified by the bot for this, and I don’t think it was worth the bot bothering. | Zippybonzo | Talk | 07:48, 29 June 2022 (UTC)

Suggested outline (2022-06-28)

Overall, the current article is well cited and has some good material, but its organization is a bit of a mess. It consists of a hodge podge of vaguely related topics, and the delineation between sections is unclear. Additionally, the article doesn't seem focused around "computer program" as a central theme. It contains several subsections now which elaborate on things of unclear relevance to computer programs -- for example the BNF discussion under "Syntax and semantics" uses linguistics as an example and is only tangentially relevant to what a "computer program" is. Recent edits have moved some things around and added material, and these have mostly been improvements I think, but there's still a ways to go.

Here is a draft suggested outline that I think would make sense. What do you think?

1. History (the earliest computer programs and overview of how the notion of a program has evolved)

2. Modern computer programs (survey of what a computer program means today)

This is a current major omission from the article. What I have in mind is illustrating through examples what today's computer programs are and look like. This section could incorporate much of the material currently under Imperative Languages and Declarative Languages as example programs.

The "hello world" section might also fit here. It's currently out of place but could instead be "examples" or something.

It would be great for this section to also include some empirical survey information such as: who writes programs today, what language(s) are they typically written in, etc.

3. Program elements

3.1 Syntax and Semantics

This is currently under "programming langauges and paradigms" but it doesn't quite fit there. It's more fundamental to what a program is.

3.2 ???

4. Programming languages and paradigms

5. Program development (current section titled software engineering & computer programming)

6. Classification of computer programs

4.1 By language

4.2 By platform (OS, architecture)

4.3 By use case (e.g., embedded systems programs, mobile phone apps, web apps, etc.)

4.4 By functional category

7. In popular culture (another section missing right now, could be nice to have)

Caleb Stanford (talk) 18:42, 28 June 2022 (UTC)

The current structure is just fine. Timhowardriley (talk) 19:02, 28 June 2022 (UTC)

I generally agree with Caleb Stanford's assessment. (Do you prefer Caleb, Mr. Stanford, something else?) This article should be mostly accessible to people with no programming experience, with occasional detail (perhaps in footnotes) to make it enjoyable for more-experienced folks. Your proposed scope and organization of the article makes sense and is highly navigable. There is just too much inappropriate content in the article at the moment, the most unfortunate being the sprawling C++ code and Makefile. Short snippets of code illustrating different levels of languages is best. I also agree with the emphasis on modern programming; as important as ALGOL, BASIC, etc., are, they are unrepresentative of what programmers do today, which again is very important for the layman. Ovinus (talk) 19:24, 28 June 2022 (UTC)

This article spends too much time covering things it shouldn't, I can agree on that much. In particular we devote way too much space to object oriented programming and technical details of C and C++ which should just be a brief summary and pointer to their own articles. I disagree that a modern definition of a computer program is much different from a historical one. We have other articles (lots of them! Software development, Computer programming, Software engineering) that can cover the process of programming or what specific tools and language variants are used today. This article should be general in scope - the parent of many child articles, in keeping with Wikipedia:Summary style. Sections on particular platforms and use cases don't really belong here. Also, no popular culture section. This article will not be helped by a collection of cruft and trivia. MrOllie (talk) 20:19, 28 June 2022 (UTC)

Good point about separate articles, thanks! Ovinus (talk) 21:14, 28 June 2022 (UTC)

Thanks, you both put it better than I did. ("Caleb" is fine!) Re: modern definition, my point is that the modern notion of a program is very different from, say, Ada Lovelace's note G or the universal Turing machine currently surveyed under History. Explaining what a program is today and what it was historically are two very different things. About classification, I think it would be nice to overview some answers to the question: what are some of the ways that programs can be categorized? Those subsections are just some ideas. Caleb Stanford (talk) 22:17, 28 June 2022 (UTC)

• The key sentence is, "Once the mechanics of basic computer programming are learned, more sophisticated and powerful languages are available to build large computer systems." Timhowardriley (talk) 22:15, 1 December 2022 (UTC)
• The most profound passage is, "Object-oriented programming developed by combining the need for containers and the need for safe functional programming. This programming method need not be confined to an object-oriented language." Timhowardriley (talk) 10:25, 5 December 2022 (UTC)
• The most difficult objective to achieve is, "To be effective, program modules need to be defined and distributed to team members. Also, team members must interact with one another in a meaningful and effective way." Timhowardriley (talk) 10:31, 5 December 2022 (UTC)
• The most insightful sentence is, "C allows the programmer to control which region of memory data is to be stored." Timhowardriley (talk) 11:38, 5 December 2022 (UTC)
• The "hodge podge" includes the sentence, "Software engineering is a variety of techniques to produce quality software." Timhowardriley (talk) 01:51, 12 February 2023 (UTC)
• The "modern notion of a program is very different from, say, Ada Lovelace's note G" because "[a]t each stage in hardware's history, the task of computer programming changed dramatically." Timhowardriley (talk) 09:33, 12 February 2023 (UTC)
• The "hello world" program in Basic at the beginning is is not "currently out of place" because "if a student didn't go on to a more powerful language, the student would still remember Basic." Timhowardriley (talk) 10:45, 22 February 2023 (UTC)
• The "most unfortunate [inappropriate content is] the sprawling C++ code" which illustrates how "[i]n the 1970s, software engineers needed language support to break large projects down into modules. One obvious feature was to decompose large projects physically into separate files. A less obvious feature was to decompose large projects logically into abstract datatypes." Timhowardriley (talk) 12:59, 6 March 2023 (UTC)

  @Timhowardriley: With kindness, can you please stop editing the talk page of this article with non-constructive edits? I understand you disagree with the assessment of the article, but it doesn't help to repeatedly edit the talk page in this way. I also apologize if you took offense to the tone of my comments or felt that any of them are not actionable. My intention was to propose, in my view, how the article can be improved. Thanks, Caleb Stanford (talk) 04:42, 7 March 2023 (UTC)

  So far, this thread has successfully dissuaded any editor from trying to implement any of these substandard suggestions. Timhowardriley (talk) 11:00, 7 March 2023 (UTC)

  Follow up: an example of a quality suggestion would be, "Dr. Alpha from Bravo University wrote a computer textbook that has a chapter titled 'Program development'. In this chapter, software engineering and computer programming are discussed." Timhowardriley (talk) 12:51, 7 March 2023 (UTC)

  Tim, your posts have not successfully dissuaded any editor; nor do they attempt to substantively address any points that were raised. Your edits here are WP:DISRUPTIVE. They are non-constructive and do not engage with the criticism. Please do not continue this behavior.

  If you have any suggestions for how to improve the article in a different way, that is appropriate for the talk page. If you think the article is perfect as is, then simply let the page be. Caleb Stanford (talk) 16:51, 7 March 2023 (UTC)

• Regarding About classification, I think it would be nice to overview some answers to the question: what are some of the ways that programs can be categorized?: The article currently describes a handful of program categories. They are: knobs and switches, first generation, second generation, third generation, fourth generation, imperative languages, declarative languages, functional languages, logical languages, object-oriented programs, application software, system software, utility programs, and microcode programs. What is your vision for an overview? Timhowardriley (talk) 17:32, 8 March 2023 (UTC)

Moved "program execution" template and top image to talk.

I'm moving the newly added "program execution" and the top image to this talk section.

Program vs. Process vs. Thread
Scheduling, Preemption, Context Switching

The "program execution" template is too large and negatively affects the article's format. The top image has interesting information, but too much of it. Also, the image should be split in two. There are two concepts present, and the joining of them is unclear. Timhowardriley (talk) 07:31, 10 June 2023 (UTC)

@Timhowardriley The distinction between "Program", "Process" and "Thread" and the steps that a line of "instruction code of program" is delivered to a CPU for processing is important. So in my opinion the image can be restored. That is, the image says "What happens that a CPU processes a program?". Thanks, Hooman Mallahzadeh (talk) 11:01, 10 June 2023 (UTC)

This edit to move the image to the OS section was a good idea. Timhowardriley (talk) 22:10, 14 June 2023 (UTC)

Prolog rewrite

This edit changed the Prolog section from material I could understand to material I cannot understand. This article is general purpose. It should explain computer programs at the most elementary level possible. The prior Prolog material was elementary. The current material is complicated. Was the prior material incorrect? Timhowardriley (talk) 20:22, 15 August 2023 (UTC)

Yes, unfortunately, the previous program was not correct. See the talk page for declarative programming. For example, the clauses cat(tom). animal(cat).are not correct Prolog. Robert Kowalski (talk) 21:24, 15 August 2023 (UTC)

Correct results for Prolog numeric grade example

The displayed results for Prolog's numeric grade example is not what the user would expect. The article says the results will display: X = 'A'. The expected results should display: X = 4. Timhowardriley (talk) 19:59, 16 August 2023 (UTC)

The example in its current form has two relations, which can be queried in isolation, or which can be queried in combination (like a join in a relational database). For example:

numeric_grade('A', 4).
numeric_grade('B', 3).
numeric_grade('C', 2).
numeric_grade('D', 1).
numeric_grade('F', 0).
numeric_grade(X, -1) :- not X = 'A', not X = 'B', not X = 'C', not X = 'D', not X = 'F'.
grade('The Student', 'A').
?- grade('The Student', X), numeric_grade(X, Y).
X = 'A',
Y = 4

Alternatively, the query could be used to define a new relation, which gives the grade of a student as a number:

numeric_grade('A', 4).
numeric_grade('B', 3).
numeric_grade('C', 2).
numeric_grade('D', 1).
numeric_grade('F', 0).
numeric_grade(X, -1) :- not X = 'A', not X = 'B', not X = 'C', not X = 'D', not X = 'F'.
grade_as_number(Student, Number) :- 
        grade(Student, Letter), numeric_grade(Letter, Number).
grade('The Student', 'A').
?- grade_as_number('The Student', Number).
Number = 4

Robert Kowalski (talk) 08:18, 18 August 2023 (UTC)

Followup: It would be nice to learn Prolog's syntax for forming functions. Could this example be converted to a function? If so, then how would the the driver program execute it? Timhowardriley (talk) 20:33, 16 August 2023 (UTC)

Here is the example in Ciao Prolog's functional syntax:

numeric_grade('A') := 4.
numeric_grade('B') := 3.
numeric_grade('C') := 2.
numeric_grade('D') := 1.
numeric_grade('F') := 0.
numeric_grade(X) := -1 :-  X \= 'A', X \= 'B', X \= 'C', X \= 'D', X \= 'F'.
grade_as_number(Student) := numeric_grade(grade(Student)).
grade('The Student') := 'A'.
grade(bob) := good.

?- ~grade_as_number(Student) = Number.
Number = 4,
Student = 'The Student' ;

Number = -1,
Student = bob

Just for fun, the query asks for all input-output pairs that satisfy the functional relationship "grade_as_number". Good for debugging.

Ciao executes the functional program by transforming it into the relational representation, with the value of the function as the last argument of the relation.