Software Preservation Group: C++ History Collection

Software Preservation Group: C++ History Collection(softwarepreservation.computerhistory.org)

44 points by quuxplusone 36 days ago | 3 comments

osullivj 36 days ago |

Is C++ the most heavily documented programming language? If so, these docs could form the most detailed description of a subset of possible software ontologies for Von Neumann architectures. By possible software ontologies, I mean the debate around OOP, DOD ArrOfStruct vs StructOfArr, out of bounds and mem safety, GC vs manual mem mgmt, atomic types vs arrays vs objects and their layout in memory, function dispatch mechanisms, genericity mechanisms etc.

tialaramex 35 days ago | |

> Is C++ the most heavily documented programming language?

I doubt it. C++ does have a huge amount of stuff written by Bjarne - its author - because it appears C++ is the only thing he ever did† and he sure does like writing so that's what all his writing is about. He's written several books about C++, various HOPL papers about C++, as well as numerous proposal papers, speeches, standards documents, letters, all about C++

C++ is also enormous and undisciplined, so while there are a lot of C++ documents there's a lot more of C++ to be documented than most languages and I suspect that out-weighs the volume of documents.

† His PhD thesis is, at least for now, only available for personal inspection and I don't care enough to travel to the holding library and read it, but it seems plausible given the topic and timing that the work for that PhD thesis is also just C++ again, or rather a distant ancestor of C++

trueno 36 days ago |

lol you know one of the crazy parts about learning c/c++ well ive learned is the amount of history that needs to be understood in order to understand why the languages and the tooling are the way that they are. proficiency almost in some ways demands you are a versed computer historian of sorts.

don't get me wrong, it's a fun rabbit hole. ive enjoyed it, i think ive gained a lot of reverence for these languages that get spoken of in such a negative light in modern times where "rust everything" is everywhere you look. i chuckle now when i see someone write about how musl had some massive performance problems, because as it turns out glibc wasnt some monolithic old piece of crap it has literally decades of heavy thought/brains/performance optimization that went into it & operating without the knowledge of where this came from = people are reinventing the same problems that have already been solved. im probably learning c/cpp completely wrong working my way up from like 2002 projects so im stuck in c98 standard and every once and then when i hope forward to compile a modern c/cpp project i have a weird appreciation for some of the modernities. but also a weird reverence for the straight forwardness of old stuff. i actually, surprisingly, dont like cmake. i dont know what to make of that, im probably not cut out to be a c/cpp developer because of that but i dunno i kinda prefer cruising through old makefiles and seeing how everything was treed together.

with that, very cool collection of history here. hard to talk about the histories without talking about the attempt at commercialization/capture of compilers and the languages though. intel and microsoft sure tried their damndest. intel doing the cpu proc info not matching intel = no optimizations for you buddy is possibly still in play today. i cant imagine any scenarios where id want to use an intel compiler the experience of getting icc/fortran installed makes me want to punch my face. but it is interesting peeking at the old compilers, icc 5 and its old documentation back in the days when the website looked like this: https://web.archive.org/web/20030130050650/http://intel.com/ i have a copy of the compiler and its interesting to tinker with.