@newt
Such threads always amaze me — because comparing "every metric" requires at least understanding what those metrics mean, but this is pure gold — comparing software such as calculator that has been just launched and that is unlikely to allocate any more memory ever to a networked software that has already deallocated all the pages it only needed at start up, but is not unlikely to allocate more if it starts being actively used 🤦
@newt
It's heavily implied here: "no input", but do they really expect for calculator to start allocating memory by gigabytes as soon as they start e.g. adding numbers up? 😲
Not that I don't think that recent versions of Windows Calculator aren't bloated — they sure are, but direct comparison only makes sense with similar kinds of software.
@newt
I'm not knowledgeable enough about how Windows software works nowadays, it used to have resources stored right in the executable files and parsing those required reading them whole, in case with mIRC, again AFAIR, everything was stored in the external files.
But what's more important: your question has nothing to do with mine — is Windows Calculator likely to allocate any more RAM? No, I don't think it is — its startup is its peak RAM usage, it affects the working set — that is the point!
@newt
Man, it seems we're both long in the tooth! It can even do graphing nowadays: https://www.ghacks.net/2020/01/17/windows-calculator-will-get-a-graphing-mode-first-look/
Is it not worth 50 megabytes?
@newt
You know me, I'm all for using TUI software and computers that are well beyond "not current", but this "new software is more bloated than old one" take could at least do mIRC more justice by comparing it to a… somewhat more comparable specimen — in equal environment, not "I just launched this and it only uses 50 megabytes, but it would soon inflate to 1.5 gigs" — why?!!! No, it won't!
@newt
Exactly — and calculator wasn't, all of its pages were accessed "recently" so they are assumed to belong to the working set, keep using it for some time and it would shrink significantly. This is the behaviour I usually observe with software of this kind — local to the computer it's running on: most of its pages can get swapped out and software keeps running nicely.
@newt
This won't necessarily be true for software like mIRC: it has network buffers, it has log buffers — you can't keep swapping this stuff in and out without software slowing down drastically.
In other words, all of these metrics: both working sets and commit charge — only make sense in the context of operating in a memory-constrained system, and for software that was just launched — they simply aren't indicative of anything.
@newt
I'm not even touching the debauchery that is web-based software. Remember how people got fooled into thinking that Chrome's memory usage is modest despite it relying on OS processes for isolation? It allocates, it processes, it deallocates fast — looks good on paper and works relatively well with non-memory-constrained system.
@newt
If something starts happening in all of the tabs simultaneously — even as simple as pages getting reloaded at the same time and you are fucked!
Its effective working set is huge — in memory-constrained system pages start getting swapped in and out perpetually and it slows down to a crawl.
@newt
OMG, just look at this: https://chromium.googlesource.com/chromium/src/+/main/docs/process_model_and_site_isolation.md
"As the early Web matured, web sites evolved from simple documents to active programs, changing the web browser's role from a simple document renderer to an operating system for programs"
I didn't know they have finally dropped the mask on browsers' role as a renderer for web documents, but they are calling this degeneracy "maturation" 🤦