You need more than 2GB memory just to check your gmail

Does anyone else think the software industry is a total shambles? Huge waste and bad quality work seem rampant and inescapable.

I just found out that gmail uses 1GB of memory just to load up the inbox. I discovered this when I tried to check my emails on my old spare linux laptop that has 2GB memory, and it took over 5 minutes to load and exhausted the RAM and started using lots of swap. Even with no other browser tabs open, and no other applications running, on a cut down light linux distro setup too.

Does anyone else find this depressing as I do? Stuff like this is more and more common and I find myself really hating having to deal with it, and even starting to resent using computers/phones/etc. It often feels slow and crap and painful to use in ways that should not be happening with the miraculous computer hardware we have in 2025.

100%

Does anyone else think the software industry is a total shambles?

Definitely. Between shoehorning in “AI” (for Paolo – LLMs) into as many things as possible (think Clippy on acid) and snarfing metrics from every telemetry sensor the software can get its grubby mitts on for the vendors to flog to the highest bidders, the software industry is a sh!tshow.

Huge waste and bad quality work seem rampant and inescapable.

They might call it various paradigms like “agile” and “constant integration”, but the basic premise all too often seems to be “throw together some crap, if it compiles, ship it, if it actually works, even better”.

This is why the EU (which will no longer apply to you Brits) wants to:

• Have some form of demonstrable quality when it comes to software security.
• Have a “right to repair”.
• Have devices supported for a certain length of time and that to be disclosed at point of purchase.

Unfortunately, being the EU, it’s often far too complex and broad for it’s own good which leads to the concept being horrifically abused (can anyone say “Legitimate Interest”?), but they’re sort of roughly heading in the right direction. I just wish they’d stop cow-towing to the leftpondians who are clearly not our friends (look at the techbros all sucking on the Great Trumpian Teat this weekend).

I just found out that gmail uses 1GB of memory just to load up the inbox. […] Even with no other browser tabs open

Am I right in guessing you were accessing GMail with a browser? Which?
I find my Firefox (on Android, rarely use a PC) is quite memory hungry, and not at all fast, but then I have a metric tonne of stuff going on in the background to sanitise the internet for me: UBlock Origin, Privacy Badger, Tampermonkey to fix annoying sites, Decentraleyes, AdNauseum, Consent-O-Matic, ClearURLs, and some others I forget the names of.

Actually, this is something that interests me regarding Iris – is there any possibility to add some sort of “hook” (even if just tossing the URL and window handle into a Service Call) to allow fetches to be vetted?

Does anyone else find this depressing as I do?

I find Apps depressing in general.

AccuWeather – 188MB (app 146MB, user data 42.19MB, cache is the rest).
The “user data” that I know if is my location (by name), some preferences, and a flag to say that I’m not a subscriber. That’s about all it needs. How the hell does that add up to 42MB?

Adobe Acrobat – 382MB (app 179MB, user data 203MB)
There shouldn’t be any user data. I don’t have an account, I’m not signed in, I just read PDFs that I download. What the hell adds up to more than the app?

Clock – 131MB (app 56.95MB, user data 73.71MB)
This is taking the piddly now. It’s basically Alarm with go-faster stripes. I have a couple of scheduled alarms (like “wake up you twat”), and it’s set to tell me the time locally and in Baltimore (it’s nearly noon there as it’s nearly six here). And the remaining 73MB of user data is…?

Meanwhile CustomRadioPlayer is 24.13MB of which the app is 23.03MB and the user data, which is something like two dozen custom stations, adds up to 1.10MB.

Remember that Android devices often come with fixed amounts of flash on the device, and it’s hit and miss as to whether there’s any SD card slot to allow it to be expanded. So wasting dozens of megabytes per app is horrible.

It often feels slow and crap and painful to use

It’s like inflation versus wages. Inflation rises faster than wages, so while you’re earning more than you were, it buys less.
Software grows faster than resources, so while you’re doing more than you were, it takes longer.

and even starting to resent using computers/phones/etc.

I don’t resent the tech, I resent the fact that more and more things are simply unable to function in any meaningful way without being tethered to the mothership; and with increasingly fewer and fewer things willing to work with open protocols (to cut out the mothership).

I also resent the fact that manufacturers increasingly feel they have the right to arbitrarily modify my devices without asking.
I have a Xiaomi phone that I mostly use for making videos and Netflix. Often when the WiFi is patchy or drops out, something called “Analytics” crashes. It has only recently started happening. So I look at the app info and see it was updated in early January. The one currently installed was installed on the 12th at 0h26, updated by itself, despite my preference for things to not self-update. It keeps gaining small amounts of data (reflected in “Storage”) but claims zero data use or battery use, while disallowing the “pause app if unused” setting. I’ve twice turned off the “Use mobile data” setting and it has turned itself back on (irrelevant for me, there’s no SIM in this phone). Shady as hell, that, isn’t it?
I think if companies want to sniff telemetry and control what goes on the user’s device without respecting the user’s choices, they are perfectly entitled to if they provide the device for free. Otherwise eff off and let the user make the choices.

But, then, we must consider that a lot of modern software comes from a country that has a highly warped concept of “freedom” and very little concept of “privacy”, so the only surprise ought to be “why are we putting up with this nonsense over here?”.

This is why I like RISC OS.

Actually, this is something that interests me regarding Iris – is there any possibility to add some sort of “hook” (even if just tossing the URL and window handle into a Service Call) to allow fetches to be vetted?

On your home network – run Pihole then everything “analytic” or advert is blocked if the DNS says the FQDN of the analytics (or whatever) is 127.0.0.1
Just be aware that some sites (looking at you Cricut) intermittently fail to load the page – because it tries to load a specific advert before moving on to the rest of the page.

Adobe Acrobat – 382MB (app 179MB, user data 203MB)

PDF readers are built into most browsers these days

I think if companies want to sniff telemetry and control what goes on the user’s device without respecting the user’s choices, they are perfectly entitled to if they provide the device for free. Otherwise eff off and let the user make the choices.

People are conned into believing that having “Private Address” (generated MAC per SSID) is to stop the user being tracked – it isn’t, it’s to stop third party link providers from tracking you so that the data collected by your ISP and the device manufacturer becomes more valuable.
Ever considered why the GPS position info on your phone is suddenly more accurate when you turn on Wi-Fi?
Even when you’re a mile from the nearest building that might have Wi-Fi.

Yep, they want to plot the relative positions of Wi-Fi hotspots and your phone – “thank you for doing our surveying for us, and paying for the privilege, sucker”

PDF readers are built into most browsers these days

Yes, my Firefox on Android will display downloaded PDFs, but I can’t get the damn thing to be a PDF reader app!

If you are accessing Gmail on Linux, you might find using a proper mail application, like Evolution, better that using the webmail interface. It cuts out the browser overhead.

Remember that Android devices often come with fixed amounts of flash on the device, and it’s hit and miss as to whether there’s any SD card slot to allow it to be expanded. So wasting dozens of megabytes per app is horrible.

Agreed. Many times.
And even with a SD card, many apps will not move there. And you cannot remove the junk that is ‘pre-installed’.
But the phone suppliers have a vested interest in all this – when phones become unusably ‘full’ users upgrade.
Even keeping my storage usage as small as possible, my first 2 smartphones were replaced because they were ‘full’ – I am now on my third.

PDF readers are built into most browsers these days

I was disappointed when Google dropped support for Win7 from the desktop Google Drive client. Then I looked more closely at it – a file synchronisation utility that just uploads your changed files to the cloud – their build contained the entirety of Chromium… presumably to display the trivial GUI.

It’s not just that nobody knows how to make something small any more… nobody knows what “small” is any more.

It’s not just that nobody knows how to make something small any more… nobody knows what “small” is any more.

So many examples of "just a ‘small’ agent on the PC – my mind says small agent is less than 1MB and preferably less than 100K – their definition of ‘small’ agent is something that is less than 100M

Anyway, if Adobe are going to do simple readers that are silly large and bloated, and the browser guys do silly large and bloated, common sense says that with what is available take the bloat in the browser and ditch the limited function bloat of the Adobe offering.

We’re all in the same bloat.

nobody knows what “small” is any more.

I don’t think it’s that they don’t know, I think it’s that they don’t care.

Their excuse, of course, will be along the lines of “we included an entire operating system and the kitchen sink in order to control what works with what because user security etc etc”.

my mind says small agent is

We grew up in a world where stuff had to fit on, at absolute most, a handful of floppies. One couldn’t spaff megabytes of memory because the machine simply didn’t have that much.
These days, I often feel that younger coders simply don’t understand efficiency because they never had to worry about such things. Rather than careful data use to reduce wastage, they’ll tell you without any sense of shame that if it’s slow it’s your hardware, get a better processor/phone/GPU/etc.

my first 2 smartphones were replaced because they were ‘full’

Can you not plug it into a PC or something (or on more modern phones, attach some USB storage) and transfer off a load of stuff? I periodically do this to my phones when they get too full of photos of Anna and bowls of linguine. ;)

That is the reason, why i like Risc Os, and the small and fast software on it.

Early mainframe programmers thought exactly the same of us microprocessor whippersnappers; “A whole 32K of RAM? When I were a lad we had to weave our own core store out of chicken wire!”

Chicken wire? That’s posh that is. When I were a lad we used pebbles and sticks and had to imagine the rest.

Etc. ;)

@ Rick

(for Paolo – LLMs)

Thanks you :) – also because I did give you a full Ai engine that builds on RO in just 15KB!!!! (relevant in this topic)

@ On the topic

Does anyone else think the software industry is a total shambles? Huge waste and bad quality work seem rampant and inescapable.

Oh, absolutely! Modern software engineering is like McDonald’s-quality food. Most people don’t even know how a CPU actually works, yet they can have a full career, earn big titles, and collect large paychecks without knowing a thing about quality software.

The problem: More features and faster

The “solution”: Layer more and more abstractions in the form of frameworks and libraries until the engineer delivering the feature only needs to write a maximum of 50 lines of code.

The “solution” side effect: There is an enormous amount of abstracted code that isn’t even executed—it’s just loaded into memory.

Then there are optimization techniques that improve performance at the cost of using more memory. There are software caches and buffers everywhere. This happens because most folks only have a few minutes to make a decision about architectural or code design changes, even though those decisions will impact the solution for months or years.

Then comes the delusional hope: “But we’ll fix it later”. Time passes, and the feature creep industry keeps pushing for more and more, stacking layers of bad choices.

And finally, the last bit (the cherry pie on the “cake”): Using interpreted languages and garbage collectors to avoid dealing with complexity and memory management as much as possible.

The result? Exactly what you’ve mentioned in the title.

Until investors keep pouring money into an industry that doesn’t want to improve and treats the next feature as the holy grail, I’m afraid things won’t change.

Early mainframe programmers thought exactly the same of us microprocessor whippersnappers; “A whole 32K of RAM? When I were a lad we had to weave our own core store out of chicken wire!”

Very true Druck, but back then we were 1KB of RAM vs 32KB. I am sure you have seen what’s going on lately too, we are with things that would work absolutely well with 64MB or even 128MB of RAM using 16GB of RAM!!! An example? Some linter for VSCode studio written in Java literally keep abusing memory to up to 16GB and more! Druck a linter XD