I see lots of complaints that software is too complex bloated and bad these days. I’m sympathetic to some extent but people lose me with the historical comparisons. For example: It’s no longer acceptable to not support Unicode. This is Good - people can use their own scripts and languages more or less seamlessly even if the UI is in English. But rendering Thai or Urdu (or even Arabic!) correctly is inescapably complex. If the historical system can’t do that, it’s not a meaningful comparison!

Likewise supporting TLS is more complex than plain text, especially once you bring in the CA system. But that doesn’t mean that we should sacrifice security for simplicity! Https everywhere is Good! mTLS is good!

In short it is easy to rage against complexity in software but much harder to quantify the amount of complexity required to function appropriately in the Real World, and harder still to build tools to appropriately manage the unavoidable complexity.

We will never have lean software, but we could still have good software.

@daxtens there's a huge amount of rose-tinted glasses, nostalgia and general mindless moaning with this kind of thing.

Not all software in the 90's was blindingly fast, in fact in windows it was normal to have a loading screen for apps because of how incredibly slow they were to start. Plenty of things were bloated, just bloated for their time (early internet explorers, 'active desktop' stuff, all of office, etc. etc.)

Software does a hell of a lot more now, works far better than it used to and is more useful for more people now (unicode being a great example).

I do object to the bloody awful trend of making everything a web app, but that's not the same as thinking everything is necessarily bloated now.

It reminds me of the 'no javascript' or 'the internet was better in the day' people, and you wonder how the hell they have suppressed their memories of how fucking awful the web was in the 90's and early 00's, I remember very well and I can assure young people that it is INFINITELY better now.

And don't get me started on how bad mobile phone software used to be...

@daxtens I think this, and other multilingual areas, have improved. Probably accessibility too. There is a lot of really important, properly hard, inherent complexity in supporting localisation and accessibility. Especially being able to _edit_ texts well!

But, it seems to me that this tremendous improvement is a tiny part of the complexity of modern computer systems.

I’m pretty sure the vast majority of that complexity isn’t similarly helpful, or is actively harmful for localisation or accessibility, because it makes local tinkering and improvement too hard.

My take is we’ve lost a lot of local user and community agency with modern computer systems. Complexity is part of the cause, imho. As are data siloing by vendors to help them become monopolies.

I don’t think anything prevents excellent localisation and accessibility as well as agency. If anything, they ought to be symbiotic. More local agency should encourage more “good complexity” — that which reflects the diversity of the human experience.

@benjohn perhaps! I’m not saying that modern computer software is good - I’m saying a drive for lean software for lean software’s sake, or a fight against complexity irrespective of its purpose, is misguided

@benjohn @benjohn also as a software engineer I think it’s really easy for me to have an overly rosy view of how hackable computers are and very hard for me to put myself in the shoes of someone who doesn’t spend all day on the computer. I don’t know if a school teacher or a carpenter or a doctor has more or less agency than they did in the past, or even how to put myself in such a different frame of reference

@daxtens I agree 💯. I've heard many people complain that a 286 with wordstar did everything they needed from a computer, but apart from a famous fantasy book writer, no one uses that because it's missing utf8, email, advanced spellcheck, mouse scrolling etc. Basically we take very advanced features for granted when complaining we have one too many.

But I still find it unacceptable that my 2020's work computer has a noticeably long latency when pressing a key in MS Teams. We're accusing bloat but it's not bloat, it's carelessness and feature checklist before user experience quality.

@aris yeah I don’t know who (if anyone) on those teams is pushing for software that can operate smoothly (quickly, consistently, predictably, reliably) enough to just fade into the background and be a subconscious extension of your mind , but I do know that they are not winning many battles!

@ljs yeah. (How good are SSDs!?!)

On old software : the thing that really gets me is old games. I tried to play the original XCOM. No tool tips on the buttons so you constantly had to refer to the manual (which I only had in PDF), and I think an extremely minimal tutorial. Brutal learning curve; I just gave up! Old games are interesting, modern games are almost universally better and more fun!

@ljs @daxtens sure, there was less JavaScript. But there was flash. Lots and lots of flash

@daxtens yes 100% about games.

I was playing the quake remaster thing recently and I got lost several times, something that used to happen to me in games ALL THE TIME.

People whine and moan ALL the time about how games are worse somehow now.

And that's before we think about how you could unwittingly have lost the entire game and not realise....

Also SSDs are amazing, biggest game changer perhaps ever for personal computing?

@christmastree @daxtens yes I remember the flash. Christ.

There was also the very secure activex

@aris @daxtens yes putting aside my distaste for electron apps, it's very clear which apps have care and consideration applied to them and which were developed with the 'that'll do' mentality.

MS teams is probably the worst software I've ever used so you know it's a particularly horrific example.

It reminds me of interminable conversations I had with people over my career who claimed that caring about software quality doesn't matter and why don't I shut up about it. That mentality leads to that kind of software.

Part of the issue I think is a desire to treat developers like fungible, uncreative cogs in a machine rather than the reality which is something more like artisans in a subjective realm, some of whom care more than others...

@ljs @christmastree @daxtens let's not forget Java applets, either, which was a very cpu-intensive way of drawing grey rectangles on a web page (that did nothing).

@monsieuricon @christmastree @daxtens oh lord, I wanted to forget that happened...

There was a time when java was just as hyped as rust is now and heralded as the answer to everything, as hard is it is to believe!

@ljs @daxtens The worst part being that the "quality doesn't matter" people may be right when the only metric that matters is income and money, because an app that does X, Y and Z badly beats an app that only does X and Y very well, when Z is important for business.

@aris @daxtens you see that's just not true. Bad devs don't get to choose what is done well or what is done badly. It's all done badly.

One of the biggest myths in software is that you can write awful software and it doesn't matter it's just 'under the hood stuff', but people only think that because they assume that whatever broken situation exists is inevitable + of no consequence.

The reality is that once something is a total mess 'under the bonnet' it not only causes an unending and unavoidable bugfest, but also makes it very hard to add new features, renders some potential new features utterly impossible to implement and requires N times as many devs to keep the shit show going than you would have required had you done it right in the first place.

But you see a suit type will look at that, not factor in 'what could have been' and talk about how it doesn't matter.

No, it always matters. Always.

And that's before we even get on to the 'externalities' of poor systems, like the horizon system in the UK. You save a few quid by paying devs less and not bothering to implement things well, then have to pay out billions because of it.