[jfindley]

Whilst that's true, there's also a question of degree. Revit, which is the source of complaint here, takes long enough to open your file you might as well go make a sandwich while it does so. The developers only found out that CPUs have multiple cores somewhere around 2017, and still haven't really worked out how to actually use more than one core for anything much. They still appear to be largely unclear what a "GPU" is, so don't bother putting anything fancy in your workstation.

It also likes to crash a lot, often taking a bunch of work with it. It's slow in a way that's probably entirely foreign to most readers of HN, and even compared to things like adobe photoshop or ptc creo, the rate of improvements is most kindly described as "glacial". In 2020, revit finally learned what a pdf was for example. Sure, large enterprise software development moves slowly. But there's slow, and then there's Autodesk.

[neutronicus]

Ha. I work on a competitor and our story is similar.

We're a little more aggressive with multi-threading but you have to understand these code-bases are massive legacy things whose value comes from a ton of business logic accreted over 25 years.

I assure you I've heard of multiple CPU cores (before this I did high-performance physics simulations across thousands of cores) but I still don't have the first idea how to get something to execute on a background thread in this codebase.

[MiroF]

Multithread is not as easy as you think for a legacy application this big.

[EdStanyer]

Working on an new project in a similar industrial market to most AD products, we've adopted a massively multi-threaded approach to development. But I'm starting to wonder if the product will come to fruition before I'm pushing trolleys 'round at Tesco.