Hacker News Reader

Comment by andrepd | Hacker News Reader

> The general guideline will be to keep Blender functionally compatible with 2.8x and later. Existing workflows or usability habits shouldn’t be broken without good reasons – with general agreement and clearly communicated in advance.

Oh boy. How I wish so much that other software developers would follow this principle... It seems nearly every software I use and rely on has to change its appearance and interface every 6-12 months, breaking familiarity for no objective reason, and simply because "it looks better" to look at (and not necessarily to use!) to the subjective eyes of someone.

View on HN

I wish they didn't. A lot of very popular software is stuck with counter-intuitive interfaces that pose huge entry barrier for anyone approaching them. Only for the people who learned 20 year old idiosyncrasies to feel at home.

The question I always ask myself as an outsider, is this actually weird and outdated, or is it something that, once you get used to, actually makes people able to work more optimally. Sometimes those power tool design decisions are just bad old decisions, sometimes they really do enable the user. Look at Vim, not for everyone, but if you are willing to learn to invest in its crazy, specific style of user interface people can fly in a way other interfaces don't seem able to quite keep up with.

There's always path to making things more approachable no matter how powerful they are.

You could alter vim in a way that people who have years of expeirience of computer use and browser use know at least how to quit goddamn thing without googling it. Or maybe even create short text, save it, open it again, move blocks of text around. For an average person vim has 100% less utility than the notepad.

The power of vim doesn't come from obscure keyboard shortcuts. It comes from editing using parametrized commands (as far as I understand). You could make modern editor with the same power as vim where a person can just sit in front of and start working with immediately and gradually learn that things she's doing manually she could do faster using command mode. And those commands might be the same as in vim because once you go beyond area of shared intuition you can do whatever you want.

The problem of vim is that it started in the era where shared intuition didn't cover basic text editing. And this area grew since then but vim refused to acknowledge this.

Is that a problem, or just reality? One thing I think of a lot these days is the "domain community and the beginner problem". Specifically, many communities I'm tangentially related to seem to overcompensate for beginner comfort at the risk of missing the entire depth of the domain.

I do not see it as a problem that Vim doesn't supply the features you're describing, though I do understand where you're coming from. I am affirmed in my view when we remember that vim runs in a terminal emulator, and the features you're describing are non-trivial to implement in that environment.

I don't mean to sound pedantic, but I don't see this as a problem at all. As it relates to Vim, or many other tools and domains. If Vim was incentivized to increase the size its user-base, I may agree, but this is not the case.

It's only a problem if you want more people to use the power and great ideas of vim. If not then it's not a problem at all. Just reality.

You can't always build something that lets you act like a beginner and gradually transition to the power user version, or at least people haven't always found ways to make that work. The UX challenge of allowing both experiences in the same app, especially with the ability to gradually move from one to the other, is very very complex.

I agree it's hard. Though my claim is that lack of will is more of a factor than complexity of way.

If any of vim great ideas ever enters shared intuition about computers it won't be due to development of vim.

Ability to gradually transition between beginner user towards power user is natural way all modern software is written. That's why you have menus and can point things with mouse and often drag stuff around, and have a cursor and a text box where cursor keys and home and end and delete and backspace and shift works. And you have hints and indications of keyboard shortcuts. Plenty of stuff is discoverable.

You start with shared understanding and build upon that.

Some legacy UIs like Blender evolve and adapt to broadening shared intuition others like vim fail to.

Well blender just did that exact thing going from 2.7 to 2.8 recently, although it was at least somewhat justified. But now that they've reworked it it doesn't make sense to do it again any time soon.

I feel like that move really should have been a major version bump, it was a big move across the board where many aspects of functionality changed or were removed, UI locations and naming completely changed, keyboard shortcuts and UX changes across the board, and so on.

Minor bump 2.7 -> 2.8 = everything breaks, your workflow no longer functions, you have to relearn the API, online resources and documentation no longer relevant for many aspects of the editor

Major bump 2.8/9 -> 3.0 = everything is compatible with 2.8?? Just feels like 2.8/2.9 and what's referenced in the blog post should have been version 3 to me, but maybe they had some technical reason regarding the backend and scripting APIs?

The laws of digital physics say that FOSS jank must be conserved, they just moved it to someplace people won't mind. I'd rather have weird update numbers than weird updates.

I don’t have any inside info, but the vibe I got back then was that Blender just considered the x in 2.x to be the major version (with the idea that they weren’t going to release a 3.0). Though if that was true, it seems they’ve changed their minds in the couple years since.

It was a major version number change. Per the top of the blog post here, they're switching to a new versioning system with 3.x

Prior to 3.x, the major version was the .X part and the 2 was somewhat meaningless.

E.g their versioning prior to 3.x is 2.major.minor.patch and will now switch to major.minor.patch

You're right. I'm so annoyed by firefox changing its interface every once in a while instead of coming up with a good one that they can actually keep stable for years...

If Linux kernel devs had just written v5.14 first they wouldn't have needed to bother with any of the previous 90 or so releases. The fools!

The linux kernel does not break its users, save rare exceptions (security fallout).

I've rarely seen a program maintain two UIs forever when they feel like refreshing their looks

Not that I mind UI change, but I think the comparison misses the point: if it's good enough, for some people UI breaks just cost more than they gain in the redesign. So it's not about reaching perfection. It's about finding a UI only just solid enough that it can stop breaking.

I don't necessarily agree personally, but I can understand that point of view.

Everything you need to know about Linux avoiding breaking user programs

https://lkml.org/lkml/2012/12/23/75

Blender has had built-in well-supported full-featured pie menus for a long time, but Firefox still doesn't, and it's nowhere on the roadmap.

Reply on news.ycombinator.com