Why Free Software has poor usability, and how to improve it | Hacker News Reader

Why Free Software has poor usability, and how to improve it

mpt.net.nz

20 pointsdylan10 comments

This article is somewhat misleading, in that of the 15 things listed, only the bottom three are identified as unique to free software.

Indeed, some of the most disgustingly unusable stuff I have ever seen has come from companies, not only Microsoft but "enterprise" people, groups that should have a vested interest in software quality.

A problem I didn't see listed, is that the people acquiring software may not care. Many of them absolutely don't. I've been in organizations where the people in charge of getting stuff don't have a clue what "features" or misfeatures are actually impacting the usability of the people they are supposed to serve. They don't ask, they don't factor it into the buy/download decision, and no one seems to trace the resulting productivity problems back to them.

actual work conversation from last week...

me: "Yes, but will this new software reduce the time it takes folks to do their jobs."

them: "I don't know."

me: "Do you know how long it takes them to do it now?"

them: "umm... no."

you hit the nail on the head in that the people making the buy decisions don't consider usability.

You can reduce the time it takes someone to do something, without knowing how long it takes them to do. That's what I'm working on right now.

If you make everyone's computer 100% faster, you know that you're delivering a net benefit, even if you're unable to quantify it (it's certainly not 100%).

>> Coding before design. Software tends to be much more usable if it >> is, at least roughly, designed before the code is written.

In my experience, software and web development tends to fail when early work is made with specific UI assumptions in mind. The successful projects I have been a part of have concentrated on putting basic functionality first, and then iterating around the usability design and other features.

As an opposing datapoint my experience is the exact opposite.

I think it works both ways.

Putting basic functionality first doesn't mean eschewing UI assumptions-- it just means making LESS of them. You can't make anything without making assumptions.

Too often, devs assume that iterative development means that they build the simplest version however they feel like and THEN call in the designer to work with them to clean it up, improve it, add features, and make it pretty.

IMO, the best way to do iterative development is to work with the UI designer in the designers preferred medium for early prototyping (xhtml/css, photoshop, paper, whatever) on the "basic functionality" that you mention. THEN build it and have your target audience use it. Rinse, repeat with improvements and new features.

I disagree with the premise of the article. On average, Free software seems to have better usability proprietary applications.

Furthermore, usability is rapidly improving, since developers seem to be becoming more aware of usability issues in general. Sure, there are examples of horribly unusable apps, but I find the ones that suck the most are the proprietary windows apps, even the ones from big companies, and some of the most usable and consistent apps are open source.

Even so, there is always room for improvement, and the biggest issue is simply the lack of developer time to implement all the great usability ideas that pop up.

"...Free software seems to have better usability..."

Not sure about that. Dia vs. Visio? Visio wins. Maya 3D vs. Blender? Many say Maya 3D wins. GIMP vs. Photoshop? I'd say that one was really a tie with a large pool of people just used to the way Photoshop does things. OpenOffice vs. Word? OpenOffice still has lots of interface snafus compared to Word. (But AbiWord doesn't have those.)

I contend that Open Source / Free Software almost never achieves the same level of GUI polish as the best proprietary apps without someone being paid to do it. The market forces in the volunteer-only situation almost never push it to that level.

If you contend that Open/Free and Proprietary apps as a whole are around the same average level, I would generally agree. But let's be honest here: that's setting the bar way too low. Everyone should be aiming for the top few %. And when it comes to that, the Proprietary stuff has the lion's share.

Rather than creating a large reply, I'd rather put it down to what is the single biggest problem is that, in my eyes, trumps every single thing in his essay: Ugly widgets, icons, fonts and badly thought out standardised user interfaces in both the KDE and, to a lesser extent, GNOME projects.

I think it's obvious why - every application inherits from them and their bad design decisions result in most applications having the same usability problems.

NB: The author apparently thinks that making me scroll a hell of a lot is good for usability - instead I ended up turning off CSS.

Designers are used to a top-down approach. i.e. They create a design (a full-on design spec with waterfall or a UI improvement/iteration). They like to have design authority. And, honestly, democratic design is generally pretty awful.

Devs generally get involved with OSS to make a different and to build something that THEY want to have (not to serve the the average user). Devs tend to be the "managers" of OSS projects.

So the big question is: When there is a disagreement about user experience in an open-source project, who wins? Who has authority? Obviously, you shoot for consensus, but it often is hard to achieve. Whem you can't achieve consensus in normal software development, a manager will usually pull rank and say, "Well, it feels like a design decision. Let's err on the side of going with our UX designer and the other people who agree with him/her".

I've never done design for an open-source project, but I'd wager that it's pretty democratic with a strong bias towards what the programmers want (which often don't jibe with what the designer wants).

Reply on news.ycombinator.com