Open Source Needs a New UI

http://blog.cobychapple.com/blog/2011/12/07/open-source-needs-a-new-ui/

by cobychapple • 15 years ago

71 71 15 years ago

blog.cobychapple.com

lukeholder• 15 years ago

There is an obvious reason for this, most of the well designed projects have financial backing (a company that consults around the product/library) and as such have the resources. They want the platform/app to succeed because they finically benefit.

Good UX is good for business.

portentint• 15 years ago

It's all about time. You're writing an open-source project - I know SOME lucky folks get paid to do it, but not most - and you have to just get the damned thing working. Once you do, you have to go back to your job, or whatever else feeds your face. The UX has to wait. And wait. And wait.

vilya• 15 years ago

I think it's on a par with clean, well commented code: people with the right experience and know-how can do it as they go without any extra effort; others will need to go back and rework it to get to the same point - and that's time that they probably won't get. I guess I'm trying to be optimistic and say that it can be done, but it requires a certain amount of discipline.

tikhonj• 15 years ago

I don't think this is really the core issue. If you're not getting paid to work on an open source project, chances are you're doing it just because you simply love programming. You probably aren't an expert designer, and, more pertinently, you're probably more than happy with what you have.

In my experience, building a UI makes you blind to its shortcomings. There are methods and procedures around this, but if you just love programming, chances are you don't know them. I found this out myself when I saw my dad struggle to use even a simple chess game I designed. When you've made something, you naturally know it inside-out; nobody else is going to have nearly that much experience with it.

munaf• 15 years ago

Generally, I agree with the article. But I've been seeing a lot of projects that lead me to believe developers are becoming more design-minded. Many new JS libraries, for example, are beautifully presented.

My hunch has always been that GitHub is setting a trend for OSS developers to present content to each other in a clean and pretty way. Hopefully that extends to the apps they create.

aangjie2• 15 years ago

+1 for GitHub... I think they are doing some good.. and would love to see that as a standard for UI design of open-source projects to follow. Though am not sure..

nknight2• 15 years ago

Right now, GNOME is probably the last example you should cite as "getting it right" if you're trying to convince open source developers of something. GNOME's UI decisions of late are not particularly popular with the crowd you need to convince.

cobychapple6• 15 years ago

I realise it hasn't been popular, and I'm not the biggest fan either, but my point isn't about what's popular amongst open source developers. Also, if developers need to be convinced of the need for good design I think that's a good example of the problem right there.

nknight2• 15 years ago

You don't need to convince them of the need for good design, you need to convince them that what they're doing now isn't good design, and that they should do something different.

I'm saying, citing the project they think is an example of bad design isn't going to help, your opinion will carry less weight.

uokesita• 15 years ago

I like to view it as, gnome at least do some 'renewals' and changes to their site and stuff, oposite as the other sites and software.

Jach2• 15 years ago

I don't know about the author, but I think Gnome 2 is really the best desktop design out there. Gnome 3 on the other hand... I'd agree that right now it's not looking very good.

mikecsh• 15 years ago

Perhaps there is a difficulty in getting designers involved with open source software projects. As a freelance web developer you could overhaul a few of the promotional websites for the software you use and find poorly promoted. Perhaps better tools are needed to make it easier for designers to contribute - maybe this could be addressed in someoway by someone like Github.

naner3• 15 years ago

Yes, we know. Since you're a designer why not offer your time to some open source project you use instead of just offering your criticism.

cobychapple6• 15 years ago

You mean like I did for http://mentoring.is (as discussed here: http://blog.cobychapple.com/blog/2011/09/18/a-distributed-me...)?

naner3• 15 years ago

No, I don't think so... isn't that just a page that says "hey, contact me if you want me to mentor you"?

I meant reaching out to an open source project that you use in an effort to become a contributor.

ugh• 15 years ago

Why do people always say that like it’s a solution? The problem is too big for a single person to solve.

jdthomas2• 15 years ago

I think this has a lot to do with not having to sell anything to anyone. OSS projects are, in general, less concerned with converting users than commercial software -- you like our project join us and use it else 'meh'. While, it would be nice to have a shinny surface on oss projects, I think it is unnecessary.

As someone who recently (for work) switched from using Linux boxes to a MacBook Pro, I would generally take the less shiny functionality over the pretty pictures any day. (\cough\ keyboard shortcut to lock screen \cough\) :)

namank3• 15 years ago

Ctrl+Shift+Eject.

to turn off the screen. To lock it, set your laptop to it to lock after 5 secs and then use the above.

jdthomas2• 15 years ago

HOLY CRAP!! Thank you! ... Wow, I must be retarded or not have looked hard enough as the second google result now shows that. :D Anyway Mucho gracias.

jcromartie• 15 years ago

And Command + Option + Eject to sleep it.

Also, mouse hotcorners work.

namank3• 15 years ago

A very fair point and I think this is mostly because we, as an industry/society, have only recently begun to pay attention to UX.

I'm not going to buy arguments that say good software has (needs) enterprise backing. Open software, more often than not, follows well established industry trends. UX is only just becoming a trend.

The same arguments that ten years ago applied to open software can today be applied to good UX. Thats that. Thats all there is to it.

Give it time, it'll happen. Better yet, MAKE it happen...be the first one to revamp Cyberduck's interface.

omouse• 15 years ago

Unfortunately, the better UI doesn't mean a better product. Working with Magento for example has been a pain in different ways. It sure does look pretty though even when you're debugging things! ;-)

philjackson2• 15 years ago

I made an effort with http://apiaxle.com. Very much a work in progress...

schraeds• 15 years ago

Looks pretty good. I am a UI/UX designer and would love to help out for free to boost my portfolio. A few ideas.. Center large gear image on icon, move smaller to right a bit and reduce in size. Smooth out banding a bit on metal texture. Change button text from "Try it Now" to something like "Download Now", "Learn More", "Get Started" or a combination of the above. Current button seems like free trial and might scare some users off vs being free. Email in my profile if you want to talk more!

philjackson2• 15 years ago

Thanks, schraeds. The idea was that people would pay for my support/development so I would like people to suspect a certain degree of 'commercialness' which is why I chose 'try it now'.

rhlee• 15 years ago

This reminds me of what David Siegel (creator of Gnome Do & Docky) said once at a Ubuntu Developer Summit. Each open source project should have a design advocate whose sole role is constantly make sure that as much though goes into usability as the code.

_dps3• 15 years ago

I think a large source of the UX quality gap for volunteer-driven Open Source products is the grossly underestimated cost of getting good UX. I submit (without proof) that volunteer Open Source efforts are great at competing on easily quantified dimensions (foos per cycle, feature parity) and not-as-great at competing on dimensions that take anything other than free time and a computer to evaluate.

It's very expensive to truly evaluate UX quality of complex products, except at a bare minimum level of "can users do what they want without being harassed". Measuring UX quality means finding the common appeal in a potentially large and heterogeneous group (as a heuristic: the more heterogeneous the audience's tastes, the less satisfying volunteer-based solutions tend to be). To find the common appeal in a large heterogeneous group you have to interact with a lot of people and measure them repeatedly because people's tastes for complex products is not something that can be well-modeled by point-in-time measurements (e.g. it took me a few weeks to discover my MacBook's 4-finger touchpad gestures).

Apache is an excellent competitor in the web server market (very narrow, highly technical audience). GIMP has a decent, but not huge following (somewhat broader audience, but still specialists). Open/Libre Office, GNUmeric, and GNUcash are not going to replace their commercial competitors any time soon.

bluekeybox3• 15 years ago

> To find the common appeal in a large heterogeneous group you have to interact with a lot of people and measure them repeatedly because people's tastes for complex products is not something that can be well-modeled

Or you could just go out and hire a good designer. Your comment just illustrates the problem why a lot of computer geeks don't get taste: they assume it's something you go out and measure.

Any good designer will tell you: you do not measure taste and design for two simple reasons: (1) public taste is constantly changing, and (2) as soon as you can measure something, it's outdated. I call it the Heisenberg uncertainty of taste. If you go to the trouble of polling 10 million people, you will be 3 years behind; measuring the opinion of 100 million individuals will set you back 6 years behind the curve or maybe more.

The kind of minimalistic web design that is commonly seen right now and which Facebook, Apple, Google, and others use was hot among the designer circles about seven years ago. If you tried measuring public taste at that time, you would get a result sheet listing PHP.net as "the" design to copy. You don't measure good design, you define it by producing something that's good and usable in some basic aspects yet not seen before and interesting in others.

aangjie2• 15 years ago

+1 for the Heisenberg uncertainty of taste.. -1 for hire a good designer.. Not because it's a bad idea, but in the current context(Open-source) it's has a lesser probbability of working..... Though would like to see someone try and experiment...

_dps3• 15 years ago

I think this is largely unfalsifiable. I have not seen evidence that designers, on average, know what people will want in advance. You absolutely need a designer to interpret what people tell you about their experiences with a prototype, but I don't think you can tell a designer "I'm building a tractor for your garden. Tell me how it should work." and expect an a priori good design from it.

I'm open to being disproven on this matter, and I'm obviously biased toward empiricism, so if you have suggestions on references I'd be appreciative.

jebblue3• 15 years ago

My wife has used GnuCash for years. Switching from Monday was easy. I think Money isn't even produced any more right? Forget complex stuff like Intuit, not for home finances. The best thing about GnuCash is that nice checkbook like look.

thinker• 15 years ago

I've been thinking about this situation myself this last month and wanted to help give back to the OSS community as a designer/front-end developer - but I didn't know where to start.

I'm considering building a simple matchmaking site that lets open source project developers request for design/front-end help and for designers/front-enders to offer up their time as well.

Would anyone be interested in this if I built it or has this already been done?

cobychapple6• 15 years ago

I think this would be a great project!

mcantelon• 15 years ago

Great idea. Haven't heard of this being done.

SeanLuke2• 15 years ago

To see what's wrong with UIs, we need look no further than X. Consider the myriad of UIs which run on X, many of which aren't compatible with one another: starting with Athena, through Motif widgets, through competing Gnome and KDE, GNUStep, and so on. There must be dozens of major toolkits historically.

What caused this mess? Openness and the hacker nature did.

A UI needs a bully: someone who says that you will do text fields this way, and if you don't like it, tough. Consistency is absolutely crucial in a successful UI. In Windows, the bully is Microsoft. In OS X the bully is Apple. Open systems by their very nature don't have a single dictator.

A UI also needs a bully to force simpler designs. X is replete with applications that have every single possible configuration option. It's part of the tradition. The problem is that this makes the applications extraordinarily unfriendly to the 99% of the population who want to get work done, in favor of the 1% hackers who like to tweak. It's a rare X application which doesn't suffer from this malady: indeed X does itself. There's no one to tell these application designers that their propeller-hats are screwed on too tight.

I don't know if there is a way to fix this. Simplicity and consistency are the two most important features of a UI, and the open nature of Linux, X, etc. simply run against them. When a bully arrives to the Linux world (such as Google in Android) you start seeing quality again. Without it, you wind up with Maemo.

philjr• 15 years ago

More of a benevolent dictator really

saucerful2• 15 years ago

> Consider the myriad of UIs which run on X, many of which aren't compatible with one another

If you are referring to window toolkits, they are all compatible in the sense that you can run different ones side-by-side. If you are referring to window managers, you only need one. If you are referring to a consistent look, the two major ones (GTK and Qt) each have themes which render their widgets in the other toolkit (GTK-Qt and QGtkStyle, respectively). I can vouch for GTK-Qt. It works.

jebblue3• 15 years ago

Totally agree, UI's do not need to be overseen by an overlord of design. I can right now start up a KDE app or a Gnome app and they happily co-exist. Try to pull of something like that on oh Windows, Mac or Android. At least I don't think I've ever seen anything like that on those OS's.

wmf2• 15 years ago

The different toolkits are not visually compatible (I realize many people don't care, especially when the alternative is to do without), and in some cases they're not conceptually compatible (e.g. how do I paste here?).

SeanLuke2• 15 years ago

> If you are referring to window toolkits, they are all compatible in the sense that you can run different ones side-by-side.

So what? I can run X Windows on my Mac using Apple's X app. That doesn't mean they're compatible with one another. I can't drag-and-drop between them. I can't transfer font or color preferences between them. I can't even properly cut and paste between them.

As it turns out, this is, amazingly, also the case for widget toolkits on X -- on the same windowing system!. There are major toolkits which are missing interoperability. Drag and drop is particularly bad. But in some cases even cut and paste is an issue.

> If you are referring to a consistent look, the two major ones (GTK and Qt) each have themes which render their widgets in the other toolkit (GTK-Qt and QGtkStyle, respectively).

Ah, the plaintive cry of the X person who still thinks that graphical interfaces are skins. So 1985, thanks for the memories.

mgunes• 15 years ago

What caused this mess?

Some historical deliberation on that question, in a broader context, by

Matthew Paul Thomas: http://web.archive.org/web/20051030040127/http://mpt.phrasew...

Havoc Pennington: http://ometer.com/free-software-ui.html

cobychapple6• 15 years ago

Interesting read—thanks for posting those links!

acabal• 15 years ago

Exactly right. As an aside, I think the key is to make sure the bully is actually right about what they're bullying.

For example both Gnome (via Gnome Shell) and Ubuntu (via Unity) have already realized that they need to be bullies to see through their grand new UI visions. That much can be inferred by how both projects are interacting with the community and the complaints/praise it generates.

But in my opinion though, the bullies behind both Gnome Shell and Unity are objectively wrong in many of their design decisions, and as a result both products are deeply and perhaps even irreversibly flawed. Unfortunately, a bully by definition must continue bullying, damn the torpedoes; and that's how we end up with products that cause such a deep rift in the community, when in fact the opposite--majority praise at how things have improved--should be happening.

Bullies are essential to good UI. It's a tragedy, though, when the bullies are wrong.

john61• 15 years ago

What is wrong with the UI of Maemo?

the_mat• 15 years ago

"most of these sites look stuck in a much older paradigm of thinking about web design"

This needs some explanation. The vim site looks dated, but clearly I'm not enough of an insidery web guy to know why some of these sites are stuck in a much older paradigm.

In fact, I'm kind of lost on the purpose of the entire rant. UI is important? Is that the takeway? Why even write an article about that?

Jach2• 15 years ago

If the vim site is outdated, so is Amazon's. (Okay, Amazon's less so, but I think the taste issue is more from the tabular layout than the lack of gradients or rollovers. Just my thoughts though.) vim itself has a fantastic UX.

As for the rant, I think people need to be reminded of such obvious things quite often. "What does you in is not failure to apply some high-level, intricate, complicated technique. It's overlooking the basics. Not keeping your eye on the ball." -Jerry Cleaver

DanBC• 15 years ago

The Amazon site is just hideous on so many different levels. Just awful. As well as looking not great it's hard to use and some aspects (search) are broken.

The Vim site looks a bit dated, but it serves its purpose. I can't really comment without seeing someone do a makeover. When I'm looking for information I prefer a bunch of nicely marked up text rather than snippets of text scattered over various shiny image heavy pages.

The Vim website even tells people how to get involved with making it better - (http://vimonline.sourceforge.net/)

reledi• 15 years ago

I agree that UI in FOSS is often terrible, but as munaf mentions below, developers are becoming more design-minded and it's slowly changing.

It's hard to get designers to contribute to FOSS, I don't know why. A lot of developers simply don't care or do care but don't know enough about design to do a great job (I fall in the latter category but I think I'm quickly improving).

The recent trend is that FOSS projects want to have good UI, but they need help. Designers don't need be involved with a specific FOSS project, they can for example create beautiful icons and openly release them.

Toolkits such as Bootstrap by Twitter and Foundation by ZURB have been a big help.

lazerwalker• 15 years ago

I think what's tough is that FOSS design generally ends up as design by committee, with a tendency towards introducing more configurability (and thus more complexity) as a means of appeasing everyone when dissension arises.

From an engineering standpoint, that's part of what makes FOSS great. From a design standpoint, that sounds like a gig from hell.

swanson• 15 years ago

I've always wondered why designers don't point to open source as a portfolio like some developers do.

As a developer, I would imagine that there aren't many better ways to land a job at a tech company than designing a logo or brand identity for a cool open source project. There is no shortage of projects on github that don't even have something as simple as a logo or an About page. To me, seeing a portfolio of tech-related material sure beats 25 brochure-style sites for a local winery or wedding photographer.

chii• 15 years ago

"designing a logo or brand identity for a cool open source project. "

a logo or a brand does not a good UI make.

What you need is indepth research and intuition into what a user "feels" at home with - especially with programs targeted at the non-technical user. This is an amazingly difficult skill to acquire, and thus, is worth lots of money. I don't see a good UX designer volunteering their time for an OS project, when there are many a profitable opportunities out there for them to exploit their skills.

ghempton• 15 years ago

Designers live in a different world. Unlike many developers, they depend on proprietary tools (Adobe) and don't interface directly with open source software in a contributive or decision making capacity.

danudey• 15 years ago

I couldn't help but notice that most (all?) of the projects he listed as good examples are open-source projects funded by companies.

Magento, Firefox, Wordpress, PhoneGap; even jQTouch is maintained by an individual who does mobile consulting for a living. In this case, it's not a group of volunteers who are pitching in for a project, then someone putting together a website. The websites (and projects) in all of those cases are marketing a product that the creator of the website is making money off of (whether directly, though search revenue, consulting, etc).

Me1000• 15 years ago

Fortunately there are designers out there who are willing to help with OSS projects. Cappuccino is a great example of this. Sofa has made 3 versions of Aristo (our theme) for us; they also helped design many other open source apps.

Even though Sofa is no longer around, there are many very good designers out there willing to help work on open source projects. You just have to ask.

b1daly• 15 years ago

On the topic of visual design and UX, these are two (not quite orthogonal) different domains. The area in which they meet, design that helps UX and UX that doesn't look off putting, is usually missed. IMO, software used by professionals can get away with poor graphic design as long as it provides high functionality and efficiency. For software intended for beginners or occasional users the graphic design needs to be stronger to encourage the user to keep trying!

makecheck• 15 years ago

Some investment in good UI is certainly nice. But it may not be clear to non-programmers just how much work good UIs can be (so it should be no surprise if volunteer projects don't spend the time).

There are some fairly obvious tasks, like artwork and putting some thought into what the behaviors should be. There are less obvious things like the fact that huge side effects often pop up.

How much time will you spend on artwork? Every version of Mac OS X looks different, so your custom widget may have 4 different looks and you just know you'll need a 5th next year. Maybe it's a cross-platform program. Maybe the entire thing needs to change depending on where in the world it will be used. And this doesn't even begin to look into issues with programming images (e.g. tracking the mouse inside odd shapes or figuring out how to render things at different window sizes).

Setting UI behavior is also deceptively simple. It can take a very long time to be sure that you've thought of everything even if you're reproducing a variation of UI that's pretty standard already. You often interact with existing things, e.g. a window manager or mouse event system, and you can be seriously screwed if you find a deficiency in a dependency.

What about when your entire framework breaks because you need a customization? (This seems to happen in Cocoa all the time.) For example, guess what: customizable toolbars and resize-from-any-edge come "for free" in standard windows but they are completely unavailable for custom windows. Some changes can look simple on paper and explode when you actually try to implement them. Nothing ruins a weekend like discovering that a simple change would break all kinds of tangentially-related UI elements that were already working. So what do you do...recreate half the OS yourself just so one piece of UI can be a little more interesting or appropriate to a task?

Open Source Needs a New UI

Related Stories