GUIdebook: Graphical User Interface galleryHome > Articles > Interview with Nathan Lineback
Go backArticlesInterview with Nathan Lineback

Interview with Nathan Lineback
Welcome to the interview with Nathan Lineback, creator of GUI Gallery Link points to external site – the oldest and the most broad in scope online museum of Graphical User Interfaces.

The obvious first question is: when did the fascination with GUIs start, and how did it turn into a website?

First, I was lucky that my very first computer, the TI-99/4a, was a heavily graphical machine. I did not get my mind wrapped around the “command line is everything” way of thinking that some people seem to follow.

I think my fascination in desktop GUIs really started when I first used Apple II DeskTop running on a IIgs in high school. The way it organized files and folders just made sense and heavily influenced how I thought a file manager and application user interface should work.

Later, when I was in college, I got my first glimpse of Windows 3.0 and how hideous it was. I was appalled at the horrible Program Manager shell and separate file manager. Fortunately I didn’t have to use it that much. Like most people back then I used various DOS based file managers. It seriously bothered me that my 286 with an entire megabyte of ram and a 65 meg hard drive couldn’t do what the Apple II Desktop could.

Another event that inspired me was when a friend of mine gave me two dead monochrome 19-inch monitors. These monitors had icons burned in to the screen, and I realized someone besides Apple or Microsoft had what looked like a usable desktop GUI. I had no idea what the monitors were from, except they were branded “Xerox” – it was only a few years back I discovered with certainty it was from a later model Xerox Star system.

A Xerox 6085 monitor identical to those described by Nathan Lineback
This image can be zoomedA Xerox 6085 monitor identical to those described by Nathan Lineback
What really started to worry me was that people started believing that Microsoft Windows was the first and only GUI. Especially after Windows 3.1 came out, everyone was so dazzled with all the screen savers and sound effects they seemed not to care how bad Program Manager was. Of course the “smart” people “knew” that Apple invented the GUI.

Digging in to the history myself, I was surprised how little attention was given to GUIs other than Macintosh and Windows. I did find some info on the Xerox Alto and a completely incredible 1983 Byte Magazine article showcasing “Microsoft Windows,” a product that at the time (91/92) was only just becoming popular. And the most ignored GUI of all seemed to be VisiCorp Visi On, only occasionally being mentioned whenever someone brought up the subject of the origins of Microsoft Windows. I wanted to show people what I had found, but I didn’t have the ability to do anything about it at the time.

My website started off in 1998 originally as an Anti-Microsoft site. I had actually come to like Windows 95, but Microsoft was shoving IE 4 down my throat. Needless to say I wasn’t happy, and this time I was going to tell the world.

Some time after that I stumbled across a website called “Uncreative Labs” that had Windows 1.01 up for download. I grabbed it and was amazed it was even worse than Windows 3.0. So I made some screenshots and added it as a “Gawk at Windows 1.01” section to my anti-microsoft section. I eventually added screenshots of Apple II Desktop and other versions of Windows, and it just grew from there.

Over the years many different people have had different ideas of how a desktop user interface should work. I feel that it is necessary to find and archive as many of these different ideas as possible. Some may be failures, others successes, and yet others are good ideas that are being held back, but may hopefully one day be rediscovered.

What are the examples of such abandoned ideas?

Well, the idea of a desktop that doesn’t embed a web browser seems to be almost abandoned.

A few specific ideas that are in my GUI Gallery and could emerge again, include the “Beginner/Intermediate/Advanced” modes found in GeoWorks 1.2, vector based icons like in Irix and Bob-like rooms. They will certainly be tried again eventually.

Probably the best idea that has been abandoned is that of simplicity. Early on, due to hardware limitations, serious GUI designers were forced to keep things simple, and to focus on usability (not that they always succeeded). Now that this is no longer an issue, designers feel free to embed web browsers, paste large shiny graphics all over the place, offer a dozen different and inconsistent ways to perform the same function, and even include advertising! Ironically, however, simplifying too much can sometimes make an interface more complicated.

Large shiny graphics seems what sells these days. Steve Jobs encouraged people to lick Aqua and now goes at length to show every little gratuitous animation in Tiger; new Longhorn prototypes have fancy effects and huge photorealistic icons. The emphasis shifted from productivity to attractiveness and joy to use. But is that necessarily a bad thing?

It is kind of like junk food. Companies like to sell it and people like to eat it but it isn’t good for them, and smart people know better. It wouldn’t be so bad except in this scenario everyone exclusively sells junk food because the companies don’t think anything else would sell.

I know some non-technical people that work in an environment that has mixed Windows 2000 Pro and Windows XP Pro computers. They refer to Windows 2000 as “professional” and Windows XP as just “XP.” This tells me that many people do realize that there is a difference.

Hmmm, wouldn’t licking Aqua cause some serious hallucinations? (smiles)

You mentioned Visi On, which is known to be your favourite GUI. You’ve been looking for it on your site since 1999, at one point joking that “Bill Gates may have personally destroyed every last copy!” How did you finally manage to find it?

Like all good things, I found it on eBay. A very nice lady who didn’t seem terribly computer-savvy and didn’t seem to know much about the software was selling it with a bunch of other VisiCorp programs.

Actually, I had been searching for Visi On and information about it since the early 90s. Almost every book, magazine article and webpage that discusses the origins of Windows mentions Visi On, but nowhere did anyone describe what Visi On was. Wanting to show how Windows was not the first, searching for it turned into a bit of an obsession. Searching an entire library and many other places over the years turned up nothing.

Early GUI Gallery from 1999 (you can see it yourself thanks to Internet Archive)
This image can be zoomedEarly GUI Gallery from 1999 (you can see it yourself thanks to Internet Archive)
As it turned out, VisiCorp actually did it to themselves, rendering every copy useless with their floppy disk copy “protection” scheme. The software was serialized to the VisiCorp mouse and/or floppy disk. Imagine your operating system refusing to work when old hardware goes bad and is replaced... oh, right we don’t have to imagine that, Windows XP does that now too!

Were there other GUIs that were especially hard to get? And now that you found your Holy Grail, is there anything else you’re after?

Packard Bell Navigator was very hard to get. The problem is that PB did not include separate installation media for the later versions. It was only pre-installed and part of their restore image. It was also always the first thing Packard Bell owners had someone delete.

A number of other GUIs such as PubTech File Organizer, Irix, GeoWorks, and GlobalView were hard to come by simply because they were uncommon. It was only through the help of very generous visitors to my site that I was able to acquire screenshots of these.

Some other very uncommon ones I would like to get, if possible, include 3-Rivers Perq, whatever GUI HP Apollo workstations had, Atari TOS 3.x and 4.x, and NeXTSTEP 1.0.

During past couple of years you have described more than 80 GUIs. When you research a new one, do you still find features or designs that surprise or delight you?

Sometimes. Of course most of the really good ideas have been copied and used elsewhere.

What probably surprises me the most is when I come across a desktop GUI that has been well thought out, organized, and is consistent throughout. Since most newer desktop GUIs seem to be designed by increasingly larger separate groups of people, this is becoming more and more rare.

Speaking of the newer GUIs, what do you think about the upcoming incarnations of Mac OS X (Tiger) and Windows (Longhorn)?

I really haven’t been following either too closely. The new version of Mac OS X sounds cool but seems nothing earth-shattering. I stopped reviewing Windows pre-release for my site a long time ago because I felt it was just increasing the hype for a product that I don’t care for. I will, of course, review the final versions whenever they come out.

I hear that Longhorn is moving towards using 3D acceleration for rendering applications. A fairly sound move given the state of modern hardware and Apple seems to already do that, but I think it has some downsides. It is common for low-end video cards to have poor video drivers, and these may have problems with longhorn. Unlike Apple, Microsoft cannot control the quality of other vendors’ hardware or software. I also wonder how a move to 3D accelerated rendering for application would affect applications running under future versions of Terminal Server. Also, knowing Microsoft, they will probably get carried away with making the desktop environment 3D like a video game. Until the day comes we all have 3D holographic displays, we still see and interact with the the video screen as 2D (even if it’s rendered by 3D hardware).

I am a little worried about how they plan to leverage this technology. Most people these days just use their computer to browse the web, and people have this crazy idea that everything should be a webby application. Considering the web is mostly 2D documents, how is Microsoft planning to force people to use this stuff now?

Seems I have no choice but to ask you about your dislike of Microsoft. Don’t you think it’s a little bit outdated now? MS Bob is turning 10 this year, Internet Explorer 4 is also quite an old story, and Longhorn actually promises some nice GUI innovations?

Yes, my IE pages are quite old now. Over the years IE has moved from being horrendously evil to almost decent to just plain obsolete. Mozilla/Firefox, Opera, Safari, and even Netscape have new versions out with vastly better standards support, while IE has not been significantly updated in ages. I doubt we will see any major improvement in any new version of IE that might come out. Too many poorly written unmaintined internal corporate web applications as well as local content on “integrated” desktop apps rely on its buggy broken behavior. If Microsoft fixed IE or even made certain enhancements, then these would break.

Even after all this time I still feel angry about what Microsoft did with IE 4, and I still disagree with all of this extreme web integration and the “everything as a webpage” philosophy. So for now I am just going to leave that as is. Besides, everyone has to hate something.

There are some Microsoft products I actually do like. I like Windows 95 OSR2 (I still use it as my main OS!), NT 4, Access 97 (front end GUI is excellent, newer versions are too buggy though), Microsoft Terminal Server totally rocks, and even Windows XP is tolerable after turning off all the webby stuff.

Did you ever wonder what the GUI scene would be like if Microsoft never issued Windows? Which one of its competitors you would prefer to take its place?

I was into computer technology back in the days before people commonly used Windows. I can tell you with fair certainty that if it were not for Windows, the average person would be using a Macintosh these days. And we would hate Apple instead.

IBM PC compatibles would probably still be around as cheap alternatives to Macintosh. Many businesses would still be loyal to PCs instead of Macs, because of their former investments in DOS apps. By today most PC users would be using some version of IBM OS/2. Back in the 90s OS/2 was slow to catch on because it used too much memory for the time, DOS support wasn’t quite what it needed to be (although actually not too bad in 2.0 and later), and the UI was not too friendly if you were used to using a Mac. I expect this still would have been the case, and as a result DOS systems would have remained popular until the late 90s.

GeoWorks/GEOS would possibly have been more popular on really low-end machines. I doubt it would have really taken off on PCs, but if a compact edition of OS/2 did not appear, then possibly many PDAs would use it instead of Windows CE.

We would still have Linux and open source but they would just be haphazardly cloning the Mac and OS/2 UIs. Who knows, IBM might still be the “big evil” and Microsoft would be backing Linux (smiles).

It is hard to say how well OS/2 PCs would compete with Apple Macintosh. It really would have depended on how seriously IBM would have taken supporting and enhancing OS/2. It is also difficult to predict people’s attitudes. Back in the day, even before Windows was in common use, there was an attitude that “PCs were better.” This was mainly because they were cheaper – you could buy a very, very inexpensive bare-bone PC and then upgrade it over time as you could afford it, using the same brand or third-party parts. Apple usually wouldn’t let you do that with a Mac.

This all assumes, of course, that nobody else would have stepped in and written a decent enough, popular, and inexpensive GUI API for the PC. I really doubt that anyone else at the time would have had the resources and motivation to do that.

This also mostly assumes, for whatever reason, that Microsoft’s lack of involvement with OS/2 2.0 was the same; in fact, Windows 3.0 almost never saw the light of day because Microsoft wanted to drop Windows in favor of OS/2 development. Only as a result of some clever folks quietly hacking features into Windows 2.x, it was saved and the direction of Microsoft OS/2 development changed into creating a new core for future versions of Windows – Windows NT. If Microsoft had stayed involved with OS/2, we would probably now have Microsoft OS/2 XP, but not quite as successful and still resulting in a larger Macintosh market share.

Of course, I would like to imagine that if Windows had never been conceived or hyped in the early 80s, Visi On might have been successful and we would all be using VisiCorp Visi On 2005 Professional right now (smiles). Unfortunately, there were other factors besides Microsoft that lead to Visi On’s demise.

Do you think Windows will still be dominant in 10 years’ time? Will GUIs even be around then?

Yes, I think Windows will still be dominant 10 years from now, but by then it should be painfully obvious to everyone that Windows is slowly going down the toilet. There are a number of reasons this will happen:

The core of Windows is close to technical perfection, they can not really improve on it any more in any way anyone would care about. In the closed source commercial software world, once a piece of software reaches this point the only way to sell new copies is to add bells and whistles – but eventually the product gets so many bells and whistles it becomes unusable. Interestingly, open source has the potential to be able to handle this by just leaving well enough alone, but from what I have observed this can fail with developer adding technical bells and whistles until the product becomes unusable.

Microsoft has lost sight of OS design philosophy and vision. This started back with Windows 98. Any sane computer scientist could tell you that an OS and a web browser are two distinctly different things and the code for the most part should be separate. Unfortunately, Microsoft mixed code and functionality to prevent one from working without the other, for the sole reason of excluding competition. The same has happened with other applications, such as Windows Media Player. The result was Windows becoming an increasingly bigger mess.

Most companies – as they get larger and experience turnover – become obsessed with business standardization, procedures, process documentation, and micromanaging employees in the illusion of increasing efficiency, thereby killing all morale. From what I have heard, Microsoft has somehow avoided this for the most part, and is supposedly a great place to work. It is impossible for this to last forever and I suspect this is already changing. I would be surprised if 10 years from now there is not a noticeable change in their work environment. (And if not I would like to know how in the heck they manage that!) Such downward change affects the quality of developed products.

Finally, there already is competition out there. Linux by itself is already a mostly viable alternative, with Wine it can even run many Windows programs. In 10 years ReactOS should be at a point that it can replace Windows on many desktops. I believe that people will be more inclined to migrate to ReactOS, as it is designed from the ground up to run Windows applications.

It may take 20 or even 100 years, but eventually Windows will fall from the control of Microsoft or be replaced. Of course Microsoft will not go down with out a fight. They are already using the broken US patent system to snatch up more software patents. You can expect with certainty that eventually Microsoft will try to use them to stifle or destroy the competition.

As for GUIs in general, they simply will be around a very long time until someone invents some kind of computer-neural interfaces. Then, instead of visually seeing and interacting with information, it will be placed directly in your brain (uh, let’s hope Microsoft doesn’t control this technology!)

And until then? Speech recognition? Smarter offsprings of Clippy? Task Gallery? Jef Raskin’s Archy?

In the immediate future I see desktop GUIs being further polluted by the web. Also, more desktop GUIs will start to render their 2D desktops with 3D hardware acceleration. Macintoshes already do this.

There probably will be more attempts at 3D environments like Task Gallery, but it is unlikely they will succeed, unless stereo vision or holographic displays or something take over. Also, it has been shown time and time again by programs like Bob and Packard Bell Navigator, that common people prefer to not to virtualize reality and to just use a computer as an appliance within real reality.

Speech and speech recognition will probably never take off. Imagine an office full of people chatting to their computers! Speech synthesis has been around forever and speech recognition has been around for a while now and neither are widely used. Heck, I used to actually use the speech synthesizer on my TI-99/4a, but our new Dell Pentium 4 machines at work don’t even have speakers!

AI has been promised for ages, but a computer is only as smart as what humans put in it, and those humans are not as smart as they think.

I do see some potential room for improvement, especially with input devices. Mice keep input simple, but after using a single nerve in your arm to click a million times over 10 years it is likely you will curse each time you painfully have to click. (I know I do.) It is obvious keyboard layouts could be improved but everyone knows QWERTY and it is impossible to just teach everyone a new layout overnight. The desktop metaphor could be revived if someone with enough common sense could scrape off the web trash that has accumulated over the years.

Unfortunately, as long as Microsoft is dominant it is unlikely any real improvement will become popular, unless they can use it to crush a competitor.

Do you consider Konfabulator and Dashboard examples of “web pollution”?

Well, here I am really talking about user interface pollution regardless of the technologies that implement them.

An example of such user interface pollution is the default Windows XP control panel. It presents the control panel as a series of webpages. Users must click through a number of these pages to get to the control panel applet they want. This is in contrast to the classic control panel where, for the most part, all of the control panel applets are listed right there in front of you. There was no reasonable reason for this change and it made it more difficult to use.

Another example: it is also becoming more common to find hyperlinks in application dialog boxes. I find this very annoying as hyperlinks are a document metaphor. Dialog boxes should usually use buttons instead.

It is true people have been creating off-the-wall flashy user interfaces since before the World Wide Web. But it seems that the web has made things worse. First, it makes everybody with a copy of FrontPage think they are user interface experts because they can create a colorful document that people oooh and ahhh at. Second, because everyone is doing it, people now feel justified doing the same thing in formal applications, even desktop applications. (Or, as an interesting thought, some people might be trying to make desktop applications look webby in order to hide the limitations of actual web applications.)

GUI Gallery as seen during this interview (February 2005)
This image can be zoomedGUI Gallery as seen during this interview (February 2005)
And for anyone that really happens to want to design an application with a UI that bugs me to heck and back they should check out my Good UI design tips – If you want to whiz off your users. Link points to external site

Let’s end on a higher note, though: what would you consider the most important software and hardware GUI inventions ever?

It is difficult to narrow it down. There were many, many important discoveries and advances from the first buttons and switches that separated the internal workings of machines from their external function, to the first megapixel graphics displays, to dancing badgers on the web.

I think the most innovative GUI invention I can think of off hand is that of executable resources. This was a very big deal on the original Macintoshes. Using a resource editor it was possible to alter almost every visual aspect of a program including icons, menus, dialogs, text strings, bitmaps, and alert message and all without recompiling the application or ever seeing a single line of code. To some extent this helped keep the job of designing the interface and writing code separate. This also made it unbelievably easy to translate applications to other languages or modify how they looked for special needs. And all of the resources were packed neatly inside the executable so all the end user would ever see was a single application file.

And what can we expect to see next on your GUI Gallery?

Well, there are still a number of GUIs and OSes that I could add. The main problem, of course, is if I can get the time (smiles). Some of the GUIs I would like to get screenshots or photos of are very rare and may not be possible unless another kind visitor donates some screenshots.

I do have a couple of things in progress right now, and some ideas that might eventually take the site in a tiny bit different direction.

By the way, I don’t claim to be an expert in UI design or anything else for that matter, I have just been around quite a bit and have seen what works and what doesn’t. I feel privileged that I have been able to work with as many different systems as I have but there is always so much more to know and experience I always wind up feeling like I don’t know anything at all.

If you think about it, it is quite mind-boggling. Computers can do a nearly infinite number of different things depending on what permutation of programming they have in them. Ultimately, designing a UI is the art of telling a user what they may not do. And there is no one right way to design a user interface. No mater how much work, research, or effort you put in to it, someone, somewhere, won’t like it.

Thank you!

Marcin Wichary

The interview has been conducted over email in February 2005. Nathan Lineback’s GUI Gallery is available at Link points to external site

Page added on 20th February 2005.

Copyright © 2002-2006 Marcin Wichary, unless stated otherwise.