A cooking sequence is always important and difficult. All depends on the solutions space at that specific time. There are times, a simple presentation could get complicated because the reviewers just want to know all the details,other times you have all the details and just can't em to understand the simple presentation.

The Paradox is either way, the root premise of a presentation is that of a buyin of an idea /product / something. Whichever way one obtains that, then ONLY it becomes a "realized" presentation!!

The more one ponders on the presentation space, the more difficult it gets..to cook up viablity in the solutions space..
  
personally I think the presentations are out and story telling is in..

Unfortunatly this is problem with any product that has user/consumer interaction, meaning most anything. Desktop programs, web apps, car dashboards, cell phones, etc...

This is the reason modularity and customizations are so popular among products that allow it. Think of desktop programs, if the user had no options some task they perform all the time may take a long time because the designer didn't think it would be used often.

For those who are familiar, this is where something like Model-View-Controller comes in. Separating the display from the input and processing leads to many displays for the same data. In your example, the text link could just be showing the first result using a predefined set of parameters for the specified algorithm. The crazy web of suggestions would be using the same algorithm with the user specifying the parameters and then seeing all of the results.

Just imagine if your car’s dashboard let you move things around, change colors, etc. This Presentation Paradox is a problem for anyone designing user interfaces.

What a lovely rant! If only half of the idiots, I mean clients, understood usability versus coolness, the web would be a such a friendlier place.

Insist that all presentations are done using a dial-up modem.

Problem solved.

Superb observation. I recommend you the book Paradox of Choice or any of Barry Schwartz videos. E.g. http://video.google.com/videoplay?docid=6127548813950043200 OR http://youtube.com/watch?v=VO6XEQIsCoM . Keep up the good work!

Opps, you forgot the CFO:

Plain app = 4 hours of time included in monthly retainer. All people involved are simple, and simply replaceable.

Full Featured app = 4 hours of retainer time PLUS $250,000 development project with 60% margins. People involved get bonuses, win awards, and discuss the new big thing.

The only thing I don't like about this rant is that it presupposes that the users of the tools are going to be at the exact same level, i.e. that they are all amateurs, all mass market users. And so you compare a complicated tool/interface with a simple tool/interface by seeing how the amateur fares on both. Of course the amateur is going to have a problem!

But sometimes your users are experts. And in that case, you want to give the powerful, though complicated and not immediately intuitively obvious, tool to the experts.

I am thinking, for example, of tools like Photoshop CS3, or Autodesk Maya, or Dreamweaver, or even something like the Eclipse IDE.

When you have a real expert that needs real power, to do real work, they darn well are going to want to use the real tool. Not some thing that gives them no control. Imagine of someone working on the Lord of the Rings films had had to use Google Sketchup to render the entire film, instead of Maya. Or are you going to tell someone to code their entire programming project in vi, instead of in a nice IDE.

Is the IDE intuitive and simply to understand at first? No. It takes time and effort. Same with Maya. But in the end, you can do things hundreds, if not thousands, of time better and easier when you learn the complicated tool.

Your characterization of search as one of those things that 100% solved with the complicated tool and 99.5% solved with the simple tool is completely untrue. Even Larry or Sergei, in John Battelle's book, said that search was only 5% solved. That 5%.. even with the simple Google interface. That was in 2003/2004, however. So let's assume that search is now 10% solved. Having a powerful tool in the hands of an expert user might get you from 10% to 35%. A huge gain. Not just half a percent.

You claim that an expert tool in the hands of an amateur user is a bad thing. It might have a great demo, and wow everyone, but in the end, it is not of real use to the amateur. And I would completely agree with you.

What I am trying to do is point out the opposite of your claims. Giving a simple tool to the expert user can be just as dangerous as giving an expert tool to the amateur user. Imagine if people had to stop using Photoshop CS3, because the interface was too complicated and non-intuitive. Imagine if everyone had to use Microsoft Paint, instead.

> But sometimes your users are experts.
> And in that case, you want to give the
> powerful, though complicated and not
> immediately intuitively obvious, tool to
> the experts.

I don't think you need to separate the two necessarily – if an amateur has huge problems with a bad interface, then a power user will still have some problems and frustrations with it, even if they know how to work around the problems.

But I definitely agree with you that some tasks require more complicated programs. The question is not "is the program complicated?" but "is the program complicated *relative to the task at hand*?"
Then, you need to look at the task space and then the solutions space offered by software makers. For instance, I was working with Lightwave 3D using a pen some years ago, and found it greatly superior in usability to working with 3D Max/ 3D Studio DOS with a mouse for my scope of tasks. Even if the task is complicated, you should offer the most simple solution available. If in the future or today tools are offered that make 3D modeling even more intuitive via cybergloves and what-not, then yeah, I would happily use them because it means less abstraction (= for my task at hand, which was not architecture or technical drawings but video games, music videos etc. – you always need to compare the tools within a certain task scope).

Software would stop evolving as soon as people settle for what's currently available arguing "because the thing behind it is complicated, so must the interface be complex." Yes, maybe it's complex, but maybe it also contains parts which are *unnecessarily* complex for the task at hand!

> Your characterization of search as one of those
> things that 100% solved with the complicated tool
> and 99.5% solved with the simple tool is completely
> untrue. Even Larry or Sergei, in John Battelle's book, said
> that search was only 5% solved. That 5%.. even with
> the simple Google interface.
>....
> Having a powerful tool in the hands of
> an expert user might get you from 10%
> to 35%.

But that's all in the back-end, *not* in the front-end (the interface). You don't need a complicated interface to make search 20% or whatever number better – you just need much more AI in the background understanding much more about the user, the queries, language, the web, etc.

> Giving a simple tool to the expert user
> can be just as dangerous as giving an
> expert tool to the amateur user.

Absolutely agreed!

> Imagine
> if people had to stop using Photoshop CS3,
> because the interface was too complicated
> and non-intuitive. Imagine if everyone
> had to use Microsoft Paint, instead.

I actually use both Corel PhotoPaint and PaintShop Pro 4 side-by-side. But I use them for different tasks – one for retouching, layers etc., the other for pixel work. PhotoPaint is bad for pixel work. PaintShop Pro is bad for anything beyond simple quick tasks and pixel work. For each task, you need to apply the right tool. (But I prefer PhotoPaint to Photoshop, which are roughly in the same task space – for my uses anyway – so you can compare the usability of each...) You mentioned SketchUp, and it's again the same – architects use it for quick sketching, then the final model would be modeled in another program... it's just a different kind of task.

MS Paint, on the other hand, I would argue is even bad for quick pixel work, so I don't think it offers anything to amateurs. Kai Krause did some interesting stuff in the "draw or design without learning much" space some years ago (like Bryce, Poser), MS Paint on the other hand is not a great solution for anyone.

PS: Even though Kai Krause rethought interfaces I still thought the usability was partly suboptimal... but at least he asked the right questions with the software!

PPS: Sometimes, it's not about amateur or power users, but just about smart people who nevertheless have better things to do with their time than figuring out a certain interface.

>I don't think you need to separate the two necessarily – if an amateur has >huge problems with a bad interface, then a power user will still have some >problems and frustrations with it, even if they know how to work around >the problems.

Hmm.. good point. Yes.

>But I definitely agree with you that some tasks require more complicated >programs. The question is not "is the program complicated?" but "is the >program complicated *relative to the task at hand*?"

Does the program complexity match, intuitively, the problem complexity? Is the program neither too simple for the problem, or too complex for the problem? Yes, I agree as well.

Where, I think we disagree is whether something like "search" is a complex problem or a simple problem. You write:

>But that's all in the back-end, *not* in the front-end (the interface). You >don't need a complicated interface to make search 20% or whatever >number better – you just need much more AI in the background >understanding much more about the user, the queries, language, the >web, etc.

I disagree that it is all in the backend. The frontend is not independent of the backend. The intelligence on the backend is limited by the expressiveness of the frontend. Now, maybe the current frontend still has lots of life left in it. But just like there is a difference between the mouse and the pen and the glove for your 3d modeling work on the frontend, there is a difference in the sort of expressiveness you can have on a search interface.

If you are talking about homepage finding, then no, the text box is all we need. But Andrei Broder wrote a paper in 2002, and talked about how only around 40% of web searches were homepage finding searches. The other 60% were transactional and informational. And should therefore require a different, more apt, more designed for the task interface.

Tim, agreed here. There may well be future search interfaces which do a better job than the text search box for certain tasks. Though we also need to see that typing text can be pretty good for communicating certain ideas, and it's even better than speech input in crowded offices etc. But for many visual types of search results, for instance, textual input often lacks efficiency. I may want something "triangle shaped that's red" but whereas I may clearly see the picture in my head, I may not be able to put this into search-engine-friendly words with Google Images.

Yes, also, on visual search.

But I think you're still kinda missing the mark if you're thinking about the advantages of text boxes being all about speech vs. text.

I think of the difference more about the type of information that you are trying to find. When you are trying to find a web homepage, then a text box is fine. But what if you are trying to find information related to the cause of the Iraq war? And suppose furthermore that you are doing original research, i.e. trying to track down all the sources and information that provide the raw materials from which you will eventually draw your conclusions? In that case, there is no 1 query that you will type into the text box. Even if the search engine had perfect natural language understanding and you typed "what caused the Iraq war?", the point is that no 1 or 10 or even 50 results will give you the answer. Your goal is to find enough information to come up with a theory of what caused the Iraq war yourself.

In that case, you are going to need an interface that does a little bit more than just a text box.

> "what caused the Iraq war?"
> ...
> In that case, you are going to need an
> interface that does a little bit more than just a text box.

Agreed... maybe something like the "virtual experts room" for instance:
http://blogoscoped.com/archive/2008-01-01-n87.html

Again, not to criticize any of your ideas, while the notion of a virtual experts room does indeed have its place, I guess I am envisioning queries on things for which there are not really any experts. You yourself are trying to become an expert, by doing the *research* via *search*.