Computer scientist and entrepreneur David Gelernter believes computers should imitate life. That means rethinking what it means to manage knowledge—and replacing the current PC and Mac desktop with a “narrative structure.” Read the article at CIOInsight.
Rethinking the GUI for the Big Picture
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Why don’t these articles have mock-ups?
All the author says is that current interfaces aren’t linear enough. As much as I hate GUIs, I think that the non-linear aspect, the fact that you can browse in multiple different dimensions, is about the only thing that computers got right so far.
Besides this, the person offers no tangible suggestions, no ideas for how to concretely rework the current GUI. Quite a pointless and redundant article.
Worthless! Show me the code.
>Worthless! Show me the code.
“Computer scientist and entrepreneur….”
Computer scientists do not always write code for the most part. They invent/discuss things. Programers do the donkey’s business. At least this is how it is in Europe.
It is like saying that Knuth’s work is worthless (while he did actually wrote some code back in the day).
He seems to think that all you have to do is arrange data in
chronological order and everything will make sense.
Today I needed to collect some images of more and less realistic
portrait sculpture. I found some in a book and others on the web.
Arranging the data in chronological order wouldn’t have helped at all
– neither the dates of the original art works nor the dates of the
photos of them, nor the dates of the scans.
In general, one of the biggest problems with data on the internet or
elsewhere is judging its quality. A great deal of it is either poor
quality or even deliberately misleading.
I don’t think this guy’s narrative would help with that.
I do remember thinking more in terms of the HTML method or a lose relational database. For instance a number of different stimuli external or internal can cause me to think of a childhood memory. I don’t have to start a birth and sift thru each memory to get the one I want. The context comes from the existing links to the data as well as constantly added new links.
Vince
On TechTV’s Big Thinkers show, they did a show on this guy. They showed him using his new GUI or atleast something that was a working prototype. It was insane. Basically imagine having Mozilla open in a 300×300 window, ok so now create 300 of those and put them in order behind each other, but each window’s Y coordinate is about 5px higher than the one in front of it (basically an infinate mirror effect). Now each one of these windows is there by date. When you select something in a window a whole stack of windows relating to what you selected would then come up behind the window you are in.
Revolutionary? No.
Ridiculious in it’s current form? Yes.
that is just my opinion, take it for what it’s worth
There is not one concept that he offers that is backed up by research. Or insight. Or vision. Or genius. Or code!
Sorry, David, the dotcom era is over. Try coming up with something real. Or find some better graduate students to steal ideas from.
#p
Yet another Dektop is Dead article. It is all useless navel contemplating untill someone actualy implements it and tests it. WIMP may not be perfect, but it is useable today. I have yet to see any real implementation of an alternative.
What does it have to do with WIMP? I think that the article has misleading name. What I understand from the article is, that author talks about knowledge management area and the ability to logically link information.)
Imagine typical police investigation 🙂 They put bits and pieces into one story and the idea is not to see the pieces they collect, but to understand the whole story, and make the conclusion.
I think that author is wrong though 🙂 Our brain is able to make conclusion upon pictures, sounds, scanned documents, but you still need some data organisation, navigation, indexing, to help your navigation. Imo there is not ideal solution to the problem yet.
I’ve heard associative data model is the way to go, but after seeing Minority Report – who knows 🙂
-pekr-
I saw an interview with TechTV. It’s a load of crap. His ideas (these seem his convervative ones) are to have a totally unorganized system, as your desk. I’ve always wondered: most of us _wish_ our desktop was like our computer, not the other way around!!! Plus, code is linear. That’s just how it is. How are you going to make code that just “places” things places? It just doesn’t make sense.
Hm…I’d also like to conduct an informal survey. Does anybody thing there is anything _radically_ wrong with the GUI as it is? I love it.
This person is trying to propose a concept that goes much further than our present desktop metaphor. Having said that, I do not see a way to do what he’s talking about at this time. He is right though, what we have since 1984 are really refinements, add-ons, more features.
What he’s talking about is true though. Our lives are not a desktop, we do not live our lives in that manner. The key is life as narrative. And it goes deeper than that – some believe that life (civilization) is language. Without language, there is no way to interpret experience. Without that ability all experience are just bits of random information. Language is narrative and narrative is the only way we have to make sense out of anything – to get the big picture. Mathamatics and scientific thought are part of language and narrative, not separated from it.
So, to achieve what he is talking about would truly be revolutionary. It is hard to wrap our minds around it because we do not have the tools yet to do this. But, you have to start somewhere.
As for the question of there being anyhing radically wrong with the current desktop/GUI, I would say no, and, in fact, it is at its very best now. But, it is still the desktop metaphor and that is not life. It works for the purposes it does have because we do understand the desktop metaphor. But, it is just the beginning – computing, in a historical sense, is so new, it is like a two day old infant. There is much, much more to come, thank goodness. To dismiss the ideas he presents is to say that the desktop metaphor is the end of computing, that there will never be anything beyond it. Now that is what would be ridiculous.
So, to achieve what he is talking about would truly be revolutionary. It is hard to wrap our minds around it because we do not have the tools yet to do this. But, you have to start somewhere.
There’s a demo of it on his company’s website. IMO, it’s an almost completely useless and frustrating method of navigation. It’s roughly akin to what my (physical) desktop is like before I file things away, and it’s impossible to find anything on my desk if I’ve been working on the same project for a while (or a few different projects at once). Sure, if I have any clue of WHEN I worked on what I’m looking for, I might be able to find it (assuming that I know when everything else in that stack was worked on). However, if I just want to find a bit of information on project X and I don’t remember if I worked on it before or after project Y, I’m better off having project X filed away as project X, not as a collection of documents worked on between 2000 and 2001 interspersed with whatever else I might have done at that time. For that matter, if what I really want is a revision I did in 2002, but I dont remember actually doing that, then I’m completely hosed.
When I do want to look at something by date, there is a date field to sort by in the details view of Explorer (and most other file managers), and that’s usually more useful than sorting everything in my filesystem by date.
I watched the TechTV show on him. His whole office was covered in papers and his desk was a mess beyond belief. He said that he knew where everything was in the piles, yet they showed him digging for papers and coming up fruitless a couple times. I think that says quite a bit.
mike
These hippies need to stop thinking about desktops and get back to hugging their trees.
Anybody have a screenshot of his “vision”?
There have been numerous attempts to portray the GUI of the future in SCI/FI novels for at least the last 20 years – by far the most common recurring theme is that of the narrative.
Especially strong focal points for the writers was elements of immersion (3D / Virtual Reality) and language as an OO glyph form to store data and design (programming) data manipulation concepts.
One example of this would be Palace by Katherine Kerr and Mark Kreighbaum. There are many others.
Perhaps we should get our ideas from the creative members of our society and then let the engineers worry how to do it
Well, Kai Krause (the guy behind Bryce) was playing around since the early nineties with “another” (if you know how Bryce looks like, you know what i mean) GUI named “Aurora”.
It would be nice to see some screenshots of Aurora…
I love these “We Must Replace GUIs Now”…..
I don’t think that guy’s suggestion is the best suggestion. What that is confusing to users right now is the overwhelming amount of menus, toolbars, dialogs and so on. Not the hard disk concept. Managing files accroading to the time spells a lot of trouble. I personally have almost 300 personal files. Now, if I were to browse them based on when I last opened them…. now that would become hard…
I don’t think the concept is dead, and should be replace. I think the implementation of that concept, whether it is Mac OS, or Windows, is bad. What would be nice is an choice for users to choose between concepts. For example, when a person is writing down phone numbers while answering the phone, he may be better off with the time based metaphor.
I would not put any real stock in this person’s application or his confused actual desktop. It is the idea. Even his example of a person’s health care narrative – he did not put that well. He made it sound like, as mentioned, just a chronology by date. That is not what it is really. It is narrative and narrative is the engine of civilization and of each person’s life.
But, I think it is really in the realm of idea right now. Attempts like his obviously do not make it real. It seems to me there must be a whole lot of stuff that would have to happen in computing to make this real. But, I do believe it is very important because I believe narrative is the context of human life, history and civilization.
This man is an idea man. He is not an imbecile.
I don’t consider myself an ‘expert’ as I am not an undergrad or grad in a field related to computer science.
Nor do I have a career especially leaned toward that (well, kind of, digital artist– but I’m not a computer scientist or anything even close).
But, it seems that this article doesn’t really say anything than “stuff should be organized to find it.” Okay, thanks.
The only idea I could think of for a radical desktop paradigm shift is have applications work more like plug-ins and windows act more like canvases. That way, you could use a paintbrush application and then a word processing application for the same document rather having Photoshop do images and Word do word processing. Or Photoshop do images and bad/non-existant word processing and Word do bad/non-existant image editing. Your applications would become more like tools than seperate workrooms. But there would be some obvious setbacks to that system.
But for Kai Kruse information.. http://www.byteburg.de/
Hmmmm, KDE seems to be heading that direction. For example, I could do word processing in KSpread, and spreadsheets on KWord…
The wonders of KParts. Wonder how good it would be if it uses CORBA…
I think what he’s saying is interesting stuff, but there are some problems. The narrative idea is fine to a degree, but I fail to see how this will *replace* the GUI. If anything, it’s complementary to the GUI. We may well want to organize data into a chronological, linear narrative as he suggests, but we need space to look at it, most likely in a window on our screen(s).
Second, a chronological narrative isn’t necessarily the only or best way to look at things. Not all data is best “mined” by time. What we really need is more control over our data, being able to organize and rearrange it in any number of ways to see it from different angles, to make new connections, and to get the most out of the available information.
In short, we’re talking about meta-data structures, searching, filtering, etc. I see nothing “revolutionary” about this, it’s simply the difficult matter of making applications that can find and sort the data and make it available to the user. It may be desirable to modify operating systems to make it easier to mine data, but we do seem to be heading that way: file attributes from the MacOS and BeOS are coming into vogue in Linux (XFS) and future versions of Windows.
The next step is writing the applications that make the most of those attributes for the users. But that mostly seems to be a matter of trial-and-error, unless someone has some special insights that haven’t occurred to me.
The idea irself has been around (not for very long though), but the implementation would be revolutionary.
Just to clarify, he is talking about replacing the desktop metaphor GUI, not doing away with a GUI altogether.
Gelertner has been pushing/developing/selling this idea for years.
http://www.scopeware.com/products/prod_overview.html
Look down near the bottom of the page for a screen shot.
All I can say is what a waste of screen space for a search engine and a preview pane.
I don’t use the “desktop” for file management, I use a 2-column file
manager. There are many of these around – for example, Windows
Commander.
They all allow you to sort each column in either alphabetical or
chronological order. (Or by file size, protection bits, owner, group,
etc.)
Many people have suggested a relational database system. These only
work if you have entered the data. If I scan in an image of a portrait
bust of Ware by Roubiliac, that will be simply an image with a file
name unless I take the time to type in data such as author, subject,
medium, copyright details, etc etc.
Nobody has time to do all this data entry. It will never happen. So
there is no way I can automatically find all the images of portrait
busts that I might have in five years time except by remembering them.
Sensible file names do help.
I do think a database would be better. I think its worth the trouble to type in some extra information.
Let’s say you went for a trip to some place and took some pictures of yourself riding a bike. You would like these pictures to be linked to other bike-riding pictures, but also linked to your other pictures of that trip and to pictures of a previous trip to the same place. A database could do this. Files and folders can’t.
You say sensible filenames help, but I always end up typing just as much information into the filename as I would type into the database. Either that or I’m using abbreviations that noone would understand but me.
the human brain is more like a hologram. it has the special property that if you remove a part of it, you do not lose any information, though your memory might be a bit more hazy and you might forget things easier.
that is how a hologram works. you cut it in half and you have 2 identical pictured of the orrigional, however, the 2 new holograms are half as sharp as the origional.
I don’t necessary agree that his is the best solution. However, I agree with the problem he states. There is so much information here and there, it is almost impossible (or requires a great deal of time) to organize and absorb it.
Often, in this information age, we lose information, or can’t locate it when necessary. How many times have I found a great resource on the net and then lose the bookmark to it, or the site is taken offline? So I agree that all this information is scattered.
A linear structure might be part of a solution, but not an end all solution. We tend to think both in a linear way (At the end of my shift I will: shutdown my workstation, clock out, get in my car, drive home, check the answering machine, drive to Vegas, sell a wallaby on the black market to Lefty Marone’) , and by association (I found the wallaby on the road by a farm in Nebrasksa. A friend of mine had a farm growing up with a barn painted the same shade of Red, the tractor was a ford. I need to change the oil on my Ford truck.).
So, a way to linearly relate topics in some sort of chronological way could be useful, coupled with a method of indexing associations to that data. I’ve seen the same sort of layered window scheme, I don’t like it. At least not with a 2-D viewport like a monitor. Until we have holographic displays, a useable virtual reality headset, or some sort of neural interface, the monitor will be a limiting factor.
Ideally, you want an application that will work automatically, sorting through files, documents, pictures, maybe even pulling stuff from the internet, and making that available to the user. It would look at all files and documents available to it, and index it.
The user can then select the criterion he wants and have the application display the available info based upon that criterion, in the format or order that the user wants, be it chronological, by theme, keywords, etc.
Of course the user should be able to manually manipulate the data, making connections and groupings that the application doesn’t make automatically, but there should be a lot less “data entry” on the user’s part than if he/she was making a relational database.
Extensive file attributes are useful because they help to organize the data, again, without as much typing as a formal database would require. A picture file could be called some meaningless name (or even an obscure abbreviation) like 1qtr.jpg, but the attributes could give more meaningful information, like a description, category, dates, etc. Somebody would have to enter the info once, but then those attributes remain part of the file, whether it’s transferred over the internet, by e-mail, by floppy disk, or however. This won’t prevent data entry, but it should seriously reduce the need for duplication of data entry.
The key to making it happen is having file systems that have extensive attributes, as BeOS did, and then having applications that actually make use of those attributes.
Basically, the idea is to be able to not just organize the data, but to fully integrate that data and make it available to the user for easy manipulation. It shouldn’t matter if the relevant data is scattered over a dozen different documents and files, be they text, spreadsheet, graphics, audio, or whatever.
Again, I don’t see this as revolutionary, it’s just a matter of making the best use of already existing computer resources and information, making the computer a real tool for the user, instead of letting users be controlled and intimidated by the computer.
I think what he’s saying is interesting stuff, but there are some problems. The narrative idea is fine to a degree, but I fail to see how this will *replace* the GUI. If anything, it’s complementary to the GUI. We may well want to organize data into a chronological, linear narrative as he suggests, but we need space to look at it, most likely in a window on our screen(s).
I think some people (possibly including the guy in this article) get too hung up on the file manager and forget that it’s not the entire GUI. As it is, Windows ships with 2 distinct file managers, each of which has numerous ways of viewing and managing those files, one more certainly wouldn’t hurt, as long as it maintained compatibility with the others (in other words, don’t start reorganizing my directory structure just to view files in a different manner). I certainly don’t want my entire interface replaced with a file manager, as file management is a very small part of what I do with my computer (especially in relationship to the amount of time I spend doing any particular thing, given the amount of time I spend programming, listening to music, watching movies/TV shows, and playing games, just on my computer).
As for the desktop metaphor, I believe that
1) it could be implemented in a much better way than available today – and I mean Windows, KDE/Gnome AND MacOS (and AmigaOS and…). Many things in today’s GUI are severly sub-par. I won’t elaborate, people tend to get religious about this.
2) the desktop metaphor is not the end-all, be-all. It has been evoluted to a point where it’s probably the best metaphor we have today, but IMHO a thing that *isn’t* a desktop should not *pretend* to be a desktop. OO programmers might see the fundamental flaw in this.
As for those who cry for code examples… get real. When someone talks about the next generation transportation system in theory, destined to solve logistical and enviromental problems of the present, and you asked me about taking a ride “for testing”, I would call you a fool.
As for those who cry for code examples… get real. When someone talks about the next generation transportation system in theory, destined to solve logistical and enviromental problems of the present, and you asked me about taking a ride “for testing”, I would call you a fool.
and if you’re still talking about it in 6 years without any research or development, who’s the fool?
That being said, the product does exist, unless they’ve put up some good smoke and mirrors. Do a quick search on the guy’s company and you’ll find that they sell ‘Scopeware’ (the product discussed in the article) through various resellers as a business knowledge management system. Unfortunately, they don’t have a trial beyond the demo on the site or anything really useful for the desktop user. He seems to like to talk about the software as if it’ll be a desktop application, but that’s not the way they market it.