November 20, 2001
6:57 AM   Subscribe

While the father of Visual Basic tells Microsoft to: "abandon the browser", a multitude of researchers are looking into next-generation interfaces (links via Joel on Software and Wired respectively). Is the software industry even close to moving on to a new paradigm?
posted by costas (26 comments total)

Is the software industry even close to moving on to a new paradigm?

No, the industry is not even in the ballpark. Web pages are an easy paradigm to understand because we conceive of them as being similar to printed pages... as documents.

A new technology to replace web sites will have to be at least 110% better than the web. If or when such a paradigm is created, even theorized, it will be big news. Dunno about you, but I haven't seen a champion come forward yet.
posted by fleener at 7:17 AM on November 20, 2001

bring on the metaverse. out with the desktop, and in with an interface that's more like.... a room, or a house, or something. then I can remember that I left *that* mp3 file in the CD player downstairs, and *that* bookmark to a good literature site is on my bookshelf, in the literature section.
posted by kv at 7:24 AM on November 20, 2001

and just what is it that Cooper thinks should replace the browser? I read that twice, and he's all gung-ho about how stupid it is and how if the people in the industry would realize that nothing would stop them from doing something better, but he doesn't offer the slightest clue about what that would be.... Could he be thinking about something like Java?
posted by mattpfeff at 7:35 AM on November 20, 2001

I have thought about next-gen interfaces on-and-off for years now --just as a user mind you. I am a data-miner and a database guy in real life, so naturally, what I think would make most sense would be a "browsable" view of a database.

Interestingly, MS seems to agree. Back in '95 or so, they were going to introduce NT4, aka Cairo, that would have been all object-oriented, and metadata-rich. It never happened (NT 4.0 was just 3.51+IE 4). However, you can see them moving slowly forward: Exchange, Access and NTFS 6 are getting the SQL server back end for data management (go search for OOFS or Cairo on The Register for some background).

Essentially, every time you open a file on Windows 2002, you will be accessing a database. Even now, if you go on a 2K/XP machine and right-click on a file, then Properties->Summary->Advanced, you get a wealth of meta-data, mostly user-customizable. Windows is only a version or two away from allowing you to access the filesystem as a DB: essentially, just as you do know, but with the possibility of files belonging to different directories, or with directories not being real collections of objects, but views of collections: i.e. a file could belong to >1 directories, simultaneously, and directories in effect become queries.

I don't know if this is "the next best thing" or even if it will work in practice. What I do know is that I have a harder time managing my 1GB of so of private files, music, etc than my customers do of managing TBs of data: our tools on the PC suck.
posted by costas at 7:45 AM on November 20, 2001

The "browser is dead" chunk of the Cooper interview is the part that leaves me lukewarm--the rest of that interview is pretty good. Two good links with good reading!
posted by gimonca at 7:47 AM on November 20, 2001

and just what is it that Cooper thinks should replace the browser?

It's hard to gauge from the interview (I think there's a little bit of professional jealousy aimed at Marc Andreessen) but I suspect he's thinking about something that takes a step back from the "browser as window on the web". I'm also keeping my fingers crossed that there's a role for remembrance agents as a challenge to the hierarchical file system: ad hoc, "emergent" structures, rather than arbitrary pre-ordained ones. The desktop as Wunderkammer, if you like.
posted by holgate at 7:51 AM on November 20, 2001

Sounds like a guy trying to drum up business.

"Most of the advances that we use today aren't technological advances -- they're product innovations"
...and Cooper Interaction Design "innovates and improves software".

I like this turn:
"Product innovations generally don't come out of the lab, or from academia"
"Internet Explorer [the browser in general] is nothing more than a master's thesis program"
...and the browser's an innovation that all of us here use daily. Or is it a technological advance?
posted by ChuqD at 7:51 AM on November 20, 2001

Tech fetishist. Saying that you can't do anything on the WWW that you couldn't do on Compuserve shows just how limited this guy's vision is. By reducing all development to its technical components he misses the really interesting parts of the story. In any advance or development, the technology component is just one part of things, and usually one of the least interesting parts.

Technology alone does not and CAN not make a revolution. Technology can be the catalyzing agent, or the enabler, but they are not the change itself. Fundamental change is much broader than that, which is something that very very few technologists have ever understood. And in their lack of understanding they often disparage artists, academics, writers, and regular folks - but they are after all the ONLY ones who provide the context in which technology exists.

Plus, the artists are always ahead of the technologists anyhow. There are hundreds of writers who have made the same point as Cooper, for instance, for a dozen or more years.
posted by mikel at 8:14 AM on November 20, 2001

posted by stavrosthewonderchicken at 8:40 AM on November 20, 2001

Essentially, every time you open a file on Windows 2002, you will be accessing a database
AT LAST!!! SELECT * FROM internet WHERE picture = 'boobies'

All he's saying is that web browsers are fantastically crap interfaces for doing anything except web browsing. A logical thought, but it is about 180 degrees away from Microsoft's position 2 years ago (MSDN users will notice that those parts are now suspiciously hard to find). He's really just out to pimp .NET - specifically the WebForms stuff in Visual Basic .NET which is precisely targeted at replacing the need for a browser when writing thin client applications.
Don't worry about the paradigm shift in the web crap, Cooper was writing about something relatively mundane. It's just that the article was written for excitable IT types and our paradigms are small and easily shifted
posted by muppetfiend at 8:52 AM on November 20, 2001

Walking away from the browser in the technology biz these days is like walking away from gas-powered cars in the automobile biz.

Sadly, the few who have tried have been buried under the weight of ripped mp3s, warez and pr0n (Napster, Hotline, et. al.) that seems to follow the "un-browser" markets...
posted by teradome at 8:56 AM on November 20, 2001

Well, I'm glad to read the saner thoughts in this thread, because I too was concerned to read that I spent a lot of my time working in, on , and around an essentially "stupid program". His Compuserve statement and the caveat about the improvement in software interfaces really seem to show his bias and his argument boiling down to semantics.

So, his argument is "Despite the phenomenal success and benefits reaped by the popular use of the browser - we should discount it as being ' stupid, a thesis, built on ancient technology and merely a conglomeration of small product upgrades to interface'" - yeah that makes sense.
posted by kokogiak at 9:10 AM on November 20, 2001

First of all, let me say that I think the 'father of visual basic' should be shot, not listened to. But I have given this topic a little bit of thought.

As a programmer, a web designer, and user I'll say this: Standard "gui" and programic interfaces can have a lot more power then pure HTML+javascript. But, creating web based interfaces is a lot easier then creating a programic one. While visual basics form editors help some what, VB forms are not very dynamic unless you really do a lot of coding.

Also, with things like CSS (well, I guess just CSS) you have the ability to really make things beautiful. So, web based interfaces can be a lot prettier. Unfortunately, when this power is wielded by people who are not artistic, the results can be ugly, and when people let their artistic impulses override usability, the results can be frustrating and annoying. You have to strike a balance. The most important thing is usability. If you can make it look good, then great.

But you have to remember, programs are not documents. And just because someone knows how to use a document doesn't mean that it's the best way to present data.

And on top of that, I believe people today in the year 2001 don't need to be spoon fed metaphors in order to use computers. We all know how to use computers now.

I think the browser presented the first real way 'out' of the Mac-style GUI, because it was something people knew how to use. The biggest hurdle to a new kind of interface for the PC is getting a lot of people to use it. The Web Page interface may not have been 'better' then the GUI in terms of pure user efficiency, but it was more flexible for developers (in some ways). In order for a new interface to be widely deployed outside of interface research labs, it's going to have to be used by something popular.

AT LAST!!! SELECT * FROM internet WHERE picture = 'boobies'

<plug>heh you should check out a project i'm working on:</plug>
posted by delmoi at 9:21 AM on November 20, 2001

I think we're done.

There will (obviously) continue to be improvements to computing, but the vast majority of people plateaued on email. They simply don't have a reason to explore much deeper into the world of computers.
posted by jragon at 9:21 AM on November 20, 2001

This article was interesting to read, but I think the guy is wrong. Who wants to access their files chronologically? The desktop metaphor isn't popular because in the 70s someone had the genius to transfer the desktop paradigm to the computer. It's popular because it's how humans have organized data for the last ... oh ... 10,000 years or so.
posted by GatorDavid at 9:37 AM on November 20, 2001

but the vast majority of people plateaued on email. They simply don't have a reason to explore much deeper into the world of computers.

I dunno. My mother's working in a Windows/Office environment for the first time -- she developed her clerical skills in an office with a room-filling mainframe and vt100 terminals -- and the questions she asks about manipulating files and folders are pretty illuminating. It shows, at least, that the metaphors aren't self-evident. And from experience, I know that the interactions are different when I'm booted into Linux or Windows, and that I'm productive in different ways. And that's something that's worth further examination.
posted by holgate at 9:37 AM on November 20, 2001

As a science fiction writer, or just as an imaginative person, I've got a pretty good idea of where computing is going and where it aught to be in the future, which is more like what kv is talking about

This is pretty long and difficult to understand, since I'm typing it in pretty crappy conditions (trying to help a kid write a term paper while useing a terrible county computer) and it's not a very good explaination, hte spellchecker wont load on the crappy school line, and the idea is mostly in it's "boy this would be cool in a story" stages. Skip to the very bottom for a summary if you'd like.

I'm thinking of something I call "the overlay" which is just a matrix laid over the world, something like the internet, only 3D and superimposed on the world. To interact with this, one would wear goggles, or glasses or have implants or something, anyways, everything has it's actual presence in the world, and attributes, data and actions in the overlay. It's best to explain this as an example:

I hand you a business card, the card contains a microserver (lets say, serverette). By taking the card into your hand, your server (attached to your body, or whatever) accepts whatever data or programs are contained on the card, things that would behaive in a similar way to present day cookies. The card has information and data tethered to the card like a standard business card, only all the data exists primarily in the overlay, and by "useing" the card, you could directly contact me, connect to my server, buy my stuff, etc. If you don't want the business card, simply throwing it into an ordinary trashcan (which would be recognised by the overlay) and all the data would be deleted from your server. Since everyone would bassically be waking around as an entity in both meatspace and the overlay, each person could message other people from where ever they were, could look up anyones names and gennerally invade anyone's privacy by sending a query to the persons server.

Summary: The internet get's superimposed on reality, everybody gets to be a webpage, and everything gets wired.

That's my unrrelated ideas on future interfaces, if you have anything to add that would educate me (I'm not in-the-know in terms of hardcore backend programming and stuff, so I don't know the how-to), or you think you can pull it off somehow or something, it would be cool if you emailed me.
posted by fuq at 9:45 AM on November 20, 2001

fuq: check out augmented reality
posted by muppetfiend at 10:01 AM on November 20, 2001


i don't agree with the sentiment that the filesystem in windows (and mostly all modern operating systems) be upgradable to a database, for the simple reason that they already are databases. (databases can be bogglingly simple abstractions: a vector or a simple linked list can be thought of as a database.) windows (and other operating systems) could modify their file systems to allow files to be linked via several directories (in fact, this is currently implementable on Unix OSs -- i think later Windows versions too -- via symbolic links). i would say directories are already "views" of real collections, since the directory exists only as an abstraction of the file system, and by real collection i take that you mean the physical contents of the files themselves stored on the hard drive (the contents being seperate from the file system's innards themselves).

by the way: alan cooper does not seem very bright.

Cooper: People are waiting for the next technology to bail them out. Each successive technology, while it brings a set of solutions, also brings a set of problems. So the net gain is nothing.

wow, all the time? gee, who knew you'd get such a harmonic balance all the time when comparing problems and solutions?
posted by moz at 10:15 AM on November 20, 2001

I hand you a business card, the card contains a microserver (lets say, serverette).

Um, no. Last thing I want to see is people walking down the street throwing physical spam at me.

There is absolutely no reason to reduce the functionality of virtual objects until they are only as functional as real-world objects. Physical business cards cost money to manufacture. Virtual business cards do not. Physical business cards take up space in your pocket if you collect enough. Virtual business cards do not. Can you imagine having to carry every business card you've ever received with you in order to access their data? No? Well, if you don't take them with you, how will your personal system know that you didn't discard them when you're out of range? What if you lost some -- say they fell out of your pocket in a taxicab? Would you lose the contact data as well because your system thought you meant to give that data to the cab driver?

Carrying the metaphor too far is a common error of amateur interface designers, which is essentially what science fiction writers are. It often makes for good reading but it usually bears little relation to reality.
posted by kindall at 10:37 AM on November 20, 2001

moz: my comment was basically summing up what MS is already planning. While I agree that a filesystem is basically a dumb database, Microsoft plans to make it into a smart one: a database that's aware of file metadata and can query them really, really fast. They will do this by making the core engine of SQL Server part of the next NT filesystem; they are already moving Exchange over to the same engine.

What does this mean? It means that MS is moving towards an object-centric storage model. Instead of storing files, the filesystem will be storing objects. Yes, I know that the distinction is small, and that it's not far from the Unix paradigm of sockets/FIFOs/devices/symlinks as files.

However, while conceptually similar to the Unix "everything is a file" paradigm, this one is the inverse: instead of "dumbing down" information to make it into a file, or bytestream, an OOFS-based OS (really, an object-based *network*) adds data (meta-data) to information, rather than stripping it away.

This allows the OS to 'forget' a lot of stuff, like hierarchical file systems (why categorize based on an arbitrary file structure when you can categorize by author, date, company, revision info, etc, etc?), application-associations (an object can be introspected and an application that can 'play' with it can be called on demand, rather than relying on extensions or resource forks), etc. Basically, data moves to the foreground, while applications move to the background. In MS' future you will manipulate data, not programs. Something very convenient for the holder of all the standard business fileformats, BTW.

I am not saying this is the best solution ever. But it will certainly change the way we work with computers in the long run. If people buy into it. I mean, how many of you actively set the extended metadata in windows (right click->Properties->Summary in 2K)? How many know that by doing this, you can make the Windows Indexing Service and filesystem searches a lot cleaner? How many have right clicked on the Explorer Column headings (Name, Size, Date, etc in the "Details" view) and saw that you can add this extended meta-data to the Explorer view?

PS: Yes, OS/2 had this a long-time ago. Yes, NeXT had something similar and OSX is going that way, albeit in a more bolted-on fashion. BeOS had even more meta-madness. But... The fact that the dominant sotware producer is doing it though makes it way more significant.
posted by costas at 10:44 AM on November 20, 2001

What's wrong with the net as it is? I think it's perfectly fine, does everything I want to, perhaps if the community part was implemented a bit more, like remember alexia? where each site had a chat room? If that was developed a bit more, be more automatic or what not.

What fuq explained you can experience in massive online rpgs, like Dark Age of Camelot, you walk around and can collect data on everything, you get messages from people and so on. One thing that I would find really annoying, if the interenet was superimposed over the real world would be talking to a person in real life and then having an instant messenger pop up in front of me, asking to chat, and then having to talk to two people and then if more and so on and so on.

I don't even think you could superimpose the internet over the real world. What would be the point? Isn't it already? with wireless pdas and such? I think what I'm getting at is the written word as well as a page have been around for a long time, I don't think there really is a better way of transfering and storing information, otherwise it would loose meaning.
posted by tiaka at 10:46 AM on November 20, 2001


While I agree that a filesystem is basically a dumb database, Microsoft plans to make it into a smart one: a database that's aware of file metadata and can query them really, really fast.

the file system is already aware of the standard file metadata (permissions, etc.), since that is all stored with the pointer to the file data in the file's entry to the file system. as for query speed, that would seem to improve find operations -- i hardly ever use the find-files feature of windows or sherlock on macintosh, so i can't say it'd be all that useful for me, but what the heck.

However, while conceptually similar to the Unix "everything is a file" paradigm, this one is the inverse: instead of "dumbing down" information to make it into a file, or bytestream, an OOFS-based OS (really, an object-based *network*) adds data (meta-data) to information, rather than stripping it away.

this doesn't sound all that revolutionary, to me; we already store file metadata in operating systems, but usually as other files or in the kernel's memory and loaded at bootstrap. (e.g.: the IP configuration for a device in unix may be stored in a file in the /etc directory.)

can i add fields to the file tables? i would assume that is how directory-like functionality would be done. associating files with several categories via field values seems interesting, at least.

Basically, data moves to the foreground, while applications move to the background. In MS' future you will manipulate data, not programs.

this seems partly true already with file extensions, as you mention. however, i question the notion that data should be paramount; often times we run programs which have no prior interaction with files that we use. Internet explorer may load configuration files, but we don't click on them to load the program. there is little stored data that needs to be referenced directly by us for that application to be used. will this change -- will we have to click on data to use IE? i don't think so, because i don't think it'd be a useful shift in paradigm.

metadata is a good idea, i agree, but i'm not really convinced that data over application is the best way to go.
posted by moz at 11:10 AM on November 20, 2001

moz: If I get this "MS future" correctly, in that world, you will not have any applications. Just swarms of components that could work on objects, probably >1 at the same time. E.g. a Word document today can contain Excel charts. In that world (I think the MS codename is NetDocs), you will not open Word, or Excel or Outlook. You will instead open the document. The OS will detect that there are word-processing objects and charting objects within this document and present you with an interface that can work on both. Not Word or Excel but a "canvas" for components from both.

Most likely, if you don't have the application/components needed to see everything in the document, NetDocs will go on .NET and grab them for you, likely for a fee. If your device can't handle that component (a cellphone opening a big Excel spreadsheet, e.g.) it will likely default to viewing the object that is being opened remotely (on some big MS server somewhere, or probably just your company's/ISP's server). It's a compelling, seamless vision of data-centric computing, everywhere. If the implementation doesn't suck, it will rock.

This is hardly innovative. Others have tried and failed. Apple with forks, BeOS with BeFS, etc. The main problem has been what you alluded to: interoperability with other systems (Mac users know what I am talking about...). However, when the force behind 90% of the PCs out there is behind this, and when they are promising a cross platform layer for object distribution (.NET) it becomes possible.

BTW, you're correct about the filesystem storing metadata: however, it can only store certain, file-specific metadata. Windows 2000 and later can store extended metadata, defined by the user or the application. That's where all the power is. For a much better description, check out this excellent article on Ars Technica about the Mac approach to metadata.
posted by costas at 11:47 AM on November 20, 2001

again, i question whether or not data-over-application will work in the long run. for one thing, it'll freak some people out (oh my gosh!! will my wife see all of these "SUPER ANAL HARDCORE XXX" site files for IE??) to have data storage required for the smallest of tasks such as web browsing, and i'm not talking about cache. i'm not saying it wouldn't be at all useful; but it won't make our lives as simple as we think.
posted by moz at 12:04 PM on November 20, 2001

just saw this on boingboing, it's like the ICQ of HTML or something. shades of martin wattenberg.

also there was a Q&A with david gelernter on feed awhile back.
posted by kliuless at 3:24 PM on November 20, 2001

« Older There's a fungus among us...  |  Inside the Vatican... Newer »

This thread has been archived and is closed to new comments