iMarc | Interactive Media Architects
  • Portfolio
  • Process
  • About
  • Communiqué
  • Contact
  • Support
  • Search

Down with the Web.

by Robert Mohns - February 16, 2006 / 11:54am View more articles

(Part 1 of 2)

Let's cut the to chase: The web is a 1960's era IBM mainframe, and it sucks.

Mainframe interfaces are characterized by a query-response mechanism. Your terminal requests a starting point and a screen is sent to your terminal — the terminal being your remote interface to the computer. You read the screen, enter data in a few fields, then send the screen back to the mainframe. The mainframe processes it, then sends you a new screen. Lather, rinse, repeat. It was little more than a paper-based system running at 300 baud.

By the 1970's, the minicomputer, pioneered in the marketplace by DEC, had changed the face of computing by selling computers with interactive terminals. Instead of exchanging screens, the user's terminal was an actual process on the minicomputer. It was still limited to text (for the most part), but you had a realtime access to the computing resources. This turned out to be a profound change. Instead of filling out forms, submitting them to "the office", and waiting for a response, you actually manipulated data directly. The paper-style interface was replaced with fast command/response cycles, enabling users to work more quickly and flexibly. Interactivity at 9600 baud.

By the 1980's, personal computers had appeared on the scene. Like the much more expensive minicomputers, they had interactive text interfaces. Unlike the minicomputers, 100% of their capacity was devoted to one user. You worked directly with the data — a first, outside the halls of academia and obscure corporate research facilities. And for the past two decades, we have worked with data ever more directly and intimately. Personal empowerment as fast as your hard drive can spin.

Then the web came along. You start by opening a terminal — although we call it a "web browser" now. It opens a "home page". You read the screen, maybe type in a few web form fields, then click "submit" or click a link. The request is sent back to the web server. The web server processes it, then sends you a new screen. Lather, rinse, repeat. It's little more than a paper-based system running over a T1.

It's a very fancy mainframe terminal. It has colors and pictures and different typefaces, and sounds, and rollover effects, and Flash geegaws and doodads, and all of that fools us into thinking we're looking at an interactive system. But we're not. We're looking at a forty year old mainframe interface that's been worked over by a graphic designer with an XGA screen.

The web sucks. Down with the web.

(Coming real soon now: Part 2)

More Articles Get the RSS Feed Post A Comment

5 Comments

by Craig Henry   #
on February 17, 2006 / 10:34am
So what are you suggesting one does with this "web" now that it's been established? By "down" do you mean "get rid of"? Because, if so - aren't you really saying "Down with Robert's job"?
by Stefan Riss   #
on February 17, 2006 / 11:52am
I have to agree with Craig: where would you like to go from here? I hope that part 2 sheds some light on that. But for now I think that the basic communication model (request/response) is about the only thing the web really has in common with the old mainframe architecture. More and more web(based)services are being integrated in our desktop/personal computer environments and then there's "Web 2.0". Though far from being the holy grail, by allowing event based interaction and partial page updates, it is the first step beyond the request/response paradigm. And I think it's a step in the right direction.
Another difference is that in the old mainframe days (which really aren't over yet), afaik, communication was more or less with only one dedicated server. The web allows you to access ressources from all over the planet - that the same principle still works here, speaks *for* the principle, not against it.
The main problem I think is that we are still still sitting in front of a screen and using a keyboard - but you can hardly blame the web for that.

Cheers,
Stefan
by Bill Bushee   #
on February 17, 2006 / 5:44pm
I'm quite curious to see where Robert takes us with part 2. Discarding the old model and creating a brave new world of the web could open up tremendous possiblities, but there are inherent road blocks to getting there.
by Andrew Stanley   #
on February 22, 2006 / 3:25pm
I think Robert hits all technologists where it hurts: we cannot break away from the Gutenberg Curse - the curse of paper, print, and books. Technology has redefined, in so many ways, the way we define our experiences and understand the immense data sets that fill our world. However, it is only in recent years that is has even begun to break free from the archetypes of written language. This mainframe world that Robert speaks of is a textual, screen and paper driven world that is bound by the Gutenberg Curse.

Technologies as simple as RIM's "Push" technology have redefined email from "at-your-workstation official communication" to "pervasive instant communication" that is in may cases closer to speech (or Instant Messaging) than email has been. This is a step - redefining the way we interconnect, the immediacy with which we communicate and how we expect to keep in touch with our world.

But web-centric technologies and paradigms have continually held dear to the form of the book - even blogs (the "new media" to some observers) are just aggregations of notes and conversations in digital space. We have yet to come up with a content distribution system that can enable thoughts, ideas, images, and the written word to meld together in the way these things do in real space. Typing will never be as visceral as speech; nor will reading be as visceral as listening. To paraphrase Lewis Black, we listen with our ears, not our eyes.

Interfaces and aggregation systems that use the amazing back-end technologies the web has fostered but allow us users to interact via our visceral, primary senses are where the Web needs to go - connecting us viscerally, intellectually, and emotionally.
by jon   #
on March 10, 2006 / 1:36pm
Forgive me for not being shocking. The first thing that came to mind after finishing your article was, well, "duh".

The reason the original query/response model was inefficient is because it was *SLOW*. Whether I'm working on a local set of data, or the data on the "mainframe" (database server, etc) is irrelevent to me as long as its *FAST*.

With DHTML (javascript/css) much more of the application logic can be processed directly in the web browser, with minimal trips to the server.

For example - Writely. The only time you need to make a trip to the server, is to save or load a new working copy of the data. How is that different than any other piece of word processing software where the document is saved on a fileserver? Oh, that's right - it isn't. The only difference is the communication protocol used to transfer those documents (NFS/CIFS/etc vs. HTTP).

I fail to see why this architectural model is so poor? You know, the Gregorian Calendar was invented in the 15th century, but it seems to be working pretty will, with only minor modifications (analogous to the the graphics on the web).

Comments have been turned off on this blog.
Read something more recent.

Statements and opinions expressed in this blog and any comments made are the private opinions of the respective poster, and, as such, iMarc LLC is neither responsible nor liable for such content.

iMarc

iMarc is a web development company in Newburyport, MA. This is our blog.
View all blogs or learn more about iMarc.

About the Author

Robert's Head Robert Mohns, Information Architect
Robert helps iMarc's clients go from business needs to web architecture and specifications.
More blogs by Robert

Search Our Blog

Recent Communiqués

  • Lunchroom Banter (Volume XVIII)
  • Building a Museum Website
  • iMarc is 11 years old today
  • Pre-holiday hijinks
  • Look and Feel
  • Gardening and Web Writing
  • Trusting Experts, Part 1
  • USC Trojans vs Oregon State Beavers
  • On My Days Off
  • Make your site 52.676% more cool
  • iMarc Grows (again)
  • Laying the subfloor
  • Inc. 5000
  • Say Hello to our new Sys Admin
  • "Not Invented Here" for a Triple Word Score

Popular Communiqués

  • Building a Museum Website
  • iMarc Grows (again)
  • iMarc is 11 years old today
  • Pre-holiday hijinks
  • Laying the subfloor
  • Lunchroom Banter (Volume XVIII)
  • Trusting Experts, Part 1
  • Look and Feel
  • Make your site 52.676% more cool
  • Inc. 5000

Recent Comments

  • Lunchroom Banter (Volume XVIII)

    By Patrick McPhail: What you'll notice is that in both scenarios I suffer.

  • iMarc is 11 years old today

    By soso: happy life day

  • iMarc is 11 years old today

    By Nick: When a company has a birthday, does that mean the CEO's had sex? …

  • iMarc is 11 years old today

    By Marino: Happy Birthday!

  • iMarc is 11 years old today

    By Josh Cole: Happy Birthday! Can't you use both your fingers and toes to count?

RSS

RSS Icon Learn about RSS and get the feed for our blog.

About iMarc

  • We build custom web sites
  • In-house strategy, design, programming, hosting
  • In business since 1997
  • We’re located in Newburyport, MA
  • Call us at (978) 462-8848

© 2008 iMarc LLC, Contact Us

Links

  • Home
  • Portfolio
  • Client Support
  • Log In
  • (icon)RSS

Meet the Team

Bill's Head Bill Bushee, Senior Developer

I handle all aspects of building a database driven dynamic website. This involves creating the initial pages and web forms in HTML, designing the database structure, and writing all of the appropriate server side scripts to handle data and manage the site's interactive features.

Learn More | Meet the Others