Teqlo - A Preview

UPDATE

The crack Marketing Team (a.k.a. Rod) made a terrific teqlo variant of the The Machine is Us/ing Us video showing how Teqlo works. (see the original here, more interesting than the YouTube version, BTW)

ORIGINAL POST

The Teqlo team has decided it is time to provide a sneak peak to show the world what they are about.  So Teqlo.com is open for preview. But before you go there, let me set  a few expectations.  This is a live, production version of the technology, but this is not a demonstration of the business.  You are going to go there and say "why would I use this routinely?"  The answer is you won't. We don't expect you to, and this is not the way we expect average people to use this technology.  We are intentionally exposing the platform part of the technology to test our ability to scale the engine as people bang on it. 

This release is a demonstration of the engine and the technology is not trivial, even though the initial use case looks like it is.   Let me illustrate with simple walk through. 

When you sign up on Teqlo, you'll get a desktop that looks a bit like some more mature products like Netvibes. This is just to give you a home page - a working canvas.  The power of Teqlo is in the Applications and in the Builder.   There is really only one Application on Teqlo.com today - Leads and Calls.  It lets you search Linkedin and DabbleDB to find people and save their names in a contact list.

The magic of Teqlo is behind the scenes and you can see a view of it in Builder.   The Builder is the work area where YOU can build your own applications from widgets, provided the widgets represent real applications like Google Calendar and not just simple RSS feeds.

Below is a series of pictures to illustrate the point.  I built a simple Application that
1. Searches Ebay for an arbitrary product (I like 4x5 cameras; so I searched for Linhof brand)
2. Lets me select the ones I want to track and then Teqlo...
3. Automatically puts them in my Google Calendar
4. Automatically puts them on a list to save for later
5. Automatically maps them on Google.

Here is the result

Now this is trivial example, but it illustrates some interesting things. First, if you go into the Builder (tab on the menu bar), you can pick from a few widgets that are in our catalog and you place them on a canvas. See the next picture.

The real key to building this little application is in what we call the Interactions Editor.  The IE (ick, we need a better name than this), allows you to say which interactions you want between any two widgets.  The IE has to know what actions and reactions are permissable for any widget and it has to validate that these two widgets can share data, i.e., share the same data microformat or other representation.  Teqlo automatically routes data between applications continuously as the data change.  So as you actually interact with the Application, such as de-select an entry, it automatically updates the interactions.

This is what's known as non-procedural programming.  The order of the rules in the Interaction Editor doesn't matter.  There is no control logic at all.  Teqlo flows the data to wherever it needs to be.

This is what makesTeqlo different. This "Application" was assembled without any javascript or any other programming by me. Granted, it is not super powerful.  But this is a function of the limited widget library, not the approach.

By the way, to give you a hint of where things are going, take note of the fact that you can have multiple "canvases."  What is a canvas? It is a page or a part of a page.  So you can have real-time applications that render different views to different users on different pages anywhere on the Web, all simultaneously.  You update the search list on Ebay and it can update a calendar (or something more interesting) on a page that I am looking at in real-time.  "Web as a platform" just moved from a Web 2.0 slogan to something concrete.  Think messaging; think feeds and events; think true collaboration.

Stay tuned.  Teqlo looks like a mashup tool today, but we have more interesting, larger, and more focused intentions.  As I said, this release is just a proof point.

Comments

[Don Geddis]

"This "Application" was assembled without any javascript or any other programming by me. Granted, it is not super powerful. But this is a function of the limited widget library, not the approach."

But surely you're not claiming that non-procedural programming has the same level of computational power as writing real programs.

This is what makes the extrapolation difficult. You present a simple example, then claim "but this is only simple, it can do MUCH more", and by connotation suggest anything that can be programmed. But surely that's too high a goal to actually achieve.

I'm having trouble conceptualizing the space of possibilities that you (or Teqlo) have in mind. Something more than mashups, yet less than real programming. I wonder where the boundaries are.

[Peter Rip]

No, you're right we are not. We are not claiming anything like that.

This is NOT a general purpose approach to real programming. But it IS a good solution for configuring simple 'cut/paste-type' interactions between web services (like the demo) and for configuring more complex real applications that invoke multiple web services. There are a lot of application where the user wants the "mostly" the functionality in the application, but wants it to behave slightly differently, e.g. use a Yahoo map rather than Google, use these 4 public data sources plus this one proprietary one, only expose data from this web service to these users (like payroll data or sales data). These kinds of variations on theme is where Teqlo has a sweetest spot. It is not for the theme itself, just the variations.

[jccodez]

Just a simple question that I am sure you as an esteemed vc would ask:

1.) what is the business model?

2.) what problem does it solve?

3.) what is your unfair advantage?

It looks a bit too "boil the ocean"

[Peter Rip]

These are fair questions. But as I said, these are questions we are not going to answer at this point. We *have* answers. But this release is simply a stress test on the technology, not a preview of the business. But the business is not an ocean boiler. We have identified a specific, addressable target user who has a real pain point that we can solve. Our unfair advantage comes from the IP we have developed - 2 issued patents. We don't have an unfair distribution advantage, yet, but we can be efficient in that regard because of our focus.

Give us a few more months and the business plan will emerge from the mist.

[Gary Culliss]

Sounds a bit like the desktop-based Proto at www.protosw.com.

[Peter Rip]

No, Gary, it's not, for two reasons. The tech is completely different and the business is not a tools play. I know all these mashup things run together at the geekspeak level. So the perception is understandable.

Again, this preview is just to debug the platform, not to preview a tool.

[Radha Popuri]

Seems to me that you, Dion and a number of enterprise 3.0 folks seem to have bought this concept of loose integration at the edges.

Seems fine to me. The only question is if you plan to have a way in the future for these loose integration edges to be fed back into the enterprise in a much more formal BPEL+SCA or JBI models.

[Elliot Turner]

Loved the video -- you guys are doing some interesting stuff!