Wednesday, April 15, 2009

A more profound Enlightenment

I've been following the Enlightenment project for years, always impressed by the strength of vision and dedication of the developers. Every once in a while I take an inventory of the graphical toolkits out there and am always disappointed by the fact that the EFL is the most progressive desktop gui system out there and yet hasn't really broken into the mainstream. All the other GUIs (QT, GTK, Windows) are built around boring components (boxes, pull-downs, radio, etc.) The concepts behind these mainstream toolkits are decades old.

When I look at the web, I see all of the excitement about Web 2.0 and the Cloud and "Linked Data". But it's all branded and contained within different application spaces. You go to GMail to access your contacts and send an SMS. You go to Facebook to update your status. You also have your status to set on GMail, AIM, and every other application. These things are really just implementations of a concept. This is especially confusing to less intuitive computer users. Users have to learn a ridiculous vocabulary to do things they already naturally understand: (eMail, Instant Message, AIM, GMail, Yahoo, Facebook, MySpace). Instead users should just have to think "I want to send this to Bob" (Email/IM) or I want to tell everyone who cares about me something (Post a Status). As the features these companies offer all coalesce, one wonders why we need to be branded at all? Why not just standardize on these features and give users more intimate access to them through their own computer.

To make things worse, this is all implemented on a HTML/Javascript layer that was never designed for it. Developers have to wrestle with browser eccentricities and code hackery becomes a necessary part of the development cycle. Instead of looking for a better platform than the browser people have over-inflated its ego (and purpose) and made plugins for it. Now we have Ubiquity, a great idea built on the wrong platform. Rather than take a step back and design a new standard rendering layer we now have Flash, Silverlight, and JavaFX. The browser is tired and overloaded, it's laden with features that belong on your desktop, not next to your web page.

If you've read sci-fi, or watched movies like Minority Report, you know what could be possible. Direct meaningful interaction with visual representations of data. I think now is the point in computer history where that vision can actually become a reality. It's simply the intersection of the browser, your desktop, the Web 2.0 services, and personal management tools like OmniFocus or mind-mapping tools (Freemind, Xmind, NovaMind, etc).

So let me try to explain this idea more concretely...

Imagine your desktop as a space with context. When you start working on a project, you create a new space/desktop for it. As you open files/email/urls it all gets associated with this context. When you decide to work an another project, you'll close this space. Later you come back to it, and everything is as you left it. When you search your computer you can search within a space or all spaces, and move or link things between spaces. A desktop will dynamically adjust to the contents. If you have 3 pictures you're working with, they'll just be thumbnails. If you're working with 1000 pictures, they'll be abstracted as a list that you can manipulate.

Now imagine that all of these things you work with have meta data and tools associated with them. Your computer has a hierarchy of objects and tools. For example, a picture can be scaled, rotated, color filtered etc. Text can have different fonts, colors, be translated. These tools are really just simple programs or scripts that are visually abstracted. Eventually there might be a database of tools you could download for different purposes. This is one of the more difficult components to design well, but I think it can be done.

Within a space you can create selections of different objects and save the selection. Once you have a selection you can act on it in different ways. You can act on their common properties. So since all objects have a creation date, you can sort by creation date. If they're pictures, you could rotate all of them.

Now expand your concept of desktop objects. Not only can they be files, but they can be objects from a database or a website. They might be widgets like you would see on any of the portals (Google, Yahoo, etc) or desktops (Google Desktop, Gnome/KDE/E widgets). They might even be objects from the local database.

Any of these objects can be acted on in certain ways. You can annotate, tag, categorize, or set a due date on them. You can also create basic elements and combine them. Rather than fire-up gEdit to take some quick notes, you just start typing notes on the desktop. You can tag these notes or set due dates for them, and they become todo items. You can type some text and then start formatting it. Then convert it to HTML or a Word Document or whatever.

If you're still with me you have some kind of image of a desktop that understands many kinds of files and data objects and can represent them visually. A desktop that might look something like what you see in sci-fi movies where you can visually drill down, make selections, apply operations, etc.

Imagine that you have a list of contacts that's deeply integrated with this desktop environment. When you open a message from someone on a space, the attachments can be moved onto your space and be manipulated as objects, you never have to open a save dialog. Also, the person becomes associated with the current context. These contacts have email accounts, im accounts, facebook accounts, etc, but you don't really care about that. You never open an email or instant message client. You simply get messages from the person and send messages. If the person is currently online through an im service the message is sent with that method. You can drag any object onto a message. The computer intelligently translates the data. So if it's a selection of rows from a database, it inlines it in the email as an html table. There's no new data formats or apis, when you get an email with an html table in it, you can drag it out into your space and manipulate it and break it apart. Imagine the power of this kind of easy communication and imagine that every object can be sent to someone. If an object is "online" it will give the person a link, if it's small it may inline it. In addition, you could invite people to collaborate on your space while you work on it. Or maybe work with online spaces stored elsewhere?

I'm also thinking that this kind of computer environment would replace a certain amount of work done with data mining and database tools. I don't do much consulting work but I've encountered a few people that had a conceptually simple problem that required a database but the tools were just too hard to use and maintain. They really just needed a few tables with very simple associations. Imagine if you could just visually setup a database and define its entities and then manipulate and search it just like any other object on the desktop. Instead of developing custom reports for every database, you give people the availability to create tables and charts using any kind of object. Say you select a list of pictures, you can then graph the picture dates on a time line. Say you have rows from a database, you can do charts with the measures found within that data.

Hopefully you understand what I'm getting at. The paradigm shift is huge, but I think it's the way computers need to go. Think about it, what does your email client or instant messaging client really give you. Aren't they just different interfaces around the same fundamental concept? If your computer was effective at organizing and archiving your email why would you even use an email client application? I believe this type of integrated desktop would completely replace your need for separate email, instant message, task management, photo management (Picasa) applications.

I've been thinking about this project for a long time now and would like to begin serious work on it. I'd like to create a collaboration space, gather interest, and start documenting more concrete ideas and organize all of this into realistic releases and milestones. I plan to start coding in 2010 after I've had a chance to talk to different experts and design the main concepts. I'm curious though if this could become the Enlightenment Desktop or maybe the next release (0.18 or maybe 1.0)?

If the developers are not interested in this vision than I'll probably just start a separate Linux desktop project that uses the EFL. Maybe call it Nirvana?

Tell me what you think. Has anyone heard of similar ideas? I've tried to find projects related to this but I don't even know what to search for. It's a bit cynical but I truly believe that this is something that won't ever come out of the big companies and can only be developed through open source. Large web companies won't be interested because it essentially obsoletes most of their products. No need for Google Docs, GMail, Picasa, Yahoo's Portal/email, etc. I'm not sure Microsoft or Apple has the vision or desire either.

Also, tell me if I'm totally crazy or not. Do some of you think about these concepts too?

No comments: