Thursday, October 27, 2005

MiteSite Chat (Tiny AJAX Chat...)

Here's the first MiteSite - MiteSite Chat. Consider it beta and it may work / fail / be gone at any given moment. I've been playing around with this for the past week. It's interesting because there is nothing better than chatting to generate lots of dynamic data. This will be an excellent test of our server framework that Alcea Fast BugTrack and Alcea Fast IssueTrack are built on. There is almost certainly going to be browser compatibility issues. I don't have a Mac lying around, and suspect it won't work. I'll be grabbing Opera in the morning for a quick test there.



This is in a skyscraper sized space, but could be reconfigured for large banner space.



Hoping to get the team to help me beautify this a bit.


Note the icon in the left allows for the MiteSite to be "torn-off" so that the user can browse to a different site, while continuing to interact with the MiteSite.


I'll be on and off this the next couple of days. Would be interested in knowing peoples thoughts.

Thursday, October 20, 2005

Mitesite: A website or webapp embedded in an ad-sized space.

[OK - I had come up with the name parvosite without talking to anyone ... especially dog owners - so the team here talked, and the new and improved named is MiteSite - I'm editing this entry to change it...]

Years ago, I worked at a company called Spidertop. The founders had an interesting idea that they called a StickyAd. The idea being that an entire website could be embedded in a banner sized space. Imagine presenting users with a live view of auctions ending on Ebay in the next ten minutes. Now imagine allowing a user to actually bid on and complete the transaction, all without leaving the site in which they saw the banner. It was an interesting idea that was ahead of its time. The interactivity had to be delivered through a Java applet. You don’t see many applets on the Internet today – and you don’t see an advertisements delivered through them – and the reason is that it takes 5 or 10 seconds for the applet to launch within a browser – and sites don’t want their users browsers frozen just to display an ad. The applet had some interesting functionality – it contained a micro handwritten xml parser, and a small interpreter that interpreted a bean scripting language that ended up being very similar to IBM’s bean markup language (BML). The applet contained the Spidertop server, which could send dynamic commands, including images. While at the company, I met with VCs in New York, and advertising companies in New York and Chicago. Pretty interesting time, but the company went under, owing my consulting company (which is essentially myself and my business partners) a good amount of money, and leaving about 10 people in the lurch, but these things happen.

Working on the Ning beta, and lightly blogging for the past year, I’ve been mulling over some viral ideas. Today I noticed that one of the top sites being linked to according to bloglines is QuizGalary – I try out a quiz: “What is my darkest secret?” Type in my name, and it comes back and tells me that I shared gum with a friend. Gee – how did it know? From a technical standpoint, I’m not really overwhelmed – but hey I do like the idea that right on the page, here’s the HTML code to jump to the quiz from your blog - go put it in there now, and instantly, a few hundred people have put this on their site.

I’m hoping to get something rough put together in the next couple of days that I think will be pretty cool. Just need to find the strength to pull an all-nighter. I believe that now, with AJAX capabilities in most browsers, that pretty compelling MiteSites can be built into a banner sized space. A mitesite can get new content while not affecting any other content on the page. What else can you do with a mitesite? They can include icons to allow them to be expanded for additional functionality. They can be popped-out – coming into their own browser window so that if I’m in the middle of browsing Ebay, and I’m finished reading Slashdot, I can move off with my primary browser without losing my spot in my mitesite.

Hoping to get something going pretty quickly, and then look at making it viral once I’ve worked out the bugs. Stay tuned.

Tuesday, October 04, 2005

24 Hour Laundry (24HL) Uncovered

I had the good fortune to be selected as a beta-tester of 24 hour laundry just a few months ago. If you haven’t heard of this organization, 24HL’s co-founders include Marc Andreesen and there was a great deal of speculation about the organization not too long ago. Their product / tool / concept: Ning.

So what is it? It is a development / computing platform for delivering social networking applications (examples of social networking applications are Flickr, ePinions, etc.) What is the best way to generate a lot of content? Create it yourself? Nope – just create a Flickr and let the unwashed masses create millions of pieces of content for you. How do you create and index all content? Let unwashed masses of developers create web apps that the truly unwashed masses then go to for content creation. As a developer I can create any content type (restaurant, photo, pet, piece of electronics, hotel, anything) – then I can create insert/update screens for all to use and share. All content is taggable / shareable / searchable - very sweet.

I believe that 24HL has the potential to be the CPAN (Comprehensive Perl Archive Network – basically the organization that manages just about all Perl libraries) of networked apps – and not just an archive, but the platform on which these are actually deployed. If you see an application (not content – an actual complete running web application) you can clone this application (creating a copy in your own development workspace) – customize, improve, activate – and you’ve now created your own Flickr.

During the beta I wasn’t certain if the 24HL team even understood the ramifications of what they had built – it could also be considered a real-world topcoder platform – build the best app you can, someone refines it and makes it better, repeat indefinitely. In the past, I believed that one of the biggest factors stopping the advancement of mainstream web applications and web services was that the average kid / developer doesn’t have access to the computing resources, programming libraries, etc that only the best funded startups might have access to. But 24HL removes this barrier to entry and now makes this a commodity.

What could be really cool about this? The ability to trace the genealogy of a piece of code. I write an AJAX chat client and web service (did I mention that they have designed Ning so that you can deliver not just HTML, but XML content), someone takes it and makes it something else and so on, until someone takes it and builds the next Flickr out of it… Cool. While seeing the genealogy of the code is not yet possible today – the folks there have given this a great deal of thought, and I have no doubt that what they have created has the possibility to change the framework in which web applications are created and delivered.

Downsides? I’m a Java developer, and the primary interface for Ning is PHP – but, PHP is not rocket science, and soon there will be thousands of sample applications from which to build your own idea.