Friday, September 30, 2005

Rich browser applications: email example

Although browsers have supported javascript for years, it's only recently (in the last year or so) that powerful/rich applications have become readilyavailable.

This screencast describes Zimbra, an open source rich browser based email client. It's worth looking at, not least to see how it compares to the OU First Class conferencing system. It would be even more interesting if the folks at FirstClass took a look at the AJAX client...

If you're interested in how traditional email clients may be extended with a tasteful, added value pop-up windowsd, watch the screencast...
If you're interested in how a screencast can be used to communicate an overview of how to use an application, watch the screencast.
If you're interested in the sorts of things you can do in a modern browser, watch the screencast.
If you have too much time on your hands ;-), watch the screencast...

For technical info, there's a couple of white papers, one about the system architecture and one about the AJAX Toolkit underpinning it all.

Friday, September 23, 2005

Separating Navigation from Content and Style?

With the latest A9 Toolbar, website designers now have an opportunity to embed site navigation information into a drop down menu from the toolbar by publishing a descriptive siteinfo.xml document at the root of their site.

The sitinfo document allows you to insert cascading menus, and site search links, into the toolbar dymanically, so whenever a user is visiting the site, the toolbar can provide a navigational system alongside any navigaiton embedded within the page.

For example, here's a snapshot of a quick test of a possible structure for some of my OUseful pages:


Now as it stands, I'm not sure how useful this is going to be. Site designers typically like to embed navigation within a browser page, conforming to colour themes and site style, and so on. Indeed, the navigational style and structure can often be one of thedefining features of a site.

Similarly for users - when they visit a site they typically look within the main window for site related information. Although it is increasingly possible for site designers to make use of browser control buttons, and exploit any installed toolbars (as A9 are hoping to do), for most users their focus of attention is within th main window.

That said, the sitinfo looks interesting in terms of the Skeletal Web. I first doodle about this when Google announced their Sitemap, suggesting that the time had possibly come for separating navigation out from content and style.

The Google sitemap definition, however, does not fit that purpose so well - it's more of a brute force technology for helping search engines locate searchable pages within a website. (Then again, it would have been interesting if the A9 siteinfo definition could have accommodated the Google sitemap structure somehow!)

With the siteinfo document, however, the emphasis is on defining navigational structures that are directed towards the user. (Again, the Google sitemap reflects the simplicity of Google search - all they want is lists of URLs, searchable by keywords). And this is where it might get interesting...

For example, consider the following:
  • The Skeletal Web Pages are published without style, e.g. as web feeds (RSS, Atom, etc.). The siteinfo document acts a bit like a Feed Annotation Stream [are these still in play - or has the idea stalled?]. That is, they place an (optional) navigational context around a single post feed, or palce a blog feed in the wider context of its parent site, for example.

  • <link rel="siteinfo" href="navigation/mynav.xml" type="text/siteinfo" /> Pull the navigational structures into a page, just like you pull in stylesheet info. (I wonder why A9 didn't try for this approach as an 'alternate' rel link?) It's not hard to imagine how this could be used either in a toolbar (as A9 have done) or a browser sidebar, for example).

  • XMLHttpRequest Pull the navigational structures into a page using XMLHttpRequest, and then render it within the page. 'Localising' the user's current location within the site will require a little workaround, admittedly.

  • Frames In frame based sites, its easy to open a 'content' page in its own window and lose all the navigational structures. Not any more...

  • Widgets, Gadgets, etc. How about having a navigation console to your website in nicely branded Konfabulator widget? Please take it as read that this is an implied lazyweb request for an navigation widget!

  • Site Views Just as it's possible to have different visual styles associated with a site, how about different navigational views (each with their own sitemap.xml definition)? For pages that do separate navigation from content, a navigation equivalent to URIid (which allows users to skin arbitrary websites with their own css styling) would allow users to add their own navigations structure to a site.

To sum up then, as it stands I thinkthat if siteinfo is usedjust by the A9 toolbar, it won't go anywhere. But if the playful community get hold of it, generalise it a bit (away from a fixed name siteinfo.xml in a fixed location (i.e. at the root of a website: and start playing with mash-ups, it could be quite revolutionary... (or not!)

Monday, September 19, 2005

Gadgets Hot On the Heels of Widgets

It wasn't so long after I got my first Konfabulator widget that Google Desktop extensions came in to play. Now it seems like the third of the tussling three (Google, Yahoo and Microsoft) want to play too: Microsoft Gadgets for the Windows Sidebar...

OpenSearch Description Format Embraces HTML Search Engines

Chasing down the announcement of the IE7 mash-up with OpenSearch (IEBlog : IE7 and OpenSearch: Behind the scenes) using the OpenSearch 1.1 format, it seems that the new spec allows developers to define an interface to an (X)HTML web search, as well as RSS search services.

For example, here's a snippet from the draft OpenSearch 1.1 spec:
<Url type="text/html"
<Param name="s" value="{searchTerms}"/>
<Param name="p" value="{startPage?}"/>
<Param name="c" value="{count?}"/>
<Param name="l" value="{language?}"/>

It'll be interesting to see whether the Firefox community embraces OpenSearch now. The last time I checked (a couple of months ago now, admittedly), I didn't come across any extensions that exploited the OpenSearch format (this was one I could see it being used), but things may have changed now... Time for a visit to the Moxilla ectensions site, methinks...

And how about Konfabulator? Are they gonna get in there too, I wonder?

Friday, September 16, 2005

Microsoft Embraces OpenSearch?

I've blogged a few times about A9's Open Search, so when this post - IEBlog : IE7 and OpenSearch: Behind the scenes - appeared on the IE blog, it caught my eye.

Here's a snippet:

It’s been a lot of fun working with A9. But, this is just the beginning. The OpenSearch spec occupies a special place because it intersects 3 important communities – browser, search and RSS. We want to hear from all of you. My dream is that OpenSearch is adopted by all the browsers and millions of sites (Internet and Intranets) and that a significant percentage of these sites also expose RSS.

I'm still waiting for the OpenSearch Konfabulator widget, or a corresponding Google Desktop extension...

A Three Way Battle? Fighting to Dominate Search, the Desktop, and the Data Web

The pace of the web at the moment is frightening - every time I take a couple of days away, there's another announcement.

I've just come back to the blogosphere after a few days away, and it seems that Google have just got in to blog searching (Google Blog Search, announcement) though many of the reports I've read on it so far have been unimpressed (I'd list them - but it's late... bad form I know but I didn't bookmark them anywhere...)

It strikes me that there are battles raging on several fronts at the moment -here's just a few things that have caught my eye:

On the Desktop

Microsoft think they own it, of course, but Yahoo!'s investment Konfabulator, and Google's new improved Google Desktop are both plays of a sort for that valuable screen real estate.

In the Persoanlised Search Arena

There's also the battle for the personalised search portal too, and again Google got in there outside the Beta this week (Personalised Google Homepage, announcement), though surely Yahoo My Web can't be far behind. Cpome to think of it, I haven't heard much about the MSN Start personal portal for a bit. In any case, none of them have got anything (IMHO) on the almost ultimately customisable Netvibes (I wonder when they'll support drag and drop placement of panels!)

For the Data Web
RSS is coming on strong, and is likely to take of further as it gets rebranded as web feeds. IE7 will make a strong play for RSS, as will Longhorn, and we all know that Firefox can make more than a little use of the format.

PS Along with Google's BlogSearch, there's also the quietly released MSN feed search and what I think was the the first feed search from the big guns, Yahoo's feed search.

PS Now why couldn't I have put it like this..? tags:

Personal Book Catalogues - Time for a 'Course Books' Service?

Sometime ago, before I got into blogging, I tussled with the idea of what Amazon Academic might look like. The leverage would come from students and academics listing courses required on a particular course in a 'course wishlist', or some such (now I realise you'd just tag the books with a course code or course ID...).

There were several possible benefits - for potential students wanting to see what sorts of stuff they'd be letting themsleves in for if they took the course (they would of course be able to browse inside a book on Amazon, or via Google print); for academics, looking to see what the rest of the academic community was recommending their students to read; for the students - possibly - having a comprehensive booklist with single click purchase or library reservation etc etc.

At the time, I thought you could get quite a long way into this by repurposing Amazon wishlists, but now there are a few 'social book cataloguing' sites that would suit the bill admirably.

Reader2 has very pretty AJAX interface that pulls in book covers from your tagged and catalogued books, and by now standard tag-based URL addressing scheme (e.g. guess what sort of books you'd find addressed like this:

Library Thing is another social book cataloguing site that again offers users the option of tagging their books, and sharing those tags with others.

When I get a chance, I'll try and pop up a screencast comparing these two sites. In the meantime, if you know of a link to a good review, or even a screencast of either of these sites in action, please post a comment, or perhaps tag it for me (psychemedia) via