Have been playing around with Wordpress.com - very nice.
For years now people have obsessed with separating style from content and have thought that 'style' is the same as 'layout'. CSS has been used for layout, which it is a very bad language for (unlike many templating languages glancing at a CSS file does not tell you what a page design will look like which ruins the whole 'view source' model that made html so successful).
XML is a much better language for layout, but we are stuck with CSS, and so will have to split CSS into separate layout and style documents.
Andy Skelton & Scott Allan Wallick's Sandbox theme is the first time I've seen something that moves towards separation of style from layout, not just style + layout from content. There is a lot that could be done with that - particularly if the semantic placeholders that have no real 'layout' component are separated out.
If that were done, and there was a convention for class names for specific types of element (e.g. a class for text can have font, size, color properties etc.) a generic css styling wizard could be built against it.
This mod re-write strategy seems to fall into the 'duh why didn't I think of that' category of really simple but perfect solutions.
People use mod rewrites to trick google into indexing stuff create pretty urls.
Rails and a bunch of other frameworks force a grammar for links based upon actions and data when really most of the time there is no distinction made, there are just name/value pairs passed to the url. e.g. color=red is no different from colorit=red.
The strategy below loops through name/value pairs as url query string parameters and actually spits out the name and the value in the url:
site.com/name1/value1/name2/value2.
There are some potential problems with null values and ordering, but I like the idea.
Create Dynamic URLs With Mod_Rewrite and PHP Functions by www.Shadow-Fox.net
Next/previous, back/forward buttons - the single most important bit of web navigation are being used to mean the opposite to their original use in browsers, because of blogs.
'Next' on Technorati means back in time and has an arrow which points to the right. 'Previous' on Techcrunch means back in time and has an arrow which points to the left.
Web browsers are in many ways as simple as the universal music playing interface that has existed since the cassette player.
Music interfaces consist of rewind, fast-forward, stop and play.
A browser UI is almost the same and consists of back (rewind), forward (fast-forward), stop (largely redundant in the browser), refresh, home/url-entry (play).
Within a web page the back and forward buttons, 'next and previous' are ubiquitous for search results and the, increasingly archetypal, blog style UI.
Because blogs are reverse chronological lists and search engines equate 'next' with less relevant or less timely they do exactly the opposite of what browsers do.
The problem is that websites are increasingly following the browser model rather than the search engine one.
For Google or most blogs 'next' means back in relevance or time and 'previous' means back in browsing history.
Because blogs happen to have reverse chronological postings the 'next' 'previous' model seems like it is compatible with a browser's history. It is, but only for this specific case.
It means that 'previous' can mean back in browsing history but more recent in time - which is confusing, to say the least.
The solution, I think, should be to assume that the browsers got their first, and that navigation history should be left/right arrows, where left is back in time.
For blog style navigation, where the very simple vertical list was developed, I would suggest that the arrows should be up or down, where back in time is a 'down' arrow and more recent is an up arrow.
If this convention were followed, there would be no ambiguity, and a single navigation device with left right and up down arrows, could be used.
In the UK in the late 80's British Telecom carried out one of the single biggest acts of design vandalism when they systematically removed the famous red telephone boxes designed by Gilbert Scott et al.
The justification for this was that they were not accessible to people in wheelchairs. This argument was impossible for people to counter and yet hid the truth - there were other ways of making phone boxes accessible that would not have required a complete change.
People argue, quite rightly, for web accessibility, but what are the results?
If you pass some of the top web sites' front pages to the W3C validator:
Yahoo - does not validate
Ebay - does not validate
Amazon - does not validate
Google - does not validate.
These have all been around for a while, however. What about the newer breed of online services?
Flickr - does not validate
Digg - does not validate
Del.icio.us - does not validate
Are all these companies wrong, or is there something wrong with current accessibility standards?
In the next part I'll look at the current state of HTML and argue for a different approach.
Get a move on Apple - please don't let a crappy half-baked service like BuyMusic.com steal your thunder and get any gullible customers before you launch your Windows music service.
Everything about Buymusic.com looks second rate; its like Tony Soprano hadn't heard of the dotcom crash and thought he could make a few bucks.
And its not just the service that sucks - the marketing manages to rip off Apple's TV ads so badly that you think you're watching a skit on Saturday Night Live, but most of all it's the product that stinks - music you can't listen to on an iPod or burn onto a CD.
I feel better now that I got that off my chest.
One of the treats of the last O'Reilly Emerging Technology conference was to see Matt Jones' excellent presentation design (along with some on stage air-guitar antics).
Within software, design is often treated either as a superficial veneer or as a reductionist process where aesthetics seem to disappear altogether. Unlike Jakob Nielsen, people like Jason Kottke and Matt Jones are true information architects.
ideal_process2.gif 773x544 pixels
Saw a demo of Chandler 0.1 today.
The most impressive aspect so far was that everything in Chandler has a URI and that any URI can be opened up and shared in real-time with other users. This seems to be extremely powerful and elegant
In '94 we did a 3d interface to Lycos where the search results were returned as a 3d model spinning around (groan) a globe. Results were shown as Cubes, Cones, Spheres and Cylinders, indicating whether the sites linked to were commercial, educational, service providers or others respectively. The size of the object represented the relevance and the color represented location, green for sites registered within the US and red for outside. The objects were slowly spinning and the speed of spin of the object represented size of document. Large documents spun slower.
The problem was that this was a toy, no matter how seductive the idea of 3 dimensional or graph based representations of search results, a list of text results is more useful for all but a handfull of specialist applications.
That is the problem I have with this and other attempts to create visual maps of search results.
The web produced some very sophisticated interfaces early on (remember Onlive Traveller) but simple interfaces are the winning formula, just ask Google.
Oh happy day!
Jason has found decent CSS styles for rounded corners. Now I can finally get around to doing a proper XHTML version of this site and ditch tables.
Albin.Net CSS: Bullet-Proof Rounded Corners
More details of Namex which allows high precision, fuzzy search for people by name.
I know that Dave probably won't agree with me but now that the big guys are interested in weblogs, weblog software need patents. Search software has long been a patent rich environment, Google has them a plenty and one of the reasons why Overture bought AltaVista was because of their patents.
Weblog software is elegantly simple and depends on innovation that is easily copyable by the bigger companies. All the more need to protect and continue to foster this innovation.
Although patents are often seen to be an anathema to the collaborative world of the developer community, they do offer protection for smaller companies. Perhaps a compromise would be a patent system analogous to the Creative Commons approach to copyright which protects against exploitation whilst preserving the developer community ethos. In other words patents which are not infringed if they are used within open source or non-profit development.
Kevin Lynch has started a weblog.
Kevin is one of the few people who can lay claim to be a true software designer or architect. A Brunel of software design.
Dave Winer boils down the argument for switching to CSS driven layout, its all about getting rid of the damn awful HTML 'table' element which has been slowing down page loads for the last few years.
"People responding to the query about a CSS version of Weblogs.Com are asking why I want to nuke the table. Performance."
Apple gear up further to removing their reliance on Microsoft products with a rumored word processor, back to the old days of Claris then.
Why follow the same product breakdown as Microsoft at all? Do you really need fully fledged applications for 'word processing', spreadsheets, presentation?
I rarely use a fully fledged 'word processor' and would rather see a completely different view of software, built around a modular framework with a plugin architecture for developers. These plugins could be local or distributed, like locally cacheable web services. I.E. you could remotely load a French spellchecker on demand whilst editing text.
Why not take the principal activities of desktop computing: editing (text, bitmaps, grids, video etc.); publishing (to HTML, XML, PDF); retrieving (a full text database filesystem) and make them modular components of a universal framework application, an extension of the operating system. Microsoft tried this a way back with OLE, but design innovation aside, Microsoft are constrained by their own packaging and revenue streams. I would trust a company like Apple to pull off an architectural rethink such as this with elegance and flair, it seems a shame that they are just cloning the Microsoft product line with added bells and whistles.
MacWhispers.com - Whispers From Around The Mac Community
Ben Hammersley.com: Safari Review
Ben Hammersley posts a to-the-point review of Apple's new browser:
"Sadly, at first glance it's shit - No tabbed browsing. Which is now an *ESSENTIAL* part of the UI, and without it a modern browser is just pants. The CSS support seems shoddy too. It's pretty though."
Please, please may this browser be shite - near total browser monopolies are great for lazy web design.
via Dave Winer. This year's Wired awards nominees are up.
What makes the Rave awards special is the focus they put on design in an industry that often treats design as a superficial afterthought.
It is especially gratifying to see that true software architectural design, as opposed to graphic or interface design, is the criterion for the software designer of the year.
Jim McClellan sees blogs as a newsgroup replacement:Survival guide 2003: Blogs as newsgroups - exactly.
Usenet had two flaws - 1. it was unintuitive for non-techies to setup a new group. 2. People are individualistic and would rather carry on a discussion by posting to their own site and linking to others. EGroups, now Yahoo Groups, solved 1., weblogs solve both 1 and 2. A distributed discussion with individuality intact.
Technical reasons aside, the web took off largely because web pages can now be designed with pretty pictures, otherwise we would all be using something that looked like Minitel.
The year is 2002, I am sitting in front of a computer that 10 years ago would have filled a large bunker and could model the aftershocks of a thermonuclear explosion and I want to find something on this computer.
I have two choices: I can use 'Start:search:for files or folders' or I can type something into a browser window. The first will clunk and whir and a couple of seconds later will search the few thousand file names on the computer but will not search the contents of those files. The second will fully search over 3 billion files from other computers and return the answer within 0.2 seconds.
Microsoft Windows is badly designed.