New puppy video

Testing if this will work.. using the "video" tag in HTML5.

Alternatively, download the video directly: kaiya_garden.ogv. Oddly it only works if I name .ogg, the .ogv version does not work in video tag!

We need Wikitranslate engine

There is Google translate, which solicits user feedback, but Google do not apply changes. The wisdom of the masses is better than a single centralised system. I would like this to expand into something which is a systran beater. Where we can click on the translation output, and make corrections, which are fed back into future translations. Like translating from Japanese and getting "Righter" instead of "Lighter"

Why web forms are bad (compared to email)

Don't you just hate it when you click on the "email" link on someone's website and it takes you to a tiny web-form which only fits about 6 words per line, cramped into a small box? After struggling to write your message you invariably click "submit" and it says Error.. you click the "back" button optimistically and of cause the whole message is gone. Webmasters got overexcited and tried to run internet messaging through a web page, when really it should be left as standard email or IM.

Short run down of why web-forms are bad for messages:

• Tiny form boxes, pledgebank is a great example, limited to 7 words per line.
• Often arbitrary restrictions on the text that can be written. Lincolnshire NHS trust have a secret blocked character list including ",', and ; then they limit to 255 characters.
• Often fails to "Submit" and when you click back of cause it has lost all your text.
• Leaves me with no record in my Sent folder of what I've written.
• No way to forward the message I have sent to a friend, or another department at the organisation.
  

All these problems make it very difficult to send a website a message telling them one of their pages has broken links or typos etc. Web-forms can always be a secondary contact option, but email should always be the first option as it is the best. See Email 101 for tips on writing to the point.

Firefox3 Intelligent Search Location Box

The smart web location box in Firefox3 will be much more useful when it takes advantage of the stats it has to hand. This is essentially a development of the idea I posted back in 2004 Technique to Facilitate Intelligent Functionality Tailored to Each User. I see the Search box as dupicating space, the browser should really have a single "Location search box". When the addresses in the bar start with http:// then they are URLs, when they are keywords separated by spaces they should be treated as a search query.

One other thing I would love to see is a "Save page" (snapshot) button which just grabbed all the content and stored it in an HTML WebDoc archive, these could be recalled later from a history. A prime use of this would be when I have booked my flights, but my email confirmation has not arrived yet.. I need to either print the web-page.. or hang-on and keep the page open until an email confirmation is processed so I don't use track of the confirmation reference.

These archived pages are static (the generated HTML layout) in that they can't be modified, they can be printed or forwarded as emails in their format as they appeared when the snapshot was taken with the click of the "Save page" button.

In the future, when the off-line file-store, email inbox/sent/drafts and online storage of documents is eventually unified... these snapshots of pages would show up as "Easyjet.com Flight booking -- saved 15 May 2008" in the "Saved" folder. So we can return to the archived pages for future reference should the airline loose track of our tickets! It will even be possible to forward them as the archive Easyjet_Flights.webdoc file I am sure ;)

Mobile Web Problem

Back in the 90s when we were all still creating our websites in HTML 3.2 we optimised pages for dial-up and compressed images down to the smallest gif we could. One unfortunate side effect of the Broadband boom in the UK worldwide is that websites don't cater for the optimal page sizes which mobile devices necessitate.

Even with the bandwidth problems of popular sites (BBC News front page is 278KiB) until last week the BBC News site did fit on my 800 pixel wide Nokia 770. However they have changed it now, so a minimum display width of 1000 pixels is required. We can of cause switch to the low graphics version.. but when the version before fitted, it is a shame we have to go back to a nearly text-only web page if I choose to browse on a mobile device.

A lot of web designers (including the BBC?) make the mistake of looking at the screen resolution of their visitors and assuming that people browse full-screen, when many people do not maximise their browser windows.. so that 1280x1024 display window is actually only about 800x600.

If you look at the resolutions of mobile devices you will see all the current Sony Ericsson models run at 240x320 resolution, and Nokia models the same. Apple iPhone is slightly higher at 320x480. LG KU990 Viewty comes in at 240x400

The other thing for website designers is to remember is that a 240pixel wide display which measures 2 inches across is 120dpi (compared to a normal desktop 72dpi), so if you display your text at 10pt, that will look 40% smaller (why aren't font sizes specified in cm on screen ?)

So web designers, remember there mobile market for browsing is growing all the time, optimise for small page bandwidth, and page width/height no more than 800px (my site comes in at 768px ;)

The other problem is sites with broken HTML, like the BBC News site, 375 errors. That is shoddy! (I should point out that blogger which generates my site has left 169 errors on the page, so I'm not in the clear either!).

Web-apps for Email GPL

I use GMail myself, but does anyone know of a decent Web-app using AJAX I could install on my server and use for email and IM (XMPP etc) while out of my office? Sometihng as good as the current GMail would be needed to get me to switch!

Flakey GNU Flash (Gnash) causes Firefox to hang

I'm running the latest Kubuntu 7.10 which comes with Firefox 2 and through I would see what all the fuss is about with Gnash (aka GNU Flash). Gnash is the GNU/FSF branded project to implement decoding support of Adobe's Proprietary Flash file format that is sadly so common on the net these days (seems contrary to the No MS-Word documents strategy you're probably thinking too?)

Anyway I followed the instructions:

apt-get install mozilla-plugin-gnash

Gnash installed ok, so I headed over to youTube to hear a new track by a band I heard on the radio earlier. I hadn't' thought to save my draft email in another tab.. pretty risky this software stuff.. My laptop started chugging and churning, I ran "top" and could see gtk-gnash was using 1.6% of RAM in its two process threads, and 20% CPU in one process and 79% in the other one! This went on for 5 mins before I managed to close the tab (after several warnings from Firefox about a script which had stopped responding). There were no errors reported, and the clip never worked. I did get a glimpse of the spinning loading icon youTube uses though, before it all went grey.

Bizarrely, there is a context menu item in Gnash (from within the browser) where you can "Quit".. so I did this and it went down to one process taking up 80% of CPU time!

So it looks like.. unfortunately for the Gnash developers.. that the implementation is presently as flaky ass the GNU+FSF strategy to support Adobe is.

3D on the web (VRML, X3D and COLLADA)

3D on the web is one area that is still absent, I'd like to see that change over the next couple of years, especially as the tech has been around since the 90s. We've got the open format VMRL and X3D, which via the script tag support Javascript code. There is the Sony Computer Entertainment COLLADA format which is now maintained by the Khronos Group.

So to pull it all together browsers really need to add native support for these formats. While their are "plugins" such as the excellent OpenVRML that is still just a plugin, and not an integrated component of the browser like SVG or HTML. Perhaps someone at an innovative net business would be what's needed now to get things rolling.

Rich Web Typography - Just around the corner?

It's 2006 and it still feels like we're in 1995 as far as fonts on the WWW go. Time has moved on, we've stopped calling it the WWW and just say the web now. It's integrated into the fabric of our lives, we don't think about what's happening when we access information, it's just something we rely on.

Fonts on the web haven't kept up with the pace of change, the fonts which can be used in web pages are still limited to what ships with peoples computers! When the tech is in place, I want to see web sites providing fonts just like they provide an image or a vector graphic at present ;)

There are ways we can get custom fonts into pages, but they are either tied to a specific browser or require the user to download and install the TrueType Font themselves as administrator of their PC beforehand. Other approaches use Adobe's Flash as the means of rendering Anti-Aliased fonts in a browser, sIFR 2.0 is one framework which takes this approach; it does gracefully display using standard XHTML if the Flash plug-in is not installed, so users should not see any Flash errors at least.

Another key point is that users and companies need fonts they are allowed to distribute, home users might not be able to afford to buy special fonts. This means home users will either have to make their own (a time consuming affair, even more so for Asian languages!), or just put up with the bog standard Helvetica and Times fonts which most Linux, Mac and Windows installs come with (you need to maintain the common denominator between systems right! ;) One popular community produced free font is Gentium, I've also been using DejaVu Sans, check them out!

There is already a way to pull in our fonts, take this CSS example I wrote from the W3C's CSS2 spec, (dropped this from the CSS2.1 spec though!):

/* Define Maxus font to be downloaded if needed */
@font-face
{
font-family: "Maxus"; src: url("maxus.ttf") format(TrueType)
}

The W3C gurus have been discussing fonts too. Here's what I can see needs to be prioritised to get this in place:

• Common browsers need to support a finalised spec, pure and simple.
• As fonts take a long time to make, there aren't as many great ones going for nothing as their could be, so that area needs some targeted work by a collection of vendors, Universities or designers etc.

As a side note, these developments could create a whole market for fonts to be licensed for use on a website, or given away free under a non-commercial licence to home users etc.

So to summarise, let's focus on the open CSS2 specification we have. Plan of action: Web authoring tools vendors like Adobe need to support it, the W3C to promote it and web browsers to implement the support! ... possible? Yes. Will it ever happen? If no better solution comes along first, I hope it will!