As I had a lot of spam problems and other attacks I upgraded the blog to a new wordpress. This is why this looks terrible and unfinished but I guess most of you haven’t realized it as you read the RSS feed anyway (which I consider the right way to read a blog).

I don’t have the time to make this pretty right now, but all the content is stil here. Please give me a shout when something stopped working.

Having just spent another hour deleting trackback spam advertising a non-existent medication I am thoroughly sick and tired of playing catch-up with the evildoers of the web and turned off trackbacks on this blog. I can get the same information from my logs, although it is a shame that I have to go there.

The initial idea of blogging was awesome: you publish information, get immediate feedback with comments and you’ll also get notified by trackbacks when other bloggers talk about your stuff or build upon it.

However, this seems to be over. If you have a blog that managed to get high up in Google’s pagerank (and you didn’t get demoted last month for daring to try to make some money with it) you will get the following:

• trackbacks from web sites that don’t exist (for whatever reason)
• trackbacks from web sites that were built for SEO (like http://learnitself.info/ – check the footer – SEO WordPress theme Ad Flex Blog ( Using v0.8.9.6a – v0.8.9.8h available ) by VK Solutions sponsorized by Digital Proof – sponsorized???)
• trackbacks from blogs that scrape other blogs (both Ajaxian and Smashing magazines are cloned throughout the web when they have new posts)
• comments from clumsy spammers
• comments that were written on order ( as Marco ranted on about lately – online and IRL :) ) to advertise a certain site

And that’s the evil spammers of this world. There is however a lot more noise on the line:

• trackbacks from people who consider a list of their latest del.icio.us bookmarks a blog post
• comments from people who don’t really want to comment but advertise their own blogs by hinting they have a solution
• comments from people who do not read up on the context but only on what they get by scanning over the post (this could be because of them reading the RSS and not getting that a post is just pointing to a larger article or product – you get a lot of those on sites like Ajaxian)
• comments from people who are fanboys and -girls of certain technologies and by default consider everything else wrong and people who talk about it morons
• comments from people who try to annoy you or take you on as you are “the expert”

This is nothing new either (I’ve written about this three years ago) but it has become so much you have to spend more time swallowing very short and rude remarks than writing your blog.

As said I am turning off trackbacks, as I am not showing URLs they were pretty pointless here anyway. I will instead publish URLs that I consider good enough to link back to. For the moment I am still allowing comments as I do get the occasional good one (in a vast sea of rubbish) but I am toying with the idea of getting rid of those, too. This would be really sad as I did start blogging because of the option to get immediate feedback from people anyone can read.

I am not at all subscribing to the school of people who claim comments are dead and if you wanted to comment blog about it yourself or twitter about it and we use technorati as the feedback consolidator. I don’t expect everybody to blog or use twitter and consider it very dangerous indeed to go down that route as it does smell of inbreeding to me. People find our blog posts via search engines and should not have to be part of the “blogging scene” to be able to comment.

That said, we need a remedy to fix the social aspect of blogging, one that is easy to access and secure enough to stop the spammers. That is very tough indeed, cause anything that is findable is also spammable.

[tags]blogging,spamming,spam,trackback,bastards,socialblogging,social web,social[/tags]

I like Autocomplete as a design pattern. It allows you to find things a lot faster than using a normal search box and saves you having to go through a search result page. Autocomplete means progressively enhancing a normal search box to provide a faster and more channeled way to find information. Good autocomplete controls do that in a totally unobtrusive manner and load the extra information only on demand and when the user requires it.

However, there is another use case I haven’t seen covered properly yet: what if I want to offer a defined, small amount of options and a full search but not rely on JavaScript? The solution is to use a list of links followed by a search box:

Apples
Bananas
Carrots
Dates
And so on



Search:

Wouldn’t it be cool to use this as a fallback when JavaScript is not available and use this information to seed an autocomplete control when it is? Well, you can.

Naturally at work I keep running into the YUI Autocomplete control and this one has the option to use a JavaScript array or even a JavaScript function as the data source for the options. We can easily use this to get the information from the list of options:

function doAutoComplete(){
var optionsLinks = opts.getElementsByTagName('a');
var ds = [];
for(var i=0;optionsLinks[i];i++){
ds.push(optionsLinks[i].firstChild.nodeValue);
};
var acDataset = new YAHOO.widget.DS_JSArray(ds);
acDataset.maxCacheEntries = 0;
var ac = new YAHOO.widget.AutoComplete('s','autocompcontainer', acDataset);
};

The thing about YUI autocomplete that always annoyed me is the amount of dependencies you have to include to get the whole thing working. This is of course by design as it builds on other parts of the YUI, but it was a bit of an annoyance to me. There is however a really cool way to make YUI controls load the YUI components necessary: you can use an object called YAHOO_config and define a listener method. This listener method gets called every time you include a YUI component and gets its name as a parameter. Couple this with dynamically created script nodes and you can call the dependencies one after the other:

function YAHOO_config_ac_run(o){
if(o===undefined){
addJS('yahoo-dom-event/yahoo-dom-event.js');
} else {
switch(o.name){
case 'yahoo-dom-event':
addJS('animation/animation-min.js');
break;
case 'animation':
addJS('autocomplete/autocomplete-min.js');
break;
case 'autocomplete':
var l = document.createElement('link');
l.type = 'text/css';
l.rel = 'stylesheet';
l.href = 'http://yui.yahooapis.com/2.3.0/build/autocomplete/assets/skins/sam/autocomplete.css';
document.getElementsByTagName('head')[0].appendChild(l);
YAHOO.util.Dom.addClass(document.body,'yui-skin-sam');
doAutoComplete();
break;
};
};
function addJS(url){
var head = document.getElementsByTagName('head')[0];
var s=document.createElement('script');
s.src = 'http://yui.yahooapis.com/2.3.0/build/' + url;
s.type='text/javascript';
head.appendChild(s);
};
};
YAHOO_config = {
listener:YAHOO_config_ac_run
};
YAHOO_config_ac_run();

Putting it all together you get an autocomplete control with a list as a fallback. Try turning JavaScript on and off to see the difference.

[tags]autocomplete,yui,progressiveenhancement,unobtrusive,webdevtrick[/tags]

This was part of my presentation at the Open Hack Day in India and I just got the time to write it up.

Here you’ll learn how to get Flickr photos into your JavaScript solutions without having to resort to using the full API. As this is a hack you will only get the latest 20 photos, if you need more detailed data like restricted to sets or more at once you’ll need to resort to the flickr API.

I mentioned it earlier for some of the Flickr badges shown here and my 24 ways article on how to use Flickr in JavaScript but it is good to have a quick step-by-step to learn how to use Flickr with JavaScript without needing to resort to a full-fledged API call.

Step 1: Locate the photos you want to use on flickr.com

Just surf around the site and find a site that does have an RSS feed. For example, my Flickr stream at http://flickr.com/photos/codepo8 or all the photos of Open Hack Day India at http://flickr.com/photos/tags/hackdayindia/.

Notice that only tags work, not search!

Step2: Click the RSS link or RSS icon (in Firefox) and change the ending of the URL call

This will get you the URL to the RSS feed, for example

http://api.flickr.com/services/feeds/photos_public.gne?tags=hackdayindia&lang=en-us&format=atom

and all you need to do to use this information directly in JavaScript is rename the “atom” at the end to “json”.

Do this and you’ll get a bunch of JavaScript instead of the atom feed:

http://api.flickr.com/services/feeds/photos_public.gne?tags=hackdayindia&lang=en-us&format=json

jsonFlickrFeed({
"title": "Photos from everyone tagged hackdayindia",
"link": "http://www.flickr.com/photos/tags/hackdayindia/",
"description": "",
"modified": "2007-10-12T14:08:47Z",
"generator": "http://www.flickr.com/",
"items": [
{
"title": "Open Hack Day coverage in Indian newspaper",
"link": "http://www.flickr.com/photos/codepo8/1552753330/",
"media": {m"},
"date_taken": "2007-10-12T15:08:47-08:00",
"description": "
codepo8 posted a photo:
 
 
The Bangalore newspaper Midday put up this article about the Open Hack Day and apparently I am an ethical Hacker surrounded by Brad and David :)
",
"published": "2007-10-12T14:08:47Z",
"author": "nobody@flickr.com (codepo8)",
"author_id": "11414938@N00",
"tags": "india newspaper bangalore fame midday bradleyhorowitz davidfilo hackday christianheilmann joearnold hackdayindia"
},
{
"title": "Break It",
"link": "http://www.flickr.com/photos/code_martial/1536271756/",
"media": {m"},
"date_taken": "2007-10-06T23:18:58-08:00",
"description": "
code_martial posted a photo:
 
 
Joe Arnold does the Break Dance.
",
"published": "2007-10-10T20:43:01Z",
"author": "nobody@flickr.com (code_martial)",
"author_id": "61697474@N00",
"tags": "yahoo dance bangalore taj hackers hacking 50mmf18af hax0rz tajresidency joearnold hackdayindia"
},

// ... and so on ....
]
})

Step 3: Write your display function

As you can see, the Flickr API encloses the dataset inside a function call that executes jsonFlickrFeed, which allows you to define a function with that name:

The data of the JSON object is a tad askew, especially the encoded output inside the description is pretty much useless in JavaScript. However it is really easy to get the image data and display all the images using the media.m property of each item:

I trust you know however that using document.write is the main reason for hair loss and other problems in this world, so use some nice DOM scripting instead. This was just for the sake of this example.

Step4: Choose the size of the photos with different file endings.

The only remaining annoyance is the size of the photos. Once again, hacking the URLs of flickr helps you there. By default, flickr will choose m which is the medium size of the photos. Simply rename the something to the size you need:

• _b is the big size (1024)
• _t is the thumbnail
• _s is the 75×75 pixels square
• _m is the medium size
• removing the _? extenstion means you choose the small size.

You can also get the sizes by clicking around the image detail pages, for example http://www.flickr.com/photo_zoom.gne?id=1536271606&size=s and checking the image properties.

Happy Hacking!

[tags]flickr,hacks,openhackdayindia,JSON,howto,trick,webdevtrick[/tags]

After writing my first blog post for the Yahoo Developer Network blog it’s time to also upgrade this one here. There are 14 hours left at the Open Hack Day in Bangalore for the hackers to come up with something to woo the others and the Jury. So far I’ve been asked a lot about mail authentication APIs and if you can have a transparent XUL Canvas overlay an iframe to be able to paint on a third party web site. I am a bit scared about what people will hack here.

My presentation seems to have been well received and as always, here are the slides called Yahoo for You on Slideshare

I’ve also taken and uploaded a lot of photos and will go on doing so until I go back to the UK next Thursday.

[tags]hackdayindia,hackday,openhackday[/tags]