lunatechian (lunatech-ian)

Some weeks back, I was talking with my manager about AI and how it is such a bogus field. My manager replied that in a few years we will see applications that use AI in our daily life. However, I was quite skeptical - and I refused to agree to this. He then gave an overview of neural net and how they can learn to solve the problems. Here I pointed out that Bayesian filters can also be considered a form of AI, as they can learn from their previous data and they can make decisions, but Bayesian filtering is mathematics and not AI. At this he replied that most of AI is mathematics and only some part of it is hocus-pocus and hand waving.

This brings me what I have been thinking for a long time.Joel write

A very senior Microsoft developer who moved to Google told me that Google works and thinks at a higher level of abstraction than Microsoft. "Google uses Bayesian filtering the way Microsoft uses the if statement," he said.

. I had always suspected this and had also felt that this was the way to go. A few months back, we had a presentation by a researcher (not a Yahoo! employee), who was working on extraction and summarization of documents. He had a formula that he was applying on the sentences of the documents to find the weight of the whole sentence and then finally if the weight of the sentence was above some limit, it showed up in the summary. I was skeptical about this approach - my belief is that the Bayesian approach can be used to classify documents. Luckily, there is a project that seems to provide a framework on which things can be built further.

I was reading through the mails in the php-general mailing list and came across this mail by Rasmus about AJAX

I find a lot of this AJAX stuff a bit of a hype.  Lots of people have
been using similar things long before it became "AJAX".  And it really
isn't as complicated as a lot of people make it out to be.  Here is a
simple example from one of my apps.  First the Javascript:

function createRequestObject() {
    var ro;
    var browser = navigator.appName;
    if(browser == "Microsoft Internet Explorer"){
        ro = new ActiveXObject("Microsoft.XMLHTTP");
    }else{
        ro = new XMLHttpRequest();
    }
    return ro;
}

var http = createRequestObject();

function sndReq(action) {
    http.open('get', 'rpc.php?action='+action);
    http.onreadystatechange = handleResponse;
    http.send(null);
}

function handleResponse() {
    if(http.readyState == 4){
        var response = http.responseText;
        var update = new Array();

        if(response.indexOf('|' != -1)) {
            update = response.split('|');
            document.getElementById(update[0]).innerHTML = update[1];
        }
    }
}

This creates a request object along with a send request and handle
response function.  So to actually use it, you could include this js in
your page.  Then to make one of these backend requests you would tie it
to something.  Like an onclick event or a straight href like this:

  <a href="javascript:sndReq('foo')">[foo]</a>

That means that when someone clicks on that link what actually happens
is that a backend request to rpc.php?action=foo will be sent.

In rpc.php you might have something like this:

  switch($_REQUEST['action']) {
    case 'foo':
      / do something /
      echo "foo|foo done";
      break;
    ...
  }

Now, look at handleResponse.  It parses the "foo|foo done" string and
splits it on the '|' and uses whatever is before the '|' as the dom
element id in your page and the part after as the new innerHTML of that
element.  That means if you have a div tag like this in your page:

  <div id="foo">
  </div>

Once you click on that link, that will dynamically be changed to:

  <div id="foo">
  foo done
  </div>

That's all there is to it.  Everything else is just building on top of
this.  Replacing my simple response "id|text" syntax with a richer XML
format and makine the request much more complicated as well.  Before you
blindly install large "AJAX" libraries, have a go at rolling your own
functionality so you know exactly how it works and you only make it as
complicated as you need.  Often you don't need much more than what I
have shown here.

Expanding this approach a bit to send multiple parameters in the
request, for example, would be really simple.  Something like:

  function sndReqArg(action,arg) {
    http.open('get', 'rpc.php?action='+action+'&arg='+arg);
    http.onreadystatechange = handleResponse;
    http.send(null);
  }

And your handleResponse can easily be expanded to do much more
interesting things than just replacing the contents of a div.

-Rasmus

Update: Disallowed comments on the entry

I have added the "internet keywords" feature to the "2-in-1" search. If you now type "news India", it will do a search on both Google and Yahoo news. There are keywords for image and mysql too. Thinking of adding a "meme" keyword, which will search the Technorati and feedster search engines.

The code for this lies here and is in public domain.

If you have been using Linux, I think you must have been frusturated with the lack of good fonts when using Emacs. It gets very hard to distinguish between "l" (the letter l) and 1 (number 1) and between comma and semicolon when programming or examining large amount of code.

The Bitstream Inc have done an excellent work in producing and relesaing under a very liberal license, the bitstream-vera fonts. These fonts are really beautiful and easy to read. To use these fonts with your emacs, just add this line to your ~/.Xdefaults file

 Emacs*font: -bitstream-bitstream vera sans mono-medium-r-*--*-100-*--*--*- 

Of course, you need to have the bitstream-vera fonts installed on your Linux box

Tinymce is a nice WYSIWYG browser based HTML editor. It transforms the way the <textarea > element appears. Handy for providing users a way to edit their content. Does not work on safari

One of the strong points of PHP is its well annotated online manual, which is peppered with useful code snippets and hints. How many times have you pointed your browser to php.net/manual/ to look up that at-the-tip-of-your-mouth function's argument list ? If you or your company use the manual frequently, it might be a good idea to mirror the manual in your internal network. I have mirrored the manual in my office, and the two benefits that I see are

• Faster manual lookup - Living in a bandwidth starved country(India), this is a major win for us.
• A peek at well written code, again a major win if you are always looking for new and nifty way for doing things

The process of how to go about mirroring the manual is surprisingly well documented at Mirroring The PHP Website. As suggested by the documentation, You may want to exclude out certain non essential things ,like the manual in different languages. I also excluded the tar.gz and the zip files in the extra directory.

After downloading the manual pages, I created the virtual host (a simple copy-paste with some minor modifications) for the manual and that was it. To keep the manual uptodate I run the rsync command once a week.

For the record, my rsync command is rsync -avzC --timeout=600 --delete --delete-after --include='manual/en/' --include='manual/en/**' --exclude='manual/**' --exclude='distributions/manual/**' --exclude='distributions/*.exe' --exclude='extra/*.zip' --exclude='extra/*.tar*' --exclude='distributions/*' rsync.php.net::phpweb ./ Of course, you will need to change this to suit your needs.

Now available for your downloading pleasure, is my GoogleSoap class. have you eveer felt the sudden need to use the Google's web APIs through PHP ? Worry not as help is at hand. You can use the GoogleSoap class to make your job easier. I have put the code and related documentation/examples in the public domain, so you can hack it up further. Some rudimentary instructions are included in the README file, I will add more to it as and when time permits me. Any additions/suggestions are always welcome.

Get the code here