So Google made a plugin for Internet Explorer that forces any IE to behave like Chrome, and by doing that rids the world of the evil we call Internet Explorer 6. Just to let you know, saying that name aloud here at the Nivas Office causes spikes to protrude from the walls and impale you in the eyes. No, no, really.

But there is legacy problem with this Chrome Frame that will just prevent it from saving the developer's ass. And the problem is not Internet Explorer 6 engine or whatever it is that makes it so outdated. It's what is on the EXTERNAL of that IE6 that counts.

You see, the main reason for people staying on IE6 is that they work in a corporation where IT administrators locked the possibility of installing new software, or there is a global corporate policy that specifically says: "Browsers on computers will be IE6". Second reason is that people are just not tech-aware and are just afraid to install new software or just do not know how (we, the power users, really take some knowledge for granted - YES it is really hard to install new software if you are someone who uses computer 30 minutes per month). Third reason people do not switch form IE6/7 is that the developers keep supporting those platforms. I talked about that in previous post.

Basically meaning that Chrome Frame can and will be installed only by those people who can, and know how to, install Chrome or Firefox in the first place!

What Google did here is approach the problem from the 100% Engineering perspective. They know that IE6 sucks, so they fixed that by making software. In an older, but still golden, supreme blog post by respected designer Douglas Bowman who worked at Google it is clearly written: "When a company is filled with engineers, it turns to engineering to solve problems. Reduce each decision to a simple logic problem. Remove all subjectivity and just look at the data." And the world does not function like that.

Easy! :) A while ago, I ranted about fulltext stopwords in mysql.

To repeat - the stopword list is a list of most used words in english language. Common words such as “some”, "little", "let" or “then” are stopwords and do not match if present in the search string. Basically, fulltext searching for any of the stopwords would return (almost) all the entries, so MySQL ignores those words to reduce result pollution and for efficiency. If the word starts with stopword, that word is returned, however if your word IS stopword, you are out of luck. Here is a list of build in stopwords in MySQL 5.0/5.1. Yeah, well this is so very exciting, but whole universe DOES NOT speak and write just english.

Last time we had problems with stopwords, there was no way to disable them without recompiling the mysql engine from the source. On production servers where you need to stay away from recompiling custom builds for maximum compatibility that was no go.

Today we again encountered stopword problem. Client complained that he cannot find video from artist "Let 3" or artist "Little Boots" (Victoria Hesket). To my surprise, in no time - I've found out a prefectly legal and easy to apply solution. Just add this to your .cnf file, restart mysql engine and rebuild indexes:

ft_stopword_file = ""

I addition to that, if you haven't allready, lower the min word to 3 (ft_min_word_len=3) to be able to search for 3 letter words.

Happy searching! :)

Today I have had an technological revelation of the highest sort. The kind of when light shines from heavens and you can almost hear the techno-angles choir go AAAAAAAAAAAaaaaaaa.

So there I was in my bed today, having breakfast. I popped open iPhone (more on that in some future post) and it told me that new Facebook App is available. The 3.0.2 version. So I upgraded. But, it turns out that there is some bug, and the upgraded App is just working somehow strange, things are out of place, the Home screen is moved too much up, and generally it is not good.

So what do I do? My reflex reaction just wanted to fire up Google to see what's going on, but since Facebook App was so fresh, there was no way people wrote about it already on blogs or wherever, and more than that, Google could not index it in those few seconds. So I fire up Twitterrific and do a quick search on Twitter. And in the next 10 minutes angels started singing.

First what happened is that Twitter started buzzing about new Facebook App 3.0.2. Few minutes later, everyone started reporting problems. And few minutes after that, there were quite a few Tweets on how to solve that problem. All of that happened in less than 10 minutes. In that time, Google was not even AWARE that Facebook had new App on the Internet, let alone that it is buggy, and even more no-no how to solve the problem.

( Delete the App and install the fresh one. If you just upgrade it, then it is screwed up)

The real time information sharing and more importantly information finding is something Google just cannot do. I was thinking how could Google even attempt to do this using only their own infrastructure, and maybe if they could parse emails from Gmail they could come up with some results. But people would instantly go into "this is my private stuff", so that is ruled out.

Digging through old articles and files as Google does is fine and necessary, but this real time information is becoming more and more important to people.

All of this just leads to a question; why isn't Twitter already under Google? There were some talks earlier but they failed. In my opinion, whatever money Twitter asks, Google can pay. And it would be well worth it.

And to keep you more puzzled what are we doing over the summer, here are some screen-shots that are "Almost visible, but not revealing too much™":

Yes, those are some of the screens for the new Music Television regional site taken directly off of our hard-drives. The site will cover an area starting from Slovenia and ending somewhere there far on the right side of the map, rather close to Black sea (Macedonia). Every country in the region will have it's own website, due to language differences but more importantly due to different, and numerous, music events in each country. The sites will cover all music types, all regional artists, all events, will have a lot of community features in start (that will grow as the website community grows), be in full support of Television programme with series description and airing times, and much more.

Expect to see this beast up and alive somewhere in late autumn 2009.

Some time ago Adobe released interesting RIA application for cross browser testing - BrowserLab. From what I can see, basically - server (with some sort of capturing service in background) captures screenshot of your website in different browser engines (IE7, FF3 Win/OSX, Safari OSX) and then Flex application displays them side by side or in onion (overlay) mode for you to scroll or zoom in/out.

I have sincere doubts in usefulness of whole "crossbrowser debugging through overlayed screenshots" concept. Here is couple of reasons:

• you can only remotely suspect which dom element is causing problems, can't be for sure, you are watching screenshot after all
• long sites die in all browser screen captures (eg. nivas.hr is REALLY long and it's a good example of horridly long page which fails to be captured)
• if webpage you are screenshotting is being generated by javascript on load event (eg embeding flash movies, ajax or whatnot), different timings caused by different rendering engine speeds will cause very different screenshots
• show me one website which will correctly display in overlay mode in all engines BrowserLab supports, and I'll eat this post. :)

But off course, I am not xhtml guru, this is taken from programmers standpoint so maybe I am missing something crucial.