A few philosophical thoughts about QlikView extensions

This post represents some afterthoughts appeared after my attempt to create extension for annotating QlikView applications -- Explainum Feeds.

If you have read some of my previous posts then you probably noticed that I'm not among the most passionate QlikView fans. I worked long time with BusinessObjects and Cognos. I've seen some other BI tools -- Microstrategy, Tableau, Oracle BI to name a few, and my long-term enthusiasm about BI made me read numerous posts and articles about products and trends in BI industry (which looks stagnating now). So I've seen various approaches to solving the same tasks and not all of them were bad.

Without any doubts, QlikTech has offered interesting approach for data analysis -- I'm talking about so-called "associative data model". This is really fresh thinking (if we can call "fresh" an idea which went live 15 years ago) in rather narrow-minded BI world which seems long ago forgot what decision-support system means (if ever knew). I believe that company which once were able to create something original, sooner or later should be able to produce something really game-changing again. It looks like since inventing the associative in-memory engine long time ago, QlikTech made nothing innovative of the same degree again. Their old-school desktop app with cumbersome, overloaded properties forms looks outdated compared to industry peers who did much better job in terms of going web and polishing ergonomics. Recent QlikTech's attempt to create and use metadata have produced something raw and basic, which is years behind of what other have done. Security system still is a cry. Etc, etc.

But one thing has really big potential -- the extensions, which appeared a year ago in QV10. Extensions can become the next QlikTech's disrupting innovation in BI world. If not overlooked and done right, surely.

For those who don't know -- QlikView extensions is a framework based on Javascript API that allows creating custom data visualization objects in QlikView applications -- custom charts, maps, various gadgets and widgets, etc. In my previous post you can find some useful links and notes that can give more technical insight.

Extensions can do for QlikView what neither R&D investment nor sales staff can -- rapid transformation into ubiquitous application platform. It is known that QlikTech had long time rather small R&D team (this is one of reasons why QlikView's UI is not so modern). But extensions can in short term involve into product development many smart people with bright ideas which could never appear in walls of R&D labs. Who would have used iPhones if only Apple made applications for them? What would have happened to Twitter and Facebook if they didn't open API? Mega-popular Farmville and Angry Birds have been created by external developers, not by Zuck's engineers.

Of course, building developers ecosystem is not an easy task. Developers need well-documented, reliable and powerful API. They need clear and rewarding cooperation model. Both users and developers need convenient marketplace that would allow users to browse, install and update extensions and allow developers to find users, get feedback and may be earn some money.

Today QlikTech is still far from all of this. When we made Explainum Feeds -- many people have shown interest to it because many need annotations in their apps. But almost all of them lost interest when they knew that it works only in slow and inconvenient WebView mode and last releases are not supported because they have bug in the extensions' Javascript API. Also, building extension wasn't an easy task in the beginning because of lack of good API documentation and very limited feature set of Javascript API itself (because too long time QlikTech made bet on Visual Basic, as we know).

I'm sure QlikTech knows about all these problems. I'm sure they also think that extensions have big potential (and they're going to offer us something new in QV11). I hope they understand that this time in order to make outstanding innovation they need not only technical performance but organizational as well (by the way, QlikTech completely ignored my request for trial keys for QV Server which we needed to test the extension -- not the best way to support extension developers).

I hope they would be able to make something game-changing once again. Because not many revolutionary innovations are happening in BI industry today. Not too many to neglect them.