Hacker News Reader

Comment by bunkat | Hacker News Reader

I'm not sure how to get excited by this. Seems weird that customers had to wait until version 3 to get a bubble chart and some basic animation events (though it seems you are still on your own to figure out the transitions). After using d3, charts built like this just feel so limiting and a bit old fashioned - I think it is a testament to the amazing work that Mike Bostock has done.

View on HN

As a customer, having few fancy charts isn't that important. But the ease with which allows customizations and the support it provides is more important. D3.js, GoogleCharts, Flotr2, etc. are too complex to customize. When it comes to building a product and not care about how to get your charts done, Highcharts is the best solution.

I think this is a common misconception. Once you wrap your head around d3, building charts is just about as easy as figuring out the required configuration for a Highcharts chart (even complex charts are generally only 30 lines of code). If you really don't care how it is done, there are solutions like nvd3.js that have created prepackaged charts. Based on the extremely limited customizations available and poor performance characteristics, along with the pricing model, I don't think I can agree that Highcharts is ever the best solution.

> Once you wrap your head around d3

Key point here. Highcharts' customers are companies and website builders that don't have, get, or take the time to wrap their heads around D3, they just need to chuck a graph or two on their site(s), then move on to the next thing. It's not idealistic, but it's the reality.

I'm personally surprised there's not a lot more commercial JS libraries like Highcharts out there, seems to me there's a big market for those.

phpChart (http://phpchart.net) is a commercial charting solution but based on jqplot, a FOSS.

FusionCharts (http://fusioncharts.com) is another successful commercial product for charts.

Also, D3.js doesn't work on IE, which is a non-started for our app. We're currently evaluating and will probably end up using Highcharts.

Thanks for the link to nvd3.

I'm in the not-enough-time-to-understand-d3 camp, but maybe that will change. In the meantime, I've found it easy to get something going fast with the highcharts tools.

I think highcharts have better support for touch devices (ie iOS) than the open source options, as far I know...

I just spent the past two months building an advanced dashboard revolving around complicated charts. After looking at both high charts and d3, it became clear that d3 was the better solution, as it allowed an order of magnitude of increased flexibility.

Not to mention that (at least compared to the previous release of HighCharts) D3's performance is just so much better. Graphing tens of thousands of data points causes my browser to freeze up for 10+ seconds at a time with HighCharts, while a D3-based graph visualization of the same data loads in under a second.

It doesn't appear to have gotten much better in this version. I took a peek at the new bubble chart (via their link to a jsFiddle) and it took about 7 seconds to render. I'm drawing similar charts using d3 with 800 data points that render in under 300 ms.

That's 7 seconds waiting for JSFiddle, not Highcharts. In my Chrome browser, Highcharts renders 800 bubbles in under 200 ms: http://jsfiddle.net/highcharts/hAyzq/

The last thing the world needs is more bubble charts. I'd like once to see one in a true scientific paper. Bubble charts are just like dowsing rods. That's why marketers and managers love them. They'll show you whatever you're looking for.

Reply on news.ycombinator.com