You'll laugh at me, but I think that one of the most critical aspects of a graphing library is whether it can anti-alias or not.
I interned with a company a few years ago and we tested graphing packages; we found that although A was superior technically to B, B was preferred by the users because it anti-aliased the graphs that it rendered. Technically it made things much more complicated, but looking better goes a very long way.
I've been using Flot a while after getting the hump about JS charting SDKs last week. A clever person I work with has ported Flot to YUI (http://github.com/bluesmoon/flot/tree/master).
I have to say I particularly like it, and I think we will be offering back some improvements to the jQuery version, such as the ability to have many more graphs on the page with destroying your browser.
I do really like the idea of Tufte graphs though. As a Tufte devotee I think I'm going to look at getting that look and feel out of flot as well.
One of the biggest problems with using Canvas to render graphs is the lack of exporting functionality. I built a reporting tool over the summer that used Flot, but ended up having to switch it with Google Charts because users couldn't copy/paste results into Word.
Image rendering may not be as dynamic as Canvas rendering, but the ability to export images and the widespread browser support still make it a very viable solution
Flot looks good but I'm concerned about its use of a Canvas emulator in IE (as opposed to writing natively to VML). Such things often don't work well, and we can't afford to have second-class behavior in IE.
Can anyone here speak to whether that is a problem or not?
Alternatively, does anybody have a different favorite charting library they'd care to mention?
I maintain a fork of flot on github. The graphs function correctly in IE (barring some here-and-there bugs which you can report). The only issue is speed (always an issue with IE), and as a result I wouldn't try too much of the interactivity... But on the whole it works great.
That's an excellent way to put it. Although sometimes you'll want to graph things that are not necessarily candidates for improvement, just to make sure the graphs stay in the acceptable range.
One of my few non-public graphs is refund rates. Normally, that is in the 2 ~ 3% range, and reducing that is not a major priority of mine. However, spikes in the graph almost always mean "You just screwed something up", so THAT information I want to see in a hurry.
I think using Google Chart API is much better than using these plugins. Lots of chart types, easy to make, gives high-quality result, and support all browsers that can display images.
If you ask me (I maintain a fork of flot on github) there is no reason to use a client-side library for graphing unless the graphs are dynamic. Requesting/rendering/etc tons of images is painful, flot can add and remove data on the fly and re-render -- and that's not even getting into the potential for hovering effects and the like (flash-style).
Flot also supports all reasonable browsers (including some unreasonable ones, like IE6). If you find a bug in either the 'official' flot or my fork (or any of the forks of my fork :)) then you should be able to report it and have it fixed quickly enough (or, fix it yourself, the code isn't especially complicated).
A couple months ago I built a reporting tool for a client using Flot, only to have to switch over to Google Charts after users started complaining.
Flot, and all other Canvas-based JS charting solutions, are really impressive applications of Canvas, but until browsers start supporting Canvas copy/paste better I'm hesitant to use them for graphing apps. Image-rendering graphing solutions support most browsers and can be easily copy/pasted into other applications, with little additional overhead.
